CN102303314A - Carbon bowl center and guide groove positioning device in industrial production and positioning method thereof - Google Patents

Abstract

The invention provides a carbon bowl center and guide groove positioning device and a positioning method thereof in industrial production, which are applied to carbon block cleaning in industry. The positioning device comprises a serial communication module, an image acquisition module, an image processing module and a camera, the serial communication module is used for communication, after the mechanical arm control system sends an acquisition signal, the image acquisition module controls the camera to acquire carbon bowl images, the image processing module processes the acquired images to obtain data of carbon bowl circular parameters and guide groove angles, and the mechanical arm control system positions the mechanical arm to clean carbon blocks according to the data. The positioning method comprises the steps of carrying out noise filtering and high-pass filtering on collected images, fitting the obtained edge images to obtain circular parameters of the carbon bowl, and carrying out radial projection and filtering on the circle to obtain a guide groove angle. The invention realizes the determination of the center of the carbon bowl and the determination of the angle of the guide groove, provides positioning parameters for the mechanical arm and can improve the working efficiency of carbon block cleaning.

Description

Carbon bowl center in the commercial production and guide groove positioner and localization method thereof

Technical field

The present invention relates to the non-ferrous metal production technical field, be specifically related to the device and the localization method thereof of the carbon bowl center and the guide groove location of aluminium electrolyzing carbon piece in the commercial production.

Background technology

The automatic cleaning of carbon piece can be enhanced productivity in the commercial production, the protection direct labor.The present invention is intended to for mechanical arm locating information is provided.

Present carbon piece cleaning is accomplished by manual work, owing to have a large amount of dusts in the factory building, workers ' health is received very big influence.Mechanical arm is the very important role of performer in industrial production automation, is widely used in the work that automatic assembling, combination and extraction object or some have difficulty and are not suitable for the manpower burden.In commercial production, can carry out the automatic cleaning of carbon piece, can enhance productivity like this, the protection direct labor with mechanical arm.But when carrying out the automatic cleaning of carbon piece, need locating information be provided for mechanical arm.It is a useful improvement that mechanical arm is applied to the cleaning of carbon piece, is the shape of carbon bowl for the particularity of its location.As shown in Figure 3; The carbon bowl of aluminium electrolyzing carbon piece is the circle with six guide grooves; Two-layer edge inside and outside this circle has, and certain deformation is arranged, localizing objects is for confirming the angle of central coordinate of circle and one of them guide groove; Carry out the cleaning of carbon piece to offer mechanical arm, but also be not directed against the positioner and the method for the cleaning of the carbon bowl of anode block in the aluminium electroloysis at present.

Summary of the invention

The objective of the invention is location, propose carbon bowl center and guide groove localization method in a kind of metal smelt production for the carbon bowl that solves anode block in the aluminium electroloysis.

A kind of carbon bowl center and guide groove positioner comprise serial communication modular, image capture module, image processing module and camera.Described serial communication modular receives the IMAQ signal that mechanical arm control system sends, and this signal is sent to image capture module, after described image capture module is received the IMAQ signal; The control camera carries out the collection of carbon bowl image, and every image will gathering in real time sends image processing module to, and described image processing module is handled every the image that receives; Concrete processing procedure is: at first to image filtering hot spot; Adopt Gaussian filter to do noise filtering again and obtain initial pictures, again initial pictures is done the high-pass filtering of 0 degree, 45 degree, 90 degree and 135 degree respectively, get four high-pass filtering results' mean value then; Obtain a breadths edge image; Edge image is comprised as the circular parameter that match obtains the carbon bowl: central coordinate of circle and radius, according to obtaining circle, obtain the projection function of initial pictures; This projection function is carried out filter filtering maximizing point; Obtain guide groove guide groove angle with respect to the center of circle in the image coordinate system at the circular place of carbon bowl, the circular parameter of the carbon bowl that image processing module will obtain from every image and guide groove angle send to mechanical arm control system through serial communication modular.

