JPS624588A - Centroid and main axial angle detector - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] <Technical Field of the Invention> The present invention is applied to, for example, a visual system of a robot, and obtains an image of an object to be recognized, and detects the positional deviation of the image with respect to a reference position. The present invention relates to a technique for detecting rotational deviation, and particularly relates to a center of gravity and principal axis angle detection device that calculates the center of gravity and principal axis angle of an object image in real time.

<Summary of the Invention> The present invention makes it possible to detect the center of gravity and principal axis angle of an image in real time by determining various moments for each horizontal scanning line of a binary image and calculating the respective cumulative sum values.

<Background of the invention> Generally, when detecting positional or rotational deviations of object images,
First, image 1 of the reference model (Fig. 10 (
1)), the coordinates of the center of gravity G0 of the image 1 on the XY seat ridge (Xao.

Y co ) and principal axis angle θ. seek. Next, an image 2 (shown in FIG. 10 (2)) of the object to be recognized is obtained using a television camera, and the coordinates of its center of gravity G (XG) are similarly obtained.

YG) and the main axis angle θ, the positional deviation amount Δ
X, ΔY, and rotational deviation amount Δθ are calculated using the following equations.

ΔX ” For detection, a method is used that uses the moment of the image to calculate the coordinates of the center of gravity and the principal axis angle.

Figure 11 is a diagram for explaining the theory of detecting the center of gravity and principal axis angle using this method.
Each pixel is made up of 6 bits, and the position of each pixel constituting the object image 3 is defined by XY coordinates.

In the figure, the coordinates of an arbitrary pixel 4 are (Xi.

Yj) (However, L+ J = In L...,
256), and the density weighting function of that pixel 4 is f (Xz
, Yj), the moment MpqC of image 3 is p.

q=0.1.2. .

2 Mzo Met thus the above 0
By executing the calculation of formula ~0, the coordinates of the center of gravity G° of image 3 (
XG, Ya) and principal axis angle θ can be detected. However, each of the above equations is a double integral.

It includes multiplication, division, and inverse trigonometric functions, and if these operations were performed using software, it would take a significant amount of time (on the order of seconds).
In particular, it is completely impossible to complete the calculation within the video signal output period for one screen, and if this was implemented using hardware, there would be problems in terms of cost and space.

<Objective of the Invention> The present invention is intended to solve the above problems all at once, and aims to provide a center of gravity and principal axis angle detection device that can detect the coordinates and principal axis angle of the center of gravity of an object image in real time. do.

<Configuration and Effects of the Invention> In order to achieve the above object, the present invention includes a lifting means for lifting an object and outputting a video signal constituting a grayscale image, and a lifting means for outputting a video signal constituting a grayscale image, and a binarizing process for the video signal to convert the video signal vertically and horizontally. A binary processing means that generates a binary image consisting of a plurality of pixels, and an arithmetic processing means that calculates the center of gravity and principal axis angle of an object based on the binary data constituting the corner pixels and the coordinates of the pixels, calculate the center of gravity and principal axis angle. A detecting device is configured, and the arithmetic processing means includes means for calculating a zeroth moment, a first moment, and a second moment regarding a horizontal coordinate address of a binary image for each angular horizontal scanning line of the video signal; means for calculating first and second moments with respect to the vertical coordinate address of the binary image for each horizontal scanning line of the signal, and correlated moments with respect to the horizontal and vertical coordinate addresses of the binary image for each horizontal scanning line of the video signal; , a means for calculating the cumulative addition value over the vertical scanning period for each of the moments, and a means for calculating the center of gravity and principal axis angle of the binary image based on the results of the cumulative addition.

According to this invention, one
Video Make full use of the signal period without wasting it,
The center of gravity position and inclination of an object image can be detected in real time. Therefore, information on object positions and angles can be obtained at high speed, and high-speed control of robots and the like can be realized.

Furthermore, the coordinates of the center of gravity and the angle of the main axis can be detected using software and a simple hardware configuration, and the invention achieves remarkable effects such as cost reduction and simplification of the circuit configuration.

