CN201029876Y - Navigation system for bone surgery - Google Patents

Abstract

The utility model relates to a medical treatment apparatus, in particular to a bone surgery guidance system used for fracture restoration surgeries of enclosed intramedullary nails. The bone surgery guidance system includes a binocular stereo vision sensing device and markers. The binocular stereo vision sensing device comprises a computer with a main frequency which is not below 1G, a locater arranged on the frame of the locater, two CCD cameras equipped in the locator, a two-path signal synchronizer. The markers are provided with at least three active light-emitting diodes arranged on different lines, wherein one marker is rigidly inserted and connected with intramedullary nails, the other marker is rigidly inserted and connected with surgery single-purpose electric drill. The aim of the utility model is to provide a bone surgery guidance system, and to achieve exact location of relative position relation between surgery instruments and focus positions of patients, leading that surgery physicians can accurately drill intramedullary nail hinge holes at one time guided by the computer screen.

Description

Orthopaedics operation navigation system

Technical field

This utility model relates to a kind of medical treatment device, relates in particular to a kind of orthopaedics operation navigation system that is used for the bone fracture restoration operation of closed intramedullary pin.

Background technology

At present, in orthopedics department operation, long bone fracture occupies very big ratio, and what at present long bone fracture is treated the most normal employing is closed intramedullary pin internal fixation technology.The intramedullary pin internal fixation is a kind of " biological " preferably internal fixation technology.Locking-type intramedullary pin internal fixation is adopted in femur and fracture of tibia operation, and effect is very good, has become the preferential method of selecting of domestic and international treatment femur and tibia fracture at present.This technology is helped the healing of knochenbruch by inserting intramedullary pin, has reduced the further damage of fracturing, and has increased the stability of fracture go-and-retum, reasonable distribution the physiological stress between sclerite, the quickly-healing that helps fracturing.When adopting this technology, conventional method is divided following step at present: 1. take X-ray according to patient's affected part situation; 2. select suitable intramedullary pin model; 3. cut joint; 4. bore open pore; 5. insert guide wire; 6. hole drilling and reaming is to the intramedullary pin diameter; 7. implant intramedullary pin; 8. beat intramedullary pin two ends strand lock screw hole; 9. fastening four screws; 10. sew up the incision.Laterally strand lock screw is used for connecting the bone and the intramedullary pin of two break-off portion, prevents rotation and the displacement of intramedullary pin in inside bone.But intramedullary pin inserts after the long bone inner chamber, because the position of strand lockhole and direction can't definitely be learned in the process of operation, therefore in the operation, the doctor often needs to use C type arm X line image to strengthen instrument, constantly take the X-ray sheet, check according to the image of X sheet whether punch position and direction be correct, overlap with the axial line that reaches strand lockhole and screw.Particularly in far-end strand cable process.Far-end strand cable---implantation of cross screw is to prevent the rotation of intramedullary pin---is generally acknowledged it is one of challenging step of tool in the knochenbruch operation for a long time.Because intramedullary pin will consistent its distortion with the skeleton cavity shape reach several millimeters, the accurate location of far-end strand cable screw hole axis be can not determine in the past.By constantly in front with the side alternately X-ray line image observe, the entrance that the surgeon adjusts electric drill with direction so that make the axis of electric drill and the screw hole axis is consistent accordingly.By inspection in advance, the benefit of boring is more and more obvious with a pair of X-ray sheet.In case the screw hole that bullport had bored by the far-end on the intramedullary pin, strand cable screw just are easy to be fixed.Its complication comprises wearing out of the breaking of unsuitable fixing, the rotation that should not have, bone, cortical layer and owing to repeatedly bores and beat or enlarge the skeleton that bullport causes and weaken.

From analysis, can find following problem to the whole surgery process:

1. x-ray image is used to determine the operation anatomical position, because image is immobilized, the actual area that shows is limited, so is the just necessary frequent use perspective in the definite position that obtains the strand lockhole, thereby causes surgeon and patient repeatedly to suffer the infringement of X-radiation.

2. in the operation during applying X-ray fluoroscopy system assist location, the doctor can only observe the monoplane view, and in the time need observing the position of operating theater instruments on many plan views, constantly the location is scanned in the position of re-adjustments C type arm, cause operation to interrupt, and waste time and energy.

3. the surgeon must be with the experience of oneself, and the reconstruct screw does not have visual real-time feedback tool to handle operation tool (electric drill) screw is squeezed in the strand lockhole exactly with respect to the position of strand lockhole brains on time and space.

This method by rule of thumb often is inaccurate, tediously long process, thereby causes inaccurate fixing, rotation improperly, bone splits, cortex to penetrate.Particularly in the case of some patient's condition complexity, because of the incorrect meeting of once punching causes many places punching or screw hole to increase bone strength is died down, patient performs the operation painful big.In addition, influenced by anatomical position and skill, the surgeon directly is subjected to x radiation x in each operation time, wherein 30%～51% time spent in the punching of strand lock screw hole at 3～30min.The strand lockhole location difficulty of intramedullary pin, be difficult to accurate locking, to cause the prolongation of operating time and influence fixed effect, these problem already cause many in the world expert professors' attention, carried out the research of surgical navigationals such as joint replacement, tooth are changed, facial rectification in succession, and also do not study at femur, the such operation of fracture of tibia, particularly in China, only the principle of this type systematic is carried out the discussion of some theoretical sides, also deeply do not carried out Research on Practical Application.Improve the operation precision, shorten operation process by means of operation guiding system, minimizing patient especially doctor is very important the open-assembly time under X ray.The research and development operation guiding system also becomes extremely urgent problem.

Summary of the invention

In order to solve the problems referred to above that exist in the existing bone surgery, goal of the invention of the present utility model provides orthopaedics operation navigation system, realization is to the accurate location of the relative position relation of operation tool and patient's lesions position, promptly to calculate the space coordinate of impact point, make surgical doctor under the guidance of computer screen, correctly once play the success of intramedullary pin strand lockhole, effectively reduce operating time, operation wound, x-ray radiation etc., and can significantly improve success rate of operation, alleviate the misery of sufferer.

In order to reach above-mentioned goal of the invention, this utility model has adopted following technical scheme:

Orthopaedics operation navigation system, comprise binocular stereo vision sensing device and marker, described binocular stereo vision sensing device comprises: dominant frequency is not less than the computer of 1G, the adjustable locator bracket of multi-directionally, be arranged on the localizer on the locator bracket, be arranged on two ccd video cameras in the localizer, image pick-up card, the double-channel signal lock unit that links to each other with image pick-up card, described marker is 2, described marker is provided with not at least 3 active illuminating diodes on same straight line, and inflexible plugging is connected with intramedullary pin on described one of them marker, and inflexible the plugging with the special-purpose electric drill of operation of described wherein another marker is connected.

As preferably, described ccd video camera is high-resolution colourful CCD video camera.

As preferably, the ccd array of described ccd video camera has 768 * 576 sensitization elements, the effective focal length f=8mm of camera lens.

As preferably, the parallax range that is arranged between two ccd video cameras in the localizer is L, L=280mm.

As preferably, described marker is triangular in shape, is respectively equipped with 1 active illuminating diode on three angles of described marker.

As preferably, described marker is star, is respectively equipped with 1 active illuminating diode on four angles of described marker.

As preferably, described locator bracket adopts universal tripod.

Below the intramedullary pin, the foundation of electric drill coordinate system of a kind of orthopaedics operation navigation system of providing for this utility model:

In the Fracture of femur surgical navigational, the binocular space positioning system obtains the space coordinates position of these index points in the operation.In order to demonstrate the spatial relation of drill bit real-time and accurately, need set up the coordinate transformation relation between them with respect to intramedullary pin.When Locating System with Binocular whenever collects piece image, after a series of filtering, feature extraction, center of gravity calculation etc., obtain the 3 d space coordinate of this each index point of moment, its relative position relation and coordinate system are as shown in Figure 3.Wherein the installation site of marker and coordinate concern as shown in Figure 4 on the intramedullary pin.