A kind of carbon bowl center and guide groove localization method specifically may further comprise the steps:

Step 1, mechanical arm control system send the IMAQ signal and give serial communication modular, and serial communication modular will receive the IMAQ signal and send to image capture module.

Step 2, mechanical arm control system control mechanical arm moves, directly over the carbon bowl that mechanical arm is positioned at will take, and image capture module control camera collection carbon bowl image then, and send the carbon bowl image of gathering to image processing module.

Step 3, image processing module are handled the image that receives:

(1) original image being carried out the filtering hot spot handles and to obtain image I: at first with the point in the image upper left corner as the origin of coordinates (O, O), horizontal direction is as the X axle, vertical direction is set up coordinate system as the Y axle, confirms the threshold value T of filtering hot spot then:

$T=\max \{t:\underset{i=t}{\overset{255}{\mathrm{\Σ}}}H\left(i\right)>0.07\}$

H (i) expression gray value is the pixel of an i shared ratio in entire image, i=0, and 1 ... 255, t representes some gray values; Pixel to original image scans by row at last, if certain pixel I (i, value j) then makes the gray value of the value of this pixel for a pixel on it greater than T, if this pixel is the first row pixel, the gray value that then makes this pixel is threshold value T;

(2) adopt Gaussian filter to do noise filtering and obtain initial pictures I ';

(3) respectively initial pictures I ' is done 0 degree, 45 degree, the high-pass filtering of 90 degree and 135 degree is asked on average the filtered of four high-pass filterings then, obtains a breadths edge image I _e

(4) to the edge image I _eDo match, obtain the circular parameter of carbon bowl:

$Q(\mathrm{\β},x,y,r)=\underset{i=0}{\overset{M}{\mathrm{\Σ}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\β}\right),x+(r-i)*\cos \left(\mathrm{\β}\right))$

$-\underset{i=-M}{\overset{0}{\mathrm{\Σ}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\β}\right),x+(r-i)*\cos \left(\mathrm{\β}\right))$

$(x_0,y_0,r_0)=\underset{(x,y,r)}{\arg \max }\underset{\mathrm{\β}}{\mathrm{\Σ}}\mathrm{abs}\left(Q(\mathrm{\β},x,y,r)\right)$

Wherein, Q (β, x, y, r) expression edge image I _eWith (x, y) for center of circle r is on the angle beta direction of circle of radius, inside and outside each M pixel poor; Angle beta is the center of circle (x; Y) with the line of the pixel of being got with cross the center of circle (x, y) and the clockwise angle of the straight line of parallel and X axle, M is a positive integer; Be illustrated in the carbon bowl circle or the number of circular outer pixel of being got, (x ₀, y ₀, r ₀) center of circle (x of the carbon bowl circle that will try to achieve for institute ₀, y ₀) and radius r ₀

(5) on initial pictures I ', inside and outside each Z pixel that the carbon bowl of obtaining is circular is done radially projecting, obtains projection function P (θ):

$P\left(\mathrm{\&theta;}\right)=\underset{i=-Z}{\overset{Z}{\mathrm{\&Sigma;}}}{I}^{\&prime;}({\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="HDA0000061762080000023.png"\; state="\{\{state\}\}">y</figure-callout>}}_0+(r_0-i)*\sin \left(\mathrm{\&theta;}\right),x_0+(r_0-i)*\cos \left(\mathrm{\&theta;}\right))$

The projection function P (θ) that obtains is obtained through filter filtering

To signal

The maximizing point obtains in the image coordinate system at the circular place of carbon bowl guide groove with respect to the guide groove angle θ in the center of circle ₀:

${\mathrm{\&theta;}}_0=\underset{\mathrm{\&theta;}}{\arg \max }\tilde{P}\left(\mathrm{\&theta;}\right)$

Whether step 4, judgement all gather four carbon bowls of carbon piece, if not, change step 2 and carry out, if the circular parameter and the guide groove angle of all carbon bowls that then will obtain send to serial communication modular.