<Description of Embodiments> FIG. 1 shows an example of the circuit configuration of a gravity center and main axis angle detecting device according to an embodiment of the present invention, and FIG. 2 is a diagram for explaining the theory of this device.

In FIG. 2, each vertical and horizontal square represents a pixel, and each pixel has a horizontal coordinate address X+, Xz, . . . .

X, s, and vertical coordinate address Y,, Y! .

...+YtSk is allocated. Image 5 is obtained by binarizing the video signal obtained by imaging the object.
It is a value image, and for example, the object part is made up of black pixels and the background part is made up of white pixels. In the figure, G indicates the center of gravity of image 5, and its coordinates (XG, YG) and principal axis angle θ are as follows:
It is expressed as a function equation of Σ, as shown in the following equation.

ΣXiΣf (Xi Yj)! J ΣYjΣf(XL, Yj) ZYj −Nj!
j Since image 5 is a binary image, the density weighting function f (Xi, Yj) described above is determined by rOJ as shown in the following equation.
It takes any value of rlJ.

Thus, in the above equations (1) to [phase], ZYj -Nj and ΣY, "NJ are the first and second moments with respect to the vertical coordinate address, Σ·Nj.

In the embodiment shown in FIG. 2, specific examples of each moment are shown on both the left and right sides of the figure.

In this embodiment, for each horizontal scanning line of the video signal, the zero-order moment Nj, the first-order moment ΣX! with respect to the horizontal coordinate address of the binary image 5! and second moment Σ
Find X, 2 in hardware, and calculate the first moment yjN and 2 with respect to the vertical coordinate address in the next horizontal scanning line.
The next moment Y''N, and the correlated moment Y+Xt regarding the horizontal and vertical coordinate addresses are calculated by software, and the cumulative addition values ΣNj, ΣY, NJ over the effective vertical scanning period are calculated for each horizontal scanning line.

are calculated respectively, and in the next vertical blanking period, the calculations of equations 1 to [phase] are executed based on the cumulative addition value, and the coordinates (XG, YG) of the center of gravity G and the principal axis θ are calculated by software.

FIG. 3 shows the video signal VD! in this embodiment. The time chart is shown below. In the figure, the upper half of the diagram shows the video signal VDt during the vertical scanning period, and the lower half of the diagram shows the video signal VD! during the horizontal scanning period. are shown respectively.

In the figure, VD indicates a vertical synchronization signal, HD indicates a horizontal synchronization signal, and one vertical scanning period (16.6 milliseconds) is 20H (however, IH means one horizontal scanning period, which is 63.5 microseconds). It consists of a vertical blanking period and the remaining valid vertical scanning period. During this effective vertical scanning period, 24
It includes two horizontal scanning lines, and each horizontal scanning line includes 256 pixel data.

In the figure, Y+, Yz, Yx, ..., Yz
az is the vertical coordinate address in FIG. 2, and
11XzIX3%""+Xzsb correspond to horizontal coordinate addresses, respectively.

Next, to explain the circuit configuration of FIG. 1, in the figure, a television camera 6 images a stationary or moving object 7, for example, from above, and receives a video signal VD that forms a grayscale image.
Output. The synchronization separation circuit 8 receives the video signal VD! Horizontal synchronization signal HD, vertical synchronization signal VD, clock signal C
K, etc. are separated and outputted to the binarization circuit 9.

The binarization circuit 9 sets a certain threshold level for the video signal VD, thereby converting the video signal VD into black and white binarization to generate a binarized image. The pixel counter 10 is for counting the number of pixels (black pixels in this embodiment) constituting the object portion for each horizontal scanning line of the video signal V D +, and the count value N is 1/ CPU (central process) is connected via O (input output) port 11.
ng Unit) 12.

The horizontal counter 13 is for allocating the above-mentioned horizontal coordinate address, and starts counting the clock signal GK after a time corresponding to, for example, 10 pixels has elapsed since the fall of the horizontal synchronizing signal HD. Adder 14 and buffer 15
is the cumulative addition value ΣX! for the horizontal coordinate address where the black pixel is located. The contents Xi of the horizontal counter 3 and the contents of the buffer 5 are input to the adder 14 and subjected to cumulative addition processing.