Definition O ₁Point is the initial point of intramedullary pin coordinate system, and its coordinate in world coordinate system is: $(\frac{x_{A_1}+x_{{\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">B</figure-callout>}}_1}}{2},\frac{{\mathrm{<figure-callout\; id="1"\; label="y\; A"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">y</figure-callout>}}_{A_{}}+{\mathrm{<figure-callout\; id="1"\; label="y\; B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">y</figure-callout>}}_{B_{}}}{2},\frac{z_{A_1}+{\mathrm{<figure-callout\; id="1"\; label="z\; B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_{B_{}}}{2}),$ x ₁The pros of axle are defined as O ₁B ₁Direction, z ₁The positive direction of axle is defined as O ₁B ₁* O ₁C ₁Direction, y ₁Forward be (O ₁B ₁* O ₁C ₁) * O ₁B ₁Direction.

If x ₁, y ₁, z ₁The unit vector of axle is respectively i ₁, j ₁, k ₁, their direction cosines in world coordinate system are used u respectively _1i, u _2i, u _3i(i=1,2,3) expression.

To x ₁Axle, u ₁₁, u ₁₂, u ₁₃Can obtain by following formula:

${\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">i</figure-callout>}}_1=\frac{O_1{B}_1}{\left|O_1{B}_1\right|}=\frac{(x_{B_1}-x_{A_1})I+(y_{B_1}-y_{A_1})J+(z_{B_1}-z_{A_1})K}{\sqrt{{(x_{B_1}-x_{A_1})}^2+{(y_{B_1}-y_{A_1})}^2+{(z_{B_1}-z_{A_1})}^2}}$

$=u_{11}i+u_{12}j+u_{13}k---(4-1)$

$\left\{\begin{array}{c}{u}_{11}=\frac{{x_B}_1-{x_A}_1}{\sqrt{{(x_{B_1}-x_{A_1})}^2+{(y_{B_1}-y_{A_1})}^2+{(z_{B_1}-z_{A_1})}^2}}\\ u_{12}=\frac{{y_B}_1-{y_A}_1}{\sqrt{{(x_{B_1}-x_{A_1})}^2+{(y_{B_1}-y_{A_1})}^2+{(z_{B_1}-z_{A_1})}^2}}\\ u_{13}=\frac{{z_B}_1-{z_A}_1}{\sqrt{{(x_{B_1}-x_{A_1})}^2+{(y_{B_1}-y_{A_1})}^2{+(z_{B_1}-z_{A_1})}^2}}\end{array}\right.---(4-2)$

Wherein:

I, j, k are the unit vector of 3 axles of world coordinate system.To z ₁Axle, u ₃₁, u ₃₂, u ₃₃Can obtain by following formula:

${\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">k</figure-callout>}}_1=\frac{O_1{\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">B</figure-callout>}}_1\&times;O_1{\mathrm{<figure-callout\; id="1"\; label="C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">C</figure-callout>}}_1}{|O_1{\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">B</figure-callout>}}_1\&times;O_1{\mathrm{<figure-callout\; id="1"\; label="C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">C</figure-callout>}}_1|}=u_{31}i+u_{32}j+u_{33}k---(4-3)$

Wherein: $O_1{\mathrm{<figure-callout\; id="1"\; label="C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">C</figure-callout>}}_1=(x_{C_1}-\frac{x_{A_1}-x_{B_1}}{2})i+(y_{C_1}-\frac{y_{A_1}-y_{B_1}}{2})j+(z_{C_1}-\frac{z_{A_1}-{\mathrm{<figure-callout\; id="1"\; label="z\; B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_{B_{}}}{2})k$

$\left\{\begin{array}{c}{u}_{31}=\&lsqb;(z_{B_1}-z_{C_1})y_{A_1}+(z_{C_1}-z_{A_1})y_{B_1}+(z_{A_1}-z_{B_1}){\mathrm{<figure-callout\; id="1"\; label="y\; C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">y</figure-callout>}}_{C_{}}\&rsqb;/{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">d</figure-callout>}}_1=a_1/d_1\\ u_{32}=\&lsqb;(x_{B_1}-x_{C_1})z_{A_1}+(x_{C_1}-x_{A_1})z_{B_1}+(x_{A_1}-x_{B_1})z_{{\mathrm{<figure-callout\; id="1"\; label="C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">C</figure-callout>}}_1}\&rsqb;/{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">d</figure-callout>}}_1=b_1/d_1\\ u_{33}=\&lsqb;(y_{B_1}-y_{C_1})z_{A_1}+(y_{C_1}-y_{A_1})z_{B_1}+(y_{A_1}-y_{B_1}){\mathrm{<figure-callout\; id="1"\; label="z\; C"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_{C_{}}\&rsqb;/{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">d</figure-callout>}}_1=c_1/d_1\end{array}\right.---(4-4)$

Here: ${\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">d</figure-callout>}}_1=\sqrt{a_1^2+b_1^2+c_1^2}---(4-5)$

To y ₁Axle, u ₂₁, u ₂₂, u ₂₃Can obtain by following formula:

${\mathrm{<figure-callout\; id="1"\; label="j"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">j</figure-callout>}}_1={\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">k</figure-callout>}}_1\&times;i_1=\left|\begin{array}{ccc}i& j& k\\ u_{31}& u_{32}& u_{33}\\ u_{11}& u_{12}& u_{13}\end{array}\right|=u_{21}i+u_{22}j+u_{23}k$

Thereby: $\left\{\begin{array}{c}{u}_{21}=u_{32}{u}_{13}-u_{12}{u}_{33}\\ u_{22}=u_{33}{u}_{11}-u_{13}{u}_{31}\\ u_{23}=u_{31}{u}_{12}-u_{11}{u}_{32}\end{array}\right.---(4-6)$

1. coordinate transformation relation

Owing to represented coordinate system o in the formula ₁-x ₁y ₁z ₁With respect to world coordinate system o _w-x _wy _wz _wRotation amount, and the some coordinate of D in world coordinate system is (x _D1, y _D1, z _D1) represented intramedullary pin coordinate system o ₁-x _1y1 _Z1With respect to world coordinate system o _w-x _wy _wz _wTranslational movement, the coordinate of D in the intramedullary pin coordinate system of setting up an office is (x ₁Dx, y ₁D ₁, z ₁D ₁), then:

$\left[\begin{array}{cccc}{x}_{D_1}& y_{D_1}& {\mathrm{<figure-callout\; id="1"\; label="z\; D"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_{D_{}}& 1\end{array}\right]={\left[\begin{array}{cccc}{x}_{{1D}_1}& y_{{1D}_1}& z_{{1\mathrm{<figure-callout\; id="1"\; label="D"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">D</figure-callout>}}_1}& 1\end{array}\right]}^1{T}_w---(4-7)$

${\mathrm{<figure-callout\; id="1"\; label="T\; W"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">T</figure-callout>}}_W_1=\left[\begin{array}{cccc}{u}_{11}& u_{12}& u_{13}& 0\\ u_{21}& u_{22}& u_{23}& 0\\ u_{31}& u_{32}& u_{33}& 0\\ \frac{x_{{\mathrm{<figure-callout\; id="1"\; label="A"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">A</figure-callout>}}_1}+x_{{\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">B</figure-callout>}}_1}}{2}& \frac{y_{A_1}+{\mathrm{<figure-callout\; id="1"\; label="y\; B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">y</figure-callout>}}_{B_{}}}{2}& \frac{z_{{\mathrm{<figure-callout\; id="1"\; label="A"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">A</figure-callout>}}_1}+{\mathrm{<figure-callout\; id="1"\; label="z\; B"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_{B_{}}}{2}& 1\end{array}\right]---(4-8)$

Matrix ¹T _wBe exactly by intramedullary pin coordinate system o ₁-x ₁y ₁z ₁(local coordinate system) is to world coordinate system o _w-x _wy _wz _wTransformation matrix of coordinates.