Step 5, serial communication modular send to mechanical arm control system with the circular parameter and the guide groove angle-data of the carbon bowl that receives, and mechanical arm control system carries out the cleaning of carbon piece according to the Data Control mechanical arm of receiving.

Advantage of the present invention and good effect are: the present invention has realized confirming of carbon bowl center and confirming of guide groove angle, for mechanical arm provides positional parameter, can improve the operating efficiency of carbon piece cleaning.

Description of drawings

Fig. 1 is the system architecture sketch map of positioner of the present invention;

Fig. 2 is the block diagram of localization method of the present invention;

Fig. 3 is the exemplary plot of an original carbon bowl image;

Fig. 4 is the waveform sketch map that projection function is carried out filter filtering;

Fig. 5 is the sketch map that positions the circular and guide groove angle of the carbon bowl that obtains in the localization method of the present invention.

The specific embodiment

To combine accompanying drawing the present invention to be done further detailed description below with implementing.

Carbon bowl center of the present invention and guide groove positioner comprise: serial communication modular 2, image capture module 3, image processing module 4 and camera 5, and as shown in Figure 1.

As shown in Figure 1; Mechanical arm control system 1 sends an IMAQ signal through serial communication modular 2 and gives image capture module 3; Image capture module 3 is received back control camera 5 collection carbon bowl images; After image of every collection completion the image of gathering being passed to image processing module 4 handles; Detect carbon bowl center and guide groove position, behind intact four images of acquisition process, image processing module 4 sends signal and returns result for mechanical arm control system 1 through serial communication modular 2.Mechanical arm control system 1 carries out the cleaning of carbon piece according to the Data Control mechanical arm location of receiving.

Camera 5 is fixed on the mechanical arm, and perpendicular to the carbon piece, it confirms that with the coordinate of mechanical arm relation mechanical arm control system 1 is confirmed moving direction through the relative coordinate of the relative mechanical arm of coordinate and carbon bowl center of mechanical arm.Because each carbon piece has four carbon bowls, take so need carry out four times.Before each the shooting, mechanical arm control system 1 can be based on the manufacturing parameter mechanically moving arm of carbon piece, directly over the carbon bowl that mechanical arm is positioned at will take.

Serial communication modular 2 is responsible for positioner of the present invention and electrical system 1 communicates, and can use programming language Visual C++6.0 to realize the function that communicates with electrical system 1.Image capture module 3 adopts MicroVision VGA174 capture card, and tool using Visual C++6.0 writes the control function, realizes obtaining of carbon bowl image.The camera 5 of image capture module 3 controls adopts Watec 902B industrial camera and TAMRON 12VM1040ASIR camera lens.Image processing module 4 tool using Visual C++6.0 write, and confirm the central coordinate of circle and the guide groove angle of carbon bowl, realize carbon bowl center and guide groove location.

4 pairs of images of image processing module are handled; Obtain the circular parameter and the guide groove angle of carbon bowl, concrete processing procedure is: at first the image that obtains is carried out the filtering hot spot and handle, adopt Gaussian filter to do noise filtering again and obtain initial pictures; Respectively initial pictures is done the high-pass filtering of 0 degree, 45 degree, 90 degree and 135 degree again; Get four high-pass filtering results' mean value then, obtain a breadths edge image, edge image is obtained carbon bowl circular central coordinate of circle and radius do match; According to obtaining circle; Obtain the projection function of initial pictures, this projection function is carried out filter filtering maximizing point, obtain guide groove guide groove angle with respect to the center of circle in the image coordinate system at the circular place of carbon bowl.

Carbon bowl center in the commercial production of the present invention and guide groove localization method, as shown in Figure 2, specifically:

Step 1, mechanical arm control system 1 send the IMAQ signal and give serial communication modular 2, and serial communication modular 2 will receive the IMAQ signal and send to image capture module 3.