The table conversion ROM 16 contains the contents of the horizontal counter 3
Direct, its square value x! Table TI for converting to 2
(shown in Figure 4) is stored, and the content of horizontal counter 3 is X! The address of the table conversion ROM 16 is accessed, and its square value Xi'' is output to the adder 17. Cumulative addition of XL! (11! Find Xt2 and CPU1
The output x2 of the table conversion ROM 16 and the contents of the buffer 718 are input to the adder 17 and subjected to cumulative addition processing.

° I to the count value Nj of the counter 0 and each buffer gold and silver
10 is taken into the CPU 12 via port 11, and the CPU
U12 executes predetermined arithmetic processing based on these data to determine the coordinates (xe, YG) of the center of gravity G of the image and the principal axis angle θ.

In addition, in the figure, F ROM (Program+wable
ReadOnly Mea + ory) 19 stores a series of programs such as positional deviation detection, and RA
M (Randon+Access Me+5or
y) has a work area for storing various data and for executing processing. Further, the communication control unit 21 includes a CPU 12
The calculation results are sent to the higher-level controller.

FIG. 5 shows another embodiment of the present invention, in which the calculator 22 and two multiplexers 23 and 24 are used to perform time-sharing processing.

Note that the other configurations in this embodiment are the same as those in the embodiment shown in FIG. 1, and the corresponding components are given the same reference numerals and their explanations will be omitted here.

However, in the circuit shown in FIG. 1, if the mode is set to learning mode and the reference model is damaged with the television camera 6, the synchronization signal etc. will be separated in the synchronization separation circuit 8, and the
The digitization circuit 9 performs binarization processing on the video signal VD to generate a binary image. This binary image output is sent to the pixel counter 10 and each adder 14.17, and in this pixel counter 10 and each adder 14.17, the above-described predetermined counting or calculation is executed.

Then, for each horizontal scanning period, an interrupt INT is generated to the CPU 12 by the horizontal synchronizing signal HD, and each time,
The contents of the pixel counter 10 and the addition results by each adder 14, 17 are taken into the CPU 12.

FIG. 6 shows the interrupt control operation in the CPU 12.

In the figure, Y indicates the count value of a vertical counter (corresponding to the vertical coordinate address) that the CPU 12 has internally, and this vertical counter is incremented every time the interrupt INT occurs. and the vertical synchronization signal VD
29. Counting starts after the time corresponding to H has elapsed.

Now, assuming that the black pixel counting operation by pixel counter 0 is completed for the Yj-th horizontal scanning line,
First, the CPU 12 reads the content N of the pixel counter 10 in step 1 (indicated by "rsTIJ" in the figure), and then
In the next step 2, the cumulative addition value N (corresponding to ΣNJ) is calculated and stored in the RAM 20. Furthermore, in step 3, the content Yj of the vertical counter in the CPU 12
is read, and in the next step 4, the CPU 12 calculates the cumulative addition value NT, (corresponding to the aforementioned ΣN,y), and stores the calculation result in the RAM 20. Next step 5
Now, in order to convert the content Yj of the vertical counter to its square value Yjz, we will use the conversion table (
(not shown) is referred to to obtain the square value YJ'. In the next step 6, the cumulative addition value NY of Y, "Nj" (the above-mentioned ΣYj"N) is calculated.

) is calculated, and the calculation result is stored in the RAM 20.
is stored in Further, in step 7, the content ΣX of the buffer 15 is taken into the CPU 12, and in the next step 8, it is cumulatively calculated, and in the next step 9, the corresponding ΣX is calculated, and the respective calculation results are is stored in the RAM 20. Furthermore, in step 10, the content copy Xi2 of the buffer 18 is
In the following step 11, the CPU 12 calculates the cumulative addition value N, (corresponding to the εεXi'') and stores the result in the RAM 20. The content Yj is r
242J, that is, whether the processing of steps 1 to 11 for the final 242nd horizontal scanning line has been completed, and if the determination is "No", the next interrupt INT The process waits until the next horizontal scanning line is reached, and the above-mentioned processes are repeatedly executed for the following horizontal scanning lines.