In like manner can be by electric drill coordinate system o ₂-x ₂y ₂z ₂Transformation matrix to world coordinate system ²T _wFor:

${\mathrm{<figure-callout\; id="1"\; label="T\; w"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">T</figure-callout>}}_w_1=\left[\begin{array}{cccc}{u}_{11}^{\&prime;}& u_{12}^{\&prime;}& u_{13}^{\&prime;}& 1\\ u_{21}^{\&prime;}& u_{22}^{\&prime;}& u_{23}^{\&prime;}& 0\\ u_{31}^{\&prime;}& u_{32}^{\&prime;}& u_{33}^{\&prime;}& 0\\ \frac{x_{A_2}+x_{{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">B</figure-callout>}}_2}}{2}& \frac{y_{A_2}+{\mathrm{<figure-callout\; id="2"\; label="y\; B"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">y</figure-callout>}}_{B_{}}}{2}& \frac{z_{A_2}+{\mathrm{<figure-callout\; id="2"\; label="z\; B"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">z</figure-callout>}}_{B_{}}}{2}& 1\end{array}\right]---(4-9)$

Wherein: $\left\{\begin{array}{c}{u}_{11}^{\&prime;}=\frac{{x_B}_2-{x_A}_2}{\sqrt{{(x_{B_2}-x_{A_2})}^2+{(y_{B_2}-y_{A_2})}^2+{(z_{B_2}-z_{A_2})}^2}}\\ u_{12}^{\&prime;}=\frac{{y_B}_2-{y_A}_2}{\sqrt{{(x_{B_2}-x_{A_2})}^2+{(y_{B_2}-y_{A_2})}^2+{(z_{B_2}-z_{A_2})}^2}}\\ u_{13}^{\&prime;}=\frac{{z_B}_2-{z_A}_2}{\sqrt{{(x_{B_2}-x_{A_2})}^2+{(y_{B_2}-y_{A_2})}^2{+(z_{B_2}-z_{A_2})}^2}}\end{array}\right.---(4-10)$

$\left\{\begin{array}{c}{u}_{31}^{\&prime;}=\&lsqb;(z_{B_2}-z_{C_2})y_{A_2}+(z_{C_2}-z_{A_2}){\mathrm{<figure-callout\; id="2"\; label="y\; B"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">y</figure-callout>}}_{B_{}}+(z_{A_2}-z_{B_2}){\mathrm{<figure-callout\; id="2"\; label="y\; C"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">y</figure-callout>}}_{C_{}}\&rsqb;/{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">d</figure-callout>}}_2=a_2/d_2\\ u_{32}^{\&prime;}=\&lsqb;(x_{B_2}-x_{C_2})z_{A_2}+(x_{C_2}-x_{A_2})z_{B_2}+(x_{A_2}-x_{B_2})z_{{\mathrm{<figure-callout\; id="2"\; label="C"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">C</figure-callout>}}_2}\&rsqb;/{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">d</figure-callout>}}_2=b_2/d_2\\ u_{33}^{\&prime;}=\&lsqb;(y_{B_2}-y_{C_2})z_{A_2}+(y_{C_2}-y_{A_2})z_{B_2}+(y_{A_2}-y_{B_2}){\mathrm{<figure-callout\; id="2"\; label="z\; C"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">z</figure-callout>}}_{C_{}}\&rsqb;/{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">d</figure-callout>}}_2={\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">c</figure-callout>}}_2/d_2\end{array}\right.---(4-11)$

${\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">d</figure-callout>}}_2=\sqrt{a_2^2+b_2^2+{\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">c</figure-callout>}}_2^2}---(4-12)$

$\left\{\begin{array}{c}{u}_{21}^{\&prime;}=u_{32}^{\&prime;}{u}_{13}^{\&prime;}-u_{12}^{\&prime;}{u}_{33}^{\&prime;}\\ u_{22}^{\&prime;}=u_{33}^{\&prime;}{u}_{11}^{\&prime;}-u_{13}^{\&prime;}{u}_{31}^{\&prime;}\\ u_{23}^{\&prime;}=u_{31}^{\&prime;}{u}_{12}^{\&prime;}-u_{11}^{\&prime;}{u}_{32}^{\&prime;}\end{array}\right.---(4-13)$

By ¹T _wWith ¹T _wWe can calculate any some three-dimensional coordinate in world coordinate system on intramedullary pin and the electric drill, we can calculate the some coordinate in the intramedullary pin coordinate system on the electric drill according to formula (4-14) again, thereby can clearly observe electric drill with respect to the position of intramedullary pin and the correctness of direction.

$\left[\begin{array}{cccc}{x}_1_2& y_1_2& {\mathrm{<figure-callout\; id="1"\; label="z"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_1_2& 1\end{array}\right]=\left[\begin{array}{cccc}{x}_2& y_2& {\mathrm{<figure-callout\; id="2"\; label="z"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">z</figure-callout>}}_2& 1\end{array}\right]\&CenterDot;T_1_2$

$=\left[\begin{array}{cccc}{x}_2& y_2& {\mathrm{<figure-callout\; id="2"\; label="z"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">z</figure-callout>}}_2& 1\end{array}\right]\&CenterDot;T_w_2\&CenterDot;T_w_{-1}^1---(4-14)$

2. application example

Known intramedullary pin and x ₁The angle of axle is α ₁, apart from o ₁x ₁y ₁Planar distance is h ₁, four screw holes that are used for clavicle are arranged on the intramedullary pin, the coordinate of centrage in the intramedullary pin coordinate system of its mesopore m is (L ₁Cos α ₁, L ₁Sin α ₁, h ₁).The coordinate system and the intramedullary pin of electric drill are similar, and chisel edge is made as L apart from the distance at coordinate center ₂, the z of center line of bit ₂Coordinate is h ₂

Obtaining each index point by the binocular space positioning system in certain test at a certain instantaneous world coordinates is:

$\left[\begin{array}{cccccccc}{x}_1& x_2& x_3& x_4& x_5& x_6& x_7& x_8\\ y_1& y_2& y_3& y_4& y_5& y_6& y_7& y_8\\ z_1& z_2& z_3& z_4& z_5& z_6& z_7& {\mathrm{<figure-callout\; id="8"\; label="z"\; filenames="Y20072010557300231.png"\; state="\{\{state\}\}">z</figure-callout>}}_8\end{array}\right]=$

$\left[\begin{array}{cccccccc}18.28& 82.17& 25.69& 88.44& 221.63& 285.64& 212.96& 275.89\\ -48.97& -56.90& -50.11& -57.59& -75.10& -81.75& -75.59& -83.65\\ 0.15& 7.72& 64.40& 71.33& 340.26& 335.10& 403.77& 298.58\end{array}\right]$

Thereby try to achieve from the intramedullary pin coordinate and be tied to being transformed to of world coordinate system:

${\mathrm{<figure-callout\; id="1"\; label="T\; w"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">T</figure-callout>}}_w_1=\left[\begin{array}{cccc}0.6994& -0.0859& 0.7096& 0\\ 0.7040& -0.0885& -0.7046& 0\\ 0.1233& 0.9924& -0.0014& 0\\ 53.3600& -53.2800& 35.7400& 1\end{array}\right]$

Be tied to being transformed to of world coordinate system from the electric drill coordinate:

${\mathrm{<figure-callout\; id="2"\; label="T\; w"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">T</figure-callout>}}_w_2=\left[\begin{array}{cccc}0.7869& -0.1240& -0.6045& 0\\ 0.6094& 0.0024& 0.7928& 0\\ -0.0968& -0.9923& 0.0775& 0\\ 248.7600& -79.3750& 319.4200& 1\end{array}\right]$

The electric drill coordinate is tied to being transformed to of intramedullary pin coordinate system:

$T_w={\mathrm{<figure-callout\; id="2"\; label="T\; w"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">T</figure-callout>}}_w_2\&CenterDot;_2{\left[T_w_1\right]}^{-1}=\left[\begin{array}{cccc}0.1321& 0.9909& 0.0252& 0.0000\\ 0.9886& -0.1299& 0.0765& -0.0000\\ 0.0725& -0.0350& -0.9968& -0.0000\\ 340.1901& -60.0197& -2.1908& 1.0000\end{array}\right]$

Obtain the equation in each comfortable intramedullary pin coordinate system of intramedullary pin and electric drill, their projection lines in each plane of intramedullary pin coordinate system draw, as shown in Figure 5, can clearly be seen that whether electric drill correctly aims at the hole on the intramedullary pin, and can obtain error angle and be presented on the screen.