Step 2, mechanical arm control system 1 control mechanical arm moves, and directly over the carbon bowl that mechanical arm is positioned at will take, image capture module 3 control cameras 5 are gathered carbon bowl images then, and send the carbon bowl image of gathering to image processing module 4.

Step 3,4 pairs of original images that receive of image processing module carry out following processing:

(1) the possible hot spot of filtering.At first with the point in the image upper left corner as the origin of coordinates (O, O), horizontal direction is as the X axle, vertical direction is set up coordinate system as the Y axle, and is as shown in Figure 5.The grey level histogram H (i) of statistical picture then, i=0,1 ... 255, H (i) expression gray value is the pixel of an i shared ratio in entire image, and through type (1) is tried to achieve the threshold value T of filtering hot spot:

$T=\max \{t:\underset{i=t}{\overset{255}{\mathrm{\&Sigma;}}}H\left(i\right)>0.07\}---\left(1\right)$

Wherein, t representes some gray values, and T is the maximum of in the Candidate Set of t, trying to achieve.At last all gray values in the original image are carried out the arest neighbors difference greater than the pixel of T, promptly scan by row, if certain pixel I (i, value j) then makes the gray value of its value for a pixel on it greater than T; If it is the first row pixel, the gray value that then makes it is threshold value T, is expressed as with formula (2):

$I(i,j)=\left\{\begin{array}{cc}I(i-1,j),& i>1\\ T,& i=1\end{array}\right.---\left(2\right)$

(i j) is the coordinate of pixel.

(2) adopt Gaussian filter that image is done noise filtering.Through obtaining image I after (1) processing, (filtering can be expressed as the Using Convolution of I and G: I '=I*G for x, y) filtering, and then I ' obtains initial pictures later on for filtering through 2-d gaussian filters device G.

Wherein, G (x y) is defined as:

$G(x,y)=\frac{1}{\sqrt{2\mathrm{\&pi;}}\mathrm{\&sigma;}}{e}^{-\frac{1}{2}(\frac{x^2}{{\mathrm{\&sigma;}}^2}+\frac{{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="HDA0000061762080000011.png"\; state="\{\{state\}\}">y</figure-callout>}}^2}{{\mathrm{\&sigma;}}^2})}---\left(3\right)$

Wherein parameter σ value is 5.6, is the optimized parameter that on described carbon bowl center and guide groove positioner, obtains.

(3) rim detection.At first respectively initial pictures I ' is done 0 degree, 45 degree, the high-pass filtering of 90 degree and 135 degree, concrete through type (4) carries out high-pass filtering:

$E_{\mathrm{\&alpha;}}(i,j)=\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{I}^{\&prime;}(i+k*\cos \mathrm{\&alpha;},j+k*\sin \mathrm{\&alpha;})-\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{I}^{\&prime;}(i-k*\cos \mathrm{\&alpha;},j-k*\sin \mathrm{\&alpha;})---\left(4\right)$

Wherein, N representes carbon bowl circle is asked the needed scope in edge, is positive integer, and the empirical value through experimentizing and obtain at described carbon bowl center and guide groove positioner in the embodiment of the invention is got 1≤N≤50; α representes angle, and value is 0 degree, 45 degree, 90 degree and 135 degree; E _α(i is that the middle coordinate of initial pictures I ' is for (i, pixel j) is done the filtered of the high-pass filtering of angle [alpha] j).

Get the mean value of four filtered then, obtain a breadths edge image, be designated as I _e:

$I_e=\frac{{\mathrm{<figure-callout\; id="0"\; label="E"\; filenames="HDA0000061762080000023.png"\; state="\{\{state\}\}">E</figure-callout>}}_0+E_{45}+E_{90}+E_{135}}{4}---\left(5\right)$

(4) carbon bowl centralized positioning.To the edge image I _eDo match, obtain the circular parameter of carbon bowl:

$Q(\mathrm{\&beta;},x,y,r)=\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\&beta;}\right),x+(r-i)*\cos \left(\mathrm{\&beta;}\right))$