Thus, steps 1~ for the final horizontal scan line.
When the process in step 11 is completed, the determination in step 12 is “YE”.
At step 13.14, the cumulative addition value N is calculated.

Using N T l and N T t, the coordinates of the center of gravity G0 (
xe. .

Yr. χ is calculated and the result is stored in a predetermined area of the RAM 20. In Step 15.16 of Matagi, the cumulative addition value N X + N yy + NX,
Using , the principal axis angle θ. is calculated and the result is stored in RAM2
0 in a predetermined area (step 17). In this case, in this embodiment, first calculate the value of Z by converting the following 0 formula, and then calculate FRO by multiplying this Z by 10 or 1/10.
The principal axis angle θ0 is calculated by accessing the address of the conversion table (FIGS. 8 and 9) stored in M19.

In addition, Fig. 8 shows the principal axis angle θ. FIG. 9 shows the conversion table T2 and the conversion curve of tan θ when the angle is 0 to 22.5", and the conversion table T3 and the conversion curve of cot θ when the principal axis angle θ is 22.5 to 45 degrees. The conversion tables T2 and T3 are used depending on the magnitude of the principal axis angle θ.

Coordinates (Xco, Yco) of the center of gravity G0 and principal axis angle θ. When the calculation of is completed, the RAM2 is finally
Each cumulative addition value N, NT, from 0.

NTz, Nx, Nv, and Nxv are cleared, and the series of processing for the reference pattern ends.

Next, the mode is set to measurement mode, and the object to be recognized is damaged using the television camera 6. In this case as well, the same image processing as in the case of the above reference model is performed, and furthermore, the pixel counter 1
0 and each adder 14, 17 counts or operations are performed.

Then, for each horizontal scan, an interrupt rNT is generated by the horizontal synchronizing signal HD to the CPU 2, and the contents of the pixel counter 10 and the addition results by the adders 14 and 17 are read each time.

FIG. 7 shows the interrupt control operation in the CPU 12 in this case. In the figure, steps 21 to 31 are cumulative addition values N, NT+, Nv, and NTz.

This is the step of calculating Nxv, Nx, which is the sixth step.
All steps 1 to 11 in the figure are similar.

and step 21 for the final horizontal scan line
After the processing in step 31 is completed, the determination in step 32 is “YES”.
Then, in steps 33 to 37, the coordinates of the center of gravity G (
xe, Ye) and the principal axis angle θ are calculated. Thereafter, the CPU 2 uses these data and the data about the reference model obtained in the learning mode to
The calculation of the formula is executed, and the image position shift amount ΔX is calculated.

ΔY and rotational deviation amount Δθ are calculated.

(Steps 38-40). Then, in the next step 41, this calculation result is output from the communication control unit 21 to the higher-level controller, and then in step 42, the cumulative addition values N, NT, are stored in the RAM 20.

NTz, Nx, Nv, and Nxv are cleared, and the entire series of processing ends.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an example of the circuit configuration of a center of gravity and main axis angle detecting device according to an embodiment of the present invention, Fig. 2 is a diagram for explaining the theory of operation of the device of the present invention, and Fig. 3 is a video FIG. 4 is a diagram showing the contents of the conversion table; FIG. 5 is a block diagram showing the circuit configuration of another embodiment; FIG. 6 is a flowchart showing interrupt control operation in learning mode; FIG. The figure is a flowchart showing the interrupt control operation in the measurement mode, FIGS. 8 and 9 are diagrams showing a conversion table and a conversion curve, and FIG. 10 is a diagram for explaining the amount of positional deviation and rotational deviation of the image. FIG. 11 is a diagram for explaining the theory of detecting the center of gravity and principal axis angle. 6...TV camera 9...Binarization circuit 10...Pixel counter 12...
・CPU13...Horizontal counter 14.17
......Adders 15, 18