According to the demonstration of computer screen, operative doctor can very clearly be found out the offset direction and the deviation angle of electric drill, and two perspective planes are arranged during practical application, sees Fig. 6, can instruct the doctor accurately to punch.

It below is the software design of orthopaedics operation navigation system of the present utility model

1. computational tool and development environment

The various algorithms of date processing finally will be realized by computer software.In the software development design, programmed environment is to develop active platform.The quality that hardware environment is selected is directly connected to efficient, software quality, the success or failure of software programming and the result of use of software of software design work.This spatial point positioning system using C++Builder6.0 and MATLAB develop.

C++Builder6.0 is the visual quick application development instrument of object-oriented of new generation (RAD, Rapid Application Development) that Inprise company (former Borland company) releases.It is used for developing application in Windows95, Windows98, Windows NT, Windows2000, is a kind of powerful, effective visual programming tools.C++Builder is a kind of C Plus Plus developing instrument, and it is a traditional C ++ the milestone of developing instrument development.It joins visual development environment, developing instrument and visual component library in the C Plus Plus, has realized real OO exploitation.The function of C++Builder has: the C/C++ developing instrument that object-based programming, I/O INTERFACE DESIGN be simple, first-class development environment is provided, provide standard and Vsual C++ compatible high, provide abundant assembly with the most powerful debugging acid, the complete new function of windows 2000 support, the Internet developing instrument is provided, distribution application system development environment tools or the like is provided.The member of integrated more than 130 various functions in the C++Builder IDE.Thereby C++Builder is a perfect combination of quick application development pattern and reusable member, and it is representing the future thrust of C Plus Plus, all plays an important role in all many-sides of computer realm.

MATLAB is a most widely used mathematical software in the world today, it is with a kind of programming language of complex matrix as basic programming unit, is a highly integrated system---integrate science calculating, image and pictorial display and processing, date processing, acoustic processing, high-caliber graphical interfaces design style.The most function of C or formula translation can easily be realized in the MATLAB language, and the interface with other high-level programming language is provided.Adopt the such specific high level language of MATLAB, not only can greatly improve programming efficiency, programming q﹠r, and can greatly shorten product development cycle.

In actual applications, how both to consider to utilize the friendly at C++Builder interface, the high efficiency that program is carried out, made full use of the terseness of MATLAB programming, the convenience of data operation again.The problem of the interface of Here it is MATLAB and C++Builder, by interface, can be very easily in C++Builder application call that the user developed by MATLAB institute development sequence, promptly realize both hybrid programmings.What the interface that MATLAB provides was commonly used has four kinds:

(1) utilizes the MATLAB engine.This method adopts client and server control mode, utilizes engine that MATLAB and C++Builder are connected.In utilization, the program that the C++Builder environment is developed down is as client, and MATLAB is as the server of this locality.The C++Builder program is to MATLAB engine transferring command and data message, and from MATLAB engine receiving data information.

(2) the compiler .mcc that utilizes MATLAB to carry.Because carry out the purpose of hybrid programming is to develop comparatively complex mathematical algorithm under C++Builder and MATLAB, the most directly solving thought is, the interface of application and other high-level languages directly will be developed the code that good MATLAB m file is compiled as executable program or translates into high-level language.Can break away from the MATLAB environment in this way, and can carry out the transmission of parameter and the demonstration of figure easily, the power of MATLAB is dissolved in the various application programs, and then be improved code execution speed greatly.What this paper used is exactly this method, and the m file that calculates transition matrix is become the executable file that can break away from the MATLAB environment with the m file conversion of calculating three-dimensional coordinate, and then, in C++Builder, call, and result of calculation is presented in the interface.

(3) utilize the Matcom compiling.Utilize Matcom (MIDEVA) the MATLAB code can be compiled as the executable file of dynamic link library (DLL), independent operating or be converted into the CPP source code.

(4) The Component Object Model COM (Component Object Model) is for communicating the standard that provides unified between component software and the application program, because COM is extensive use of, MATLAB also begins to support the COM technology, and provide a typelib file (MATLAB Application (version6.5) Type Library) that is the standard of Mlapp.tlb, two interface IMLApp and DIMLApp have wherein been defined, IMLApp is not open, the internal interface that is called by DIMLApp just is so CLIENT PROGRAM can be visited Matlab by DIMLApp.

2. the population structure of system

Need images acquired in the spatial point navigation system, obtain lot of data, and the Nonlinear System of Equations that the data of impact point are formed is found the solution its process complexity.For the software design clear thinking, readable strong, consider that simultaneously C++Builder6.0 and MATLAB language characteristic take into account the high efficiency and the practicality of programming, carried out system program design.Program structure as shown in Figure 9.

3. system program design

(1) INTERFACE DESIGN and function introduction

According to system construction drawing, adopt the C++Builder6.0 development environment to programme, the Product Interface that comes out with this environment exploitation has good man computer interface.The interface of system calibrating:

Can see that from the interface this software system is mainly finished image acquisition, the camera calibration and the function of the three-dimensional coordinate of display-object point in real time.Do simple the introduction with regard to this software system below.

Beginning to gather the incident that button triggers mainly contains: the collection of image, image binaryzation obtains the image coordinate of impact point and is presented in the text box.

Stopping to gather the incident that button triggers is that the coordinate of fixed point under the world coordinate system of setting is read in the system, prepares for calculating projection matrix as the part of data source.

Beginning to demarcate the incident that button triggers is the executable file that projection matrix is calculated in operation.This executable file is that compiling generates through .mcc by the m file of MATLAB.Execution result exists on the disk, by the space coordinates of calculating impact point is used.

The function finished of beginning navigation button is at first finished and is begun to gather the function that button is finished, the image coordinate of the impact point of acquisition, and the transition matrix that calculated according to the last step then, the executable file of three-dimensional coordinate is calculated in operation, and display result is in text box.

Pause button is to stop images acquired, stops to calculate three-dimensional coordinate.Exit button promptly logs off.

In the realization of system, the function that utilizes MATLAB to write to find the solution Nonlinear System of Equations, and it is compiled into executable file, and the method simple possible, program reliability is strong.

(2) key issue of Design of System Software

In software design, the main key issue that exists has: image acquisition problem, Flame Image Process problem and find the solution three problems of Nonlinear System of Equations.Provide the method for solution below one by one.

1. about the processing of image acquisition problem

The work of native system image acquisition utilizes the Meteor-II/Multi Channal image pick-up card of Canadian Matrox company exploitation, and this capture card has abundant program library---MIL-Lite (basic controlling storehouse), MIL (pattern recognition storehouse).Can under MS Visual C/C++, MS Visual Basic, Borland C/C++, call the powerful image acquisition of its prior function, handle function.Wherein, MIL-Lite provides free use, but it does not comprise the function of image processing and analyzing part.Native system utilizes MIL-Lite control storehouse that image acquisition is arrived computer, carries out subsequent image processing by C++Builder.

Utilize C++Builder to carry out Flame Image Process and have characteristics such as speed is fast, safety good, powerful, call the MIL-Lite of Matrox image pick-up card or the Flame Image Process function in MIL storehouse at C++Builder, can accomplish that integration is good, information is obtained intuitively, programming is convenient, realizes the image pretreatment apace, by obtaining feature, reach the final purpose of Flame Image Process and identification, the framework that its software collection is partly realized as shown in figure 10.