(6)

$-\underset{i=-M}{\overset{0}{\mathrm{\&Sigma;}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\&beta;}\right),x+(r-i)*\cos \left(\mathrm{\&beta;}\right))$

$({\mathrm{<figure-callout\; id="0"\; label="x"\; filenames="HDA0000061762080000023.png"\; state="\{\{state\}\}">x</figure-callout>}}_0,{\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="HDA0000061762080000023.png"\; state="\{\{state\}\}">y</figure-callout>}}_0,r_0)=\underset{(x,y,r)}{\arg \max }\underset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}\mathrm{abs}\left(Q(\mathrm{\&beta;},x,y,r)\right)---\left(7\right)$

Wherein, M is a positive integer, is illustrated in the carbon bowl circle or the number of circular outer pixel of being got, Q (β, x, y, r) expression I _eWith (x, y) for the center of circle, r are on the angle beta direction of circle of radius, inside and outside each M pixel poor, angle beta be the center of circle (x, y) and the line of the pixel of being got and the clockwise angle of X axle positive direction, 0≤β＜360.Abs representes to ask absolute value, Expression is to all angle summations, (x ₀, y ₀, r ₀) for passing through maximization

The center of circle (the x that the carbon bowl of trying to achieve is circular ₀, y ₀) and radius r ₀

(5) guide groove location: circular inside and outside each Z pixel of getting the carbon bowl that (4) obtain is done radially projecting, and 10≤Z≤50 obtain projection function P (θ):

$P\left(\mathrm{\&theta;}\right)=\underset{i=-Z}{\overset{Z}{\mathrm{\&Sigma;}}}{I}^{\&prime;}({\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="HDA0000061762080000023.png"\; state="\{\{state\}\}">y</figure-callout>}}_0+(r_0-i)*\sin \left(\mathrm{\&theta;}\right),x_0+(r_0-i)*\cos \left(\mathrm{\&theta;}\right))---\left(8\right)$

Wherein, θ is an angle, 0≤θ＜360, and P (θ) is the projection of initial pictures I ' on angle θ direction.The curve that projection function obtains is formed by 1 and-1 with the coefficient that filter filtering as shown in Figure 4 obtains

wave filter shown in Figure 4; On the corresponding angle of guide groove; Filter coefficient is-1; On the corresponding angle of non-guide groove, filter coefficient is 1.To signal

The maximizing point obtains guide groove angle θ with respect to the center of circle in the image coordinate system at the circular place of carbon bowl ₀:

${\mathrm{\&theta;}}_0=\underset{\mathrm{\&theta;}}{\arg \max }\tilde{P}\left(\mathrm{\&theta;}\right)---\left(9\right)$

Also can obtain the coordinate (x of guide groove _d, y _d) be:

x _d＝x ₀+r ₀*cos(θ ₀)

(10)

y _d＝y ₀+r ₀*sin(θ ₀)

Central coordinate of circle that obtains and guide groove angle are as shown in Figure 5.

Whether step 4, judgement all gather four carbon bowls of carbon piece, if not, change step 2 and carry out, if all data that then will obtain comprise that the circular parameter of carbon bowl and guide groove angle send to serial communication modular 2.

Step 5, serial communication modular 2 send to mechanical arm control system 1 with the data that receive, and mechanical arm control system 1 carries out the cleaning of carbon piece according to the Data Control mechanical arm of receiving.

Adopt apparatus and method of the present invention to carry out the location of carbon bowl center and guide groove, the embodiment data that obtain after a carbon bowl image is handled are as shown in table 1 below.

Table 1 embodiment data

Parameter	x 0(pixel)	y 0(pixel)	θ 0(degree)
				Estimated value	937	853	17
Evaluated error	2	-3	1.2

From table 1, can find out the center of circle (x that the carbon bowl that the present invention obtains is circular ₀, y ₀) compare with the circular center of circle of actually determined carbon bowl, abscissa differs 2 pixels, and ordinate differs 3 pixels, guide groove angle θ ₀Differ 1.2 degree, error is very little, thereby can explain that the present invention can realize confirming of carbon bowl center and confirming of guide groove angle, for mechanical arm provides positional parameter.