Utilize the MIL-Lite storehouse at first must distribute an application (Application), be equivalent to create the control and the execution environment of Flame Image Process.Under application, can set up a plurality of systems (System).Can set up a plurality of data buffer memorys (DataBuffer), digital converter (Digitizer) and data show (Display) under each system, image file is read in the data buffer memory, give array with the buffer memory assignment, can realize treatment of picture by processing to array.View data with just can show after data show is related by control or forms.So just finished the collecting work of image.Notice that at first the CD installation procedure that carries of operation image capture card carries out decompress(ion) to program package before creation procedure, obtains required library file.The headers and libraries file that the image acquisition of will install is set in the translation and compiling environment at C++Builder comprises to come in.

2. image processing speed problem

In the surgical navigational navigation system, need handle two width of cloth images that collect.For image processing function (comprising image segmentation, target label), finishing function needs one by one pixel to be operated, and it is very important that processing speed just becomes.

In C++Builder, usually piece image is handled (binaryzation, target label etc.), need simply scanogram from the both direction, by each pixel in the Pixels attribute access image of Tcanvas class, and the color of each pixel constantly is set by the Pixels attribute of Tcanvas class in each loop body.Because application program calls the Pixels attribute each time color is set, and all may cause repainting of image, simultaneously this call also constantly with the display memory swap data, this has obviously caused the reduction greatly of the speed of service.

To the display memory operation, common solution is an application drawing picture in internal memory for fear of continually.Its process is: set up an interim bitmap in internal memory, the image information of Image assembly is copied in the interim bitmap, then memory bitmap is carried out Flame Image Process such as same image binaryzation, target label, after pending the finishing, memory bitmap is composed the attribute to Graphic, once the result is shown.This method is avoided continually display memory being operated, thereby speed is promoted greatly.

In native system, used the method for the ScanLine attribute of Tbitmap class.This method need be set up interim bitmap equally in internal memory, but has utilized the ScanLine attribute but not Pixels attribute access data.This attribute can return pointer that point to realize the bitmap internal memory, and once visits and handle full line data of bitmap by this pointer, thus the speed of service of faster procedure greatly.Below provided and utilized the ScanLine attribute to carry out the program description of binary conversion treatment:

bool TmainForm::BinoValue(Graphics::Tbitmap*Bitmap，intyuzhi)

{

// interim memory bitmap Bitmap is pending bitmap, and yuzhi is the threshold value thresholding

BYTE*ptr；

int g＝yuzhi；

Bitmap-〉PixelFormat=pf8bit; The storage inside form of // bitmap is the pf8bit gray level image

for(int y＝0；y<Bitmap->Height；y++)

{

ptr＝(BYTE*)Bitmap->ScanLine[y]；

for(int x＝0；x<Bitmap->Width；x++)

{

if(ptr[x]>＝g)

ptr[x]＝255；

else

ptr[x]＝0；

}

}

return true；

}

To R channel image sampling 50 times, carry out Flame Image Process and consume shown in table 5-1 its average time.Identical with the R passage for G channel algorithm principle, time loss is very nearly the same.The native system time mainly consumes in image acquisition work as can be seen by table 5-1.

The table 5-1R channel image processing time consumes

Classification	Time (unit: ms)
		Binaryzation target label (containing center calculation) summation is caught image	3.376 8.899 12.275 186.460

3. find the solution the Nonlinear System of Equations problem

By system software structure Fig. 9 as can be known, obtain the centre coordinate of impact point after, carry out the calculating of projection transition matrix, and the calculating of impact point three-dimensional coordinate.In native system, utilize the friendly at C++Builder interface and the terseness of MATLAB programming, the convenience of data operation, bring into play the chief separately, solved the problem of finding the solution of Nonlinear System of Equations.Write the m file by MATLAB and finish finding the solution of equation group, the compiler .mcc that utilizes MATLAB to carry then, the MATLABm file that exploitation is good is compiled as executable program.In C++Builder, directly call.

It below is the system principle diagram of the orthopaedics operation navigation system that provides of this utility model

As shown in figure 11, at first, patient's lesions position is carried out tomoscan image CT or does the MRI image, obtain the medical image of affected part, these image datas are input to computer and according to the needs of actual operation operation tool and tissue on every side thereof are made corresponding demonstration; The affected part area-of-interest that needs when then, the two-dimensional image preliminary treatment being obtained performing the operation afterwards; The image that obtains is carried out showing on computers after three-dimensional reconstruction and some processing, make surgical doctor can rely on computer screen to observe the correctness of operation tool position and carry out preoperative diagnosis and surgical planning; In the operation process, space positioning apparatus is constantly measured patient position in the art and operating theater instruments by the index point on it, through corresponding spatial alternation, the medical image image co-registration of crossing with registration together, thereby the position that can on display, find out operating theater instruments whether with art before plan consistent.The doctor can observe the position of operating theater instruments with respect to affected part from different sides like this, just as directly carry out processing and manufacturing in the machinery on three-view diagram, reaches purpose accurately and rapidly thereby make to perform the operation.

Promptly, the orthopaedics operation navigation system that utilizes this utility model to develop, record several index points of being fixed on the intramedullary pin three-dimensional world coordinate with respect to certain world coordinate system, record simultaneously and be fixed on operation tool---the world coordinates of several index points on the electric drill, both are unified under the intramedullary pin coordinate system, can learn whether residing position of electric drill and direction be correct, thereby instruct operative doctor correctly to punch, also can directly result of calculation be passed to robot motion's control sequence, finish correct punching task automatically by robot.

Adopt the utility model of the technical program, its beneficial effect is:

One, the selection of computer

Because the generation of digital picture and processing will be finished on computers, so the performance of computer will directly have influence on the performance of whole visual system.Along with the development of PC technology, a lot of complicated date processing and Flame Image Process can realize on microcomputer.This utility model has adopted the above processor of 1G, and 256 MB of memory is based on the operating system of Windows2000.Facts have proved that such configuration can satisfy system requirements.

Two, the selection of video camera

Adopt high-resolution colourful CCD video camera will make system's image segmentation, friendly interfaces etc. are more satisfactory, the precision of system also can be higher, but consider that high-resolution needs image processing function more fast, and the present price of high-resolution colourful CCD video camera is very expensive, select suitable black-white CCD video camera also can satisfy the needs of orthopedic surgery navigation substantially, therefore, two high resolution CCD video cameras that the approaching Computar company of performance produces in this utility model, have been selected, the ccd array of ccd video camera has 768 * 576 sensitization elements, the effective focal length f=8mm of camera lens.

Three, the selection of image pick-up card

The Meteor-II/Multi Channal image pick-up card of Canada Matrox company exploitation can connect 2 road rgb signals or 6 road black-and-white video signals, by pci bus view data is real-time transmitted in system or the video memory.The Matrox image pick-up card has abundant program library MIL (Matrox imaging library), and it provides, and a series of functions are used to catch, Treatment Analysis, transmission, demonstration and preservation image.MIL-Lite is the basic controlling storehouse, is the subclass of MIL, and it does not comprise the function of image processing and analyzing part.Utilize MIL-Lite control storehouse that image acquisition is carried out follow-up processing to computer.

Four, double-channel signal lock unit

Utilize the R passage and the G passage of capture card to gather the two-way gray level image of different azimuth, in order to ensure the two-way image is at one time two different azimuth, increased in the system and added lock unit, so that synchronizing signal (composite signal of field synchronization and line synchronising signal) to be provided, make the image synchronization of two paths, thereby guarantee that the image that obtains is the scene image of synchronization different points of view.

Five, the design and fabrication of locator bracket

Locator bracket is designed in experimental stage can multi-direction adjustable frame for movement, and this utility model adopts universal tripod in the starting stage for this reason, not only can regulate the orientation of video camera, and the distance between can two video cameras of arbitrary adjusting.