Claims (8)

1. charcoal bowl center and the guide groove positioner in the commercial production is characterized in that this positioner comprises serial communication modular, image capture module, image processing module and camera; Described serial communication modular receives the IMAQ signal that mechanical arm control system sends, and this signal is sent to image capture module, after described image capture module is received the IMAQ signal; The control camera carries out the collection of carbon bowl image, and every image will gathering in real time sends image processing module to, and described image processing module is handled every the image that receives; Concrete processing procedure is: at first to image filtering hot spot; Adopt Gaussian filter to do noise filtering again and obtain initial pictures, again initial pictures is done the high-pass filtering of 0 degree, 45 degree, 90 degree and 135 degree respectively, get four high-pass filtering results' mean value then; Obtain a breadths edge image; Edge image is comprised as the circular parameter that match obtains the carbon bowl: central coordinate of circle and radius, according to obtaining circle, obtain the projection function of initial pictures; This projection function is carried out filter filtering maximizing point; Obtain guide groove guide groove angle with respect to the center of circle in the image coordinate system at the circular place of carbon bowl, the circular parameter of the carbon bowl that image processing module will obtain from every image and guide groove angle send to mechanical arm control system through serial communication modular.

2. carbon bowl center according to claim 1 and guide groove positioner is characterized in that, described image capture module adopts MicroVision VGA174 capture card to realize.

3. carbon bowl center according to claim 1 and guide groove positioner is characterized in that described camera is fixed on the mechanical arm, before each the shooting, mechanical arm control system control mechanical arm be positioned at the carbon bowl that will take directly over.

4. application rights requires carbon bowl center and the guide groove localization method in a kind of commercial production of 1 described carbon bowl center and guide groove positioner, it is characterized in that these localization method concrete steps are following:

Step 1, mechanical arm control system send the IMAQ signal and give serial communication modular, and serial communication modular will receive the IMAQ signal and send to image capture module;

Step 2, mechanical arm control system control mechanical arm moves, directly over the carbon bowl that mechanical arm is positioned at will take, and image capture module control camera collection carbon bowl image then, and send the carbon bowl image of gathering to image processing module;

Step 3, image processing module are handled the original image that receives:

(1) original image being carried out the filtering hot spot handles and to obtain image I: at first with the point in the image upper left corner as the origin of coordinates (O, O), horizontal direction is as the X axle, vertical direction is set up coordinate system as the Y axle, confirms the threshold value T of filtering hot spot then:

$T=\max \{t:\underset{i=t}{\overset{255}{\mathrm{\&Sigma;}}}H\left(i\right)>0.07\}$

H (i) expression gray value is the pixel of an i shared ratio in entire image, i=0, and 1 ... 255, t representes some gray values; Pixel to original image scans by row at last, if certain pixel I (i, value j) then makes the gray value of the value of this pixel for a pixel on it greater than T, if this pixel is the first row pixel, the gray value that then makes this pixel is threshold value T;

(2) (x y) does noise filtering and obtains initial pictures I '=I*G to adopt Gaussian filter G;

(3) initial pictures I ' is done 0 degree, 45 degree, the high-pass filtering of 90 degree and 135 degree is asked on average the filtered of four above-mentioned high-pass filterings then, obtains a breadths edge image I _e

(4) to the edge image I _eDo match, obtain the circular parameter of carbon bowl:

$Q(\mathrm{\&beta;},x,y,r)=\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\&beta;}\right),x+(r-i)*\cos \left(\mathrm{\&beta;}\right))$

$-\underset{i=-M}{\overset{0}{\mathrm{\&Sigma;}}}{I}_e(y+(r-i)*\sin \left(\mathrm{\&beta;}\right),x+(r-i)*\cos \left(\mathrm{\&beta;}\right))$

$(x_0,y_0,r_0)=\underset{(x,y,r)}{\arg \max }\underset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}\mathrm{abs}\left(Q(\mathrm{\&beta;},x,y,r)\right)$

Wherein, Q (β, x, y, r) expression edge image I _eSo that (x y) is the center of circle, and r is on the angle beta direction of circle of radius; Inside and outside each M pixel poor, angle beta be the center of circle (x, y) and the clockwise angle that is wired to X axle positive direction of the pixel of being got; M is a positive integer, is illustrated in the carbon bowl circle or the number of circular outer pixel of being got, (x ₀, y ₀, r ₀) center of circle (x of the carbon bowl circle that will try to achieve for institute ₀, y ₀) and radius r ₀

(5) on initial pictures I ', inside and outside each Z pixel of the circle of the carbon bowl of obtaining is done radially projecting, obtains projection function P (θ):

$P\left(\mathrm{\&theta;}\right)=\underset{i=-Z}{\overset{Z}{\mathrm{\&Sigma;}}}{I}^{\&prime;}(y_0+(r_0-i)*\sin \left(\mathrm{\&theta;}\right),x_0+(r_0-i)*\cos \left(\mathrm{\&theta;}\right))$

The projection function P (θ) that obtains is obtained through filter filtering

To signal

The maximizing point obtains in the image coordinate system at the circular place of carbon bowl guide groove with respect to the guide groove angle θ in the center of circle ₀:

${\mathrm{\&theta;}}_0=\underset{\mathrm{\&theta;}}{\arg \max }\tilde{P}\left(\mathrm{\&theta;}\right)$

The circular parameter and the guide groove angle of step 4, all carbon bowls that will obtain send to serial communication modular, and judge then whether four carbon bowls of carbon piece are all gathered, if not, change step 2 and carry out, if carry out next step;

Step 5, serial communication modular send to mechanical arm control system with the circular parameter and the guide groove angle of the carbon bowl that receives, and mechanical arm control system carries out the cleaning of carbon piece according to the Data Control mechanical arm of receiving.

5. carbon bowl center according to claim 4 and guide groove localization method is characterized in that, the described Gaussian filter G in (2) in the step 3 (x y) is:

$G(x,y)=\frac{1}{\sqrt{2\mathrm{\&pi;}}\mathrm{\&sigma;}}{e}^{-\frac{1}{2}(\frac{x^2}{{\mathrm{\&sigma;}}^2}+\frac{y^2}{{\mathrm{\&sigma;}}^2})}$

Wherein, parameter σ value is 5.6.

6. carbon bowl center according to claim 4 and guide groove localization method is characterized in that, (3) the described high-pass filtering in the step 3 specifically obtains through following formula:

$E_{\mathrm{\&alpha;}}(i,j)=\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{I}^{\&prime;}(i+k*\cos \mathrm{\&alpha;},j+k*\sin \mathrm{\&alpha;})-\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{I}^{\&prime;}(i-k*\cos \mathrm{\&alpha;},j-k*\sin \mathrm{\&alpha;})$

Wherein, N is a positive integer, and expression is asked the needed scope in edge to carbon bowl circle, and span is: 1≤N≤50; α representes the angle of high-pass filtering, and value is 0 degree, 45 degree, 90 degree and 135 degree; E _α(i is that the middle coordinate of initial pictures I ' is for (i, pixel j) is done the filtered of the high-pass filtering of angle [alpha] j).

7. carbon bowl center according to claim 4 and guide groove localization method; It is characterized in that in (5) described in the step 3 projection function P (θ) is carried out filter filtering, the coefficient of this wave filter is formed by 1 and-1; On the corresponding angle of guide groove; Filter coefficient is-1, and on the corresponding angle of non-guide groove, filter coefficient is 1.

8. carbon bowl center according to claim 4 and guide groove localization method is characterized in that, the span of the Z described in the step 3 is: 10≤Z≤50.

Images