Six, the parallax range between the ccd video camera

According to theoretical calculation analysis as can be known, between two ccd video cameras distance---be that baseline is big more, the positioning accuracy of system will be high more, but the length of baseline is subjected to the restriction of visual field.Distance between two ccd video cameras increases, and can cause spatial characteristic point in the visual field of a video camera and exceed the visual field of another one video camera.Therefore, the length of baseline will be determined according to actual needs.Through test of many times and analysis, finally we are defined as L=280mm with the length of base.Height and deflection are adjustable.

Seven, the design and fabrication of marker

Marker is mainly used to be fixed on the operating theater instruments in the art---intramedullary pin and operation tool---on the electric drill.It is connected on the testee rigidly, can record the locus of index point on it by the binocular vision spatial locator, thereby determine the locus and the attitude of intramedullary pin and electric drill in the operation, and then determine both relative position, to instruct operative doctor or operating robot to correct the pose of electric drill, guarantee the correctness of punching.

Marker can be integrally-built, also can have several index points to constitute.Consider the convenience of calculating and making, marker is designed to the same version, can adopt the method for adorning luminous tube less, realize the difference of marker according to different needs.If the luminous tube place is mounted to reflection sphere, can realize the passive type marker.

Marker is provided with not, and 4 active illuminating diodes on same straight line can improve locating accuracy.

Eight, systems soft ware

Systems soft ware mainly comprises following a few part: at C++Builder6.0 and MATLAB as development environment (later stage transplants it on VC again): exploitation image capture software bag; The rim detection software kit; The core calculations software kit; The camera calibration software kit; Spatial point and object are rebuild software kit; Intramedullary pin strand lockhole electric drill positioning software bag.

Description of drawings

Fig. 1 is the three-dimensional structure diagram of binocular stereo vision sensing device of the present utility model.

Fig. 2 is connected with the marker user mode structural representation of intramedullary pin and electric drill for this utility model.

Fig. 3 is intramedullary pin of the present utility model and electric drill and the coordinate system figure that goes up marker thereof.

Fig. 4 is intramedullary pin of the present utility model and marker scheme of installation.

Fig. 5 is the projection of intramedullary pin of the present utility model and electric drill X-Y plane in the intramedullary pin coordinate system, and wherein figure line is represented intramedullary pin, and dotted line is represented the electric drill centrage, upwards raises angle 7.4638 degree.

Fig. 6 is a femur strand locking bit surgical navigational indicator diagram of the present utility model, and wherein figure line is represented intramedullary pin, and dotted line is represented electric drill, upwards raises angle 7.4638 degree, the angle that rotates backward 2.0973 degree.

Fig. 7 is image border figure of the present utility model.

Fig. 8 is index point rim detection of the present utility model and stasiofax nomogram.

Fig. 9 is a software architecture diagram of the present utility model.

Figure 10 is a capture program structural representation of the present utility model.

Figure 11 is a system principle diagram of the present utility model.

The specific embodiment

1-2 describes in further detail this utility model below in conjunction with accompanying drawing:

Embodiment 1

Shown in Fig. 1-2, orthopaedics operation navigation system comprises binocular stereo vision sensing device and marker 5, and described marker 5 is 2, and described marker 5 is star, is respectively equipped with 1 active illuminating diode 6 on four angles of described marker 5.Inflexible plugging is connected with intramedullary pin 9 on described one of them marker 5, and the intramedullary pin two ends respectively are provided with two strand lock screw holes 7, and wherein another marker 5 inflexible plugging with the special-purpose electric drill 10 of operation are connected, and described electric drill head is provided with horizontal strand lock screw 8.

Described binocular stereo vision sensing device comprises: dominant frequency is not less than the computer 4 of 1G, the adjustable locator bracket 3 of multi-directionally, be arranged on the localizer 2 on the locator bracket 3, be arranged on two ccd video cameras 1 in the localizer 2, described ccd video camera 1 is high-resolution colourful CCD video camera.The ccd array of described ccd video camera 1 has 768 * 576 sensitization elements, the effective focal length f=8mm of camera lens.The parallax range that is arranged between two ccd video cameras 1 in the localizer 2 is L, L=280mm.Image pick-up card, the double-channel signal lock unit that links to each other with image pick-up card.

The orthopaedics operation navigation system that this utility model is developed, be by recording 4 index points being fixed on the intramedullary pin three-dimensional world coordinate with respect to the world coordinate system at a place, record simultaneously and be fixed on operation tool---the world coordinates of 4 index points on the electric drill, both are unified under the intramedullary pin coordinate system, can learn whether residing position of electric drill and direction be correct, thereby instruct operative doctor correctly to punch.

Below be the test and the test result analysis of orthopaedics operation navigation system of the present utility model

(1) pilot system

The surgical navigational test is carried out in laboratory.At first made human body thigh model, the intramedullary pin that will have marker inserts lower limb model inside, the testing crew electric hand drill relies on the location drawing picture and the angular metric that require electric drill translation and rotation of electric drill shown on the computer screen with respect to intramedullary pin, change the position and the direction of electric drill, check the correct qualification rate of once punching.

(2) test result analysis

As can be seen from the test results, it is not high that system accuracy is used for femur strand lock screw hole location.System still exists certain error, and its analysis of causes is as follows:

The key element that is used to navigate---video camera exists the gray-scale shift phenomenon.Video camera is placed resting state, take actionless giving out light a little, find that the gray value at every turn obtain is all different, under the such gray scale situation of 10-255, general difference is about 1-5.Test of many times found that drift value is uncertain, does not have the obvious variation rule.Be assumed to be normal distribution at random, our given gray scale of adopting image exists positive and negative 3 grey scale change for this reason.For example certain under the situation that does not change the index point position, test adopt the image barycentric coodinates of 8 index points, shown in table 4-1.

Table 4-1 test data (Table6-2test datum)

No.1data points No.2data points changes

497.3636 298.0909 497.3636 298.0909 0 0

529.1111 298.6667 529.1111 298.6667 0 0

561.3846 299.3846 561.3846 299.3846 0 0

545.2857 315.0000 545.2857 315.000000

497.0000 329.7143 496.8333 329.8333 0.1667 -0.1190

529.0000 330.8000 528.9091 330.6364 0.0909 0.1636

560.7500 331.3333 560.7500 331.3333 0 0

512.5000 346.0000 512.5000 346.0000 0 0

496.1250 361.8750 496.1250 361.8750 0 0

528.5000 362.5000 528.5000 362.5000 0 0

560.3333 363.1111 560.3333 363.1111 0 0

From table data as can be seen, the image barycentric coodinates of indivedual points have changed 0.1667 pixel.Because the variation of image coordinate will cause the variation of world coordinates, its result is shown in table 4-2.

The variable quantity (Table6-3Changes of the points coordinates) of table 4-2 world coordinates

No.1Coordinates in world coordinate system No.2Coordinates in world coordinate systemchanges

-16.6836 -56.0833 12.0922 -13.0377 -37.5935 13.3782 -3.6459 -18.4898 -1.2860

48.1110 -56.7542 14.5311 49.8053 -48.4025 13.2768 -1.6943 -8.3517 1.2543

-4.1582 -58.3063 75.9333 -3.8464 -43.8132 78.7575 -0.3118 -14.4931 -2.8242

59.4271 -58.5684 77.9283 60.3395 -73.8141 76.0596 -0.9124 15.2457 1.8687

206.7691 -63.1704 326.5482 208.3830 -55.3572 326.6389 -1.6139 -7.8132 -00907

270.6480 -63.4132 320.5725 269.1723 -78.2911 325.8943 1.4757 14.8779 -5.3218

197.7932 -73.4988 390.4225 198.1854 -695817 390.5489 -0.3922 -3.9171 -0.1264

262.3062 -75.7455 384.4929 259.9728 -87.0787 387.7583 2.3334 11.3332 -3.2654

By table 4-2, we find out that maximum variable quantity reaches 5.32mm.

(3) the inferior pixel edge extraction method of interpolation

In Flame Image Process, often to calculate the center of gravity of certain picture point.In operation guiding system, index point is designed to circular luminous point.Find from top analysis, because maximum error reached 5 millimeters more than when the gray-scale shift of video camera caused spatial point to measure, the method of dealing with problems certainly adopts the zero video camera very little, that pixel is higher that wafts, if but can on algorithm, find some ways, reduce error, a kind of economical and practical method of can yet be regarded as.

In order to find the solution the barycentric coodinates of index point, in general Flame Image Process, adopt the threshold value cutting techniques more, in order to find edge of image comparatively accurately, comparatively edge detection techniques such as Xian Jin Laplace, Canny are employed.But no matter which kind of edge detecting technology all adopts operator convolution smoothing technique, and starting point all is to find suitable edge pixel position, thus its precision the highest also can only be Pixel-level, the highest error of also having only 0.5 pixel.The position of how further to improve marginal point, thus accurately determine the position of picture centre, be a kind of economical and practical method that the binocular space positioning system improves precision.For this reason, this paper has proposed the inferior pixel edge extraction method of Gaussian function interpolation, and in conjunction with circular fitting stasiofax algorithm, make calculating significantly accurately of picture centre, and it is little influenced by the video camera gray noise, the ability significance that just suppresses noise improves, thereby final spatial point positional accuracy measurement is obviously improved.Now be described below:

Adopt image, amplify certain dot image wherein.

The intensity profile of this point is as follows;

30 30 33 39 41 38 33 30 28 28

29 35 47 87 125 105 57 34 30 29

31 44 108 253 255 255 167 54 29 29

32 55 179 255 255 255 255 133 28 24

33 57 183 255 255 255 255 155 35 25

32 45 108 249 255 255 216 80 29 27

30 32 47 97 145 139 79 40 31 30

29 28 31 36 41 40 36 33 32 29

From image itself and its intensity profile we as can be seen, the luminous point distance presents the original circular geometry of luminous point when far away and very little, but have 9 white point quadrates to distribute after will sending out greatly, the right also has the luminous point of two whites, and other pixel intensity is not quite similar.We owing to need avoid the influence of experimental enviroment bright spot, determine that according to grey level histogram edge of image point gray value is 185, thereby obtain the edge image of following Fig. 7 at the beginning.The barycentric coodinates that can try to achieve image are: x ₀=529.4000, y ₀=298.6000, the image pixel coordinate that shows in the reality is (530,299).Consider picture noise, this institute with video camera and image pick-up card gray noise positive and negative about 3.If this epigraph gray scale is 183 gray scale increase by 2～3, so, this point also will become marginal point, and the coordinate of image center of gravity will become this moment: x ₀=524.2500, y ₀=294.1875.The error of visual picture center of gravity will reach 0.5873.

Give noise jamming, it is as follows to obtain twice picture centre coordinate:

${\left[\begin{array}{cc}{x}_{0i}& y_{0\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">i</figure-callout>}}\end{array}\right]}_1=\begin{array}{}\end{array}\left[\begin{array}{cc}497.4115& 298.0364\\ 529.2518& 298.6200\\ 561.3143& 299.3135\\ 545.2715& 314.9385\\ 496.8928& 329.8216\\ 528.8153& 330.7026\\ 560.7030& 331.2759\\ 512.4466& 346.0267\\ 496.2739& 361.7932\\ 528.3968& 362.4868\\ 560.3888& 363.2982\end{array}\right]$

${\left[\begin{array}{cc}{x}_{0i}& y_{0\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">i</figure-callout>}}\end{array}\right]}_2=\left[\begin{array}{cc}497.3636& 298.0909\\ 529.1111& 298.6667\\ 561.3846& 299.3846\\ 545.2857& 315.0000\\ 496.8333& 329.8333\\ 528.9091& 330.6364\\ 560.7500& 331.3333\\ 512.5000& 346.0000\\ 496.1250& 361.8750\\ 528.5000& 362.5000\\ 560.333& 363.1111\end{array}\right]$

The difference of twice gained image center of gravity is:

${\left[\begin{array}{cc}{e}_x& e_y\end{array}\right]}_{1-2}=\left[\begin{array}{cc}0.0479& -0.0545\\ 0.1407& -0.0467\\ -0.0703& -0.0711\\ -0.0142& -0.0615\\ 0.0595& -0.0117\\ -0.0938& 0.0662\\ -0.0470& -0.0574\\ -0.0534& 0.0267\\ 0.1489& -0.0818\\ -0.1032& -0.0132\\ 0.0555& 0.1871\end{array}\right]$

Maximum error 0.1871.

If image is done spline interpolation, and then get 185, can obtain degree of accuracy and be higher than 1 inferior pixel edge as the threshold value of cutting apart.

Spline function is at (x _I-1, x _i) on expression formula be:

$S\left(x\right)=M_{i-1}\frac{{(x_i-x)}^3}{{6h}_i}+M_i\frac{{(x-x_{i-1})}^3}{{6h}_i}+(y_{i-1}-\frac{M_{i-1}{\mathrm{<figure-callout\; id="2"\; label="h\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">h</figure-callout>}}_i^{}}{6})\frac{x_i-x}{h_i}+(y_i-\frac{M_{\mathrm{<figure-callout\; id="2"\; label="i\; h\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">i</figure-callout>}}{h}_i^{}{}_2^{}}{6})\frac{x-x_{i-1}}{h_i}---(4-15)$

By above-mentioned function expression we as can be seen S (x) guaranteed interpolation piecemeal, i.e. the seriality of S (x) has guaranteed S (x) simultaneously again, and " in the seriality of node, this is just feasible to utilize this function to carry out interpolation to ask for the more realistic situation in image border.Utilize boundary condition and catch up with that send out can be in the hope of each parameter in the above-mentioned expression formula, thereby try to achieve S (x)=185 o'clock x value.This moment, the edge coordinate of gained was a sub-pix, because crucial center of gravity calculation in the later stage, so the coordinate of marginal point has higher precision.

In the practical application, calculate, not only take the machine duration, and do not have practical value if entire image is carried out cubic spline interpolation.Therefore, in this research, proposition is write down the position and the gray value of the point around the initial edge of image at the reading images coordinate time, and make marks, only these data are carried out cubic spline interpolation afterwards, ask for the edge with the form of gray level threshold segmentation again, thereby can obtain comparatively ideal marginal point and image barycentric coodinates.

Have at video camera under the situation of gray noise, still with adopt for above-mentioned twice image carry out edge extracting, because in this research, index point all adopts circular luminous point, therefore, this paper proposes again on the basis of spline interpolation sub-pix marginal point to be carried out the least square circular fitting, and its approximating method is as follows:

If the equation of circle is:

x ²-2xx ₀+y ²-2y ₀y-c＝0 (4-16)

Actual point (x _i, y _i) quadratic sum of the error that causes is:

$f=\mathrm{\&Sigma;}{(x_i^2-{2x}_i{x}_0+y_i^2-{2y}_0{y}_i-c)}^2---(4-17)$

According to principle of least square method:

$\frac{\&PartialD;f}{{\&PartialD;x}_0}=0$

$\frac{\&PartialD;f}{{\&PartialD;\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="Y20072010557300251.png"\; state="\{\{state\}\}">y</figure-callout>}}_0}=0$

$\frac{\&PartialD;f}{\&PartialD;c}=0$

:

$x_0=\frac{\mathrm{\&Delta;x}}{\mathrm{\&Delta;}}$

${\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="Y20072010557300251.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=\frac{\mathrm{\&Delta;y}}{\mathrm{\&Delta;}}---(4-19)$

$\mathrm{\&Delta;}=2(\mathrm{\&Sigma;}{x}_i^2-n{\left(\mathrm{\&Sigma;}{x}_i\right)}^2)(\mathrm{\&Sigma;}{y}_i^2-n{\left(\mathrm{\&Sigma;}{y}_i\right)}^2)-2{(\mathrm{\&Sigma;}\left(x_i{y}_i\right)-\mathrm{n\&Sigma;}{x}_i\mathrm{\&Sigma;}{y}_i)}^2$

Wherein: $\mathrm{\&Delta;x}=(\mathrm{n\&Sigma;}{\mathrm{<figure-callout\; id="3"\; label="x\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">x</figure-callout>}}_i^{}+\mathrm{n\&Sigma;}{x}_i{y}_i^2-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{x}_i^2-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{y}_i^2)(\mathrm{n\&Sigma;}{y}_i^2-\mathrm{\&Sigma;}{y}_i\mathrm{\&Sigma;}{y}_i)$

$-(\mathrm{n\&Sigma;}{x}_i^2{y}_i+\mathrm{n\&Sigma;}{y}_i^3-\mathrm{\&Sigma;}{\mathrm{<figure-callout\; id="2"\; label="x\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">x</figure-callout>}}_i^{}\mathrm{\&Sigma;}{y}_i-\mathrm{\&Sigma;}{y}_i\mathrm{\&Sigma;}{y}_i^2)(\mathrm{n\&Sigma;}{x}_i{y}_i-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{y}_i)$

$\mathrm{\&Delta;y}=(\mathrm{n\&Sigma;}{x}_i^2{y}_i+\mathrm{n\&Sigma;}{y}_i^3-\mathrm{\&Sigma;}{\mathrm{<figure-callout\; id="2"\; label="x\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">x</figure-callout>}}_i^{}\mathrm{\&Sigma;}{y}_i-\mathrm{\&Sigma;}{y}_i\mathrm{\&Sigma;}{y}_i^2)(\mathrm{n\&Sigma;}{x}_i^2-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{x}_i)$

$-(\mathrm{n\&Sigma;}{\mathrm{<figure-callout\; id="3"\; label="x\; i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">x</figure-callout>}}_i^{}+\mathrm{n\&Sigma;}{x}_i{y}_i^2-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{x}_i^2-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{y}_i^2)(\mathrm{n\&Sigma;}{x}_i{y}_i-\mathrm{\&Sigma;}{x}_i\mathrm{\&Sigma;}{y}_i)$

The centre coordinate of 11 index points that approximating method is tried to achieve, as shown in Figure 8:

The centre coordinate data that add 11 points of gained under the gray noise situation for twice are as follows:

${\left[\begin{array}{cc}{x}_{0i}& y_{0\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="Y20072010557300221.png,Y20072010557300231.png"\; state="\{\{state\}\}">i</figure-callout>}}\end{array}\right]}_1=\left[\begin{array}{cc}497.4199& 298.0507\\ 529.2384& 298.6254\\ 561.3153& 299.3106\\ 545.2790& 314.9531\\ 496.8647& 329.8412\\ 528.7624& 330.6692\\ 560.7097& 331.2733\\ 512.4397& 346.0258\\ 496.2782& 361.7889\\ 528.3864& 362.4844\\ 560.3690& 363.2831\end{array}\right]$

${\left[\begin{array}{cc}{x}_{0i}& y_{0\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="Y20072010557300221.png"\; state="\{\{state\}\}">i</figure-callout>}}\end{array}\right]}_2=\left[\begin{array}{cc}497.4012& 298.0439\\ 529.2444& 298.6147\\ 561.3162& 299.3117\\ 545.2750& 314.9373\\ 496.9097& 329.8230\\ 528.8169& 330.7067\\ 560.6581& 331.2875\\ 512.4420& 346.0234\\ 496.2761& 361.7937\\ 528.3960& 362.4857\\ 560.3987& 363.3063\end{array}\right]$

The difference of twice gained image center of gravity is:

${\left[\begin{array}{cc}{e}_x& e_y\end{array}\right]}_{1-2}=\left[\begin{array}{cc}0.0187& 0.0068\\ -0.0060& 0.0107\\ -0.0009& -0.0011\\ 0.0040& 0.0158\\ -0.0450& 0.0182\\ -0.0545& -0.0375\\ 0.0516& -0.0142\\ -0.0023& 0.0024\\ 0.0021& -0.0048\\ -0.0096& -0.0013\\ -0.0297& -0.0232\end{array}\right]$

As can be seen, although other gray noise influence of ad eundem all appears in twice image, the barycentric coodinates maximum difference of asking only has 0.0545 pixel after employing spline interpolation edge extracting and the circular fitting.As seen, adopt new algorithm not only to improve the degree of accuracy of image center of gravity, but also drawn up because the gray noise that the camera review capture card causes improves precision more than 3 times.

Embodiment 2

Orthopaedics operation navigation system comprises binocular stereo vision sensing device and marker 5, and described marker 5 is 2, and described marker is triangular in shape, is respectively equipped with 1 active illuminating diode 6 on three angles of described marker.Inflexible plugging is connected with intramedullary pin 9 on described one of them marker 5, and the intramedullary pin two ends respectively are provided with two strand lock screw holes 7, and wherein another marker 5 inflexible plugging with the special-purpose electric drill 10 of operation are connected, and 10 in described electric drill is provided with horizontal strand lock screw 8.

Described binocular stereo vision sensing device comprises: dominant frequency is not less than the computer 4 of 1G, the adjustable locator bracket 3 of multi-directionally, and described locator bracket 3 adopts universal tripod.Be arranged on the localizer 2 on the locator bracket 3, be arranged on two ccd video cameras 1 in the localizer 2, described ccd video camera 1 is high-resolution colourful CCD video camera.The ccd array of described ccd video camera 1 has 768 * 576 sensitization elements, the effective focal length f=8mm of camera lens.The parallax range that is arranged between two ccd video cameras 1 in the localizer 2 is L, L=280mm.Image pick-up card, the double-channel signal lock unit that links to each other with image pick-up card.

The orthopaedics operation navigation system that this utility model is developed, be by recording 3 index points being fixed on the intramedullary pin three-dimensional world coordinate with respect to the world coordinate system at a place, record simultaneously and be fixed on operation tool---the world coordinates of 3 index points on the electric drill, both are unified under the intramedullary pin coordinate system, can learn whether residing position of electric drill and direction be correct, thereby instruct operative doctor correctly to punch.

The above only is preferred embodiment of the present utility model, and all equalizations of being done according to this utility model claim change and modify, and all should belong to the covering scope of this utility model patent.

Claims (7)

1. orthopaedics operation navigation system comprises binocular stereo vision sensing device and marker (5), it is characterized in that,

Described binocular stereo vision sensing device comprises: dominant frequency is not less than the computer (4) of 1G, the adjustable locator bracket of multi-directionally (3), be arranged on the localizer (2) on the locator bracket (3), be arranged on two ccd video cameras (1) in the localizer (2), image pick-up card, the double-channel signal lock unit that links to each other with image pick-up card

Described marker (5) is 2, and described marker (5) is provided with not at least 3 the active illuminating diodes (6) on same straight line,

Described one of them marker (5) is gone up inflexible plugging and is connected with intramedullary pin (9),

Inflexible the plugging with the special-purpose electric drill of operation (10) of described wherein another marker (5) is connected.

2. orthopaedics operation navigation system as claimed in claim 1 is characterized in that, described ccd video camera (1) is high-resolution colourful CCD video camera.

3. orthopaedics operation navigation system as claimed in claim 2 is characterized in that, the ccd array of described ccd video camera (1) has 768 * 576 sensitization elements, the effective focal length f=8mm of camera lens.

4. orthopaedics operation navigation system as claimed in claim 1 is characterized in that, the parallax range that is arranged between two ccd video cameras (1) in the localizer (2) is L, L=280mm.

5. orthopaedics operation navigation system as claimed in claim 1 is characterized in that described marker is triangular in shape, is respectively equipped with 1 active illuminating diode (6) on three angles of described marker.

6. orthopaedics operation navigation system as claimed in claim 1 is characterized in that, described marker (5) is star, is respectively equipped with 1 active illuminating diode (6) on four angles of described marker (5).

7. orthopaedics operation navigation system as claimed in claim 1 is characterized in that, described locator bracket (3) adopts universal tripod.

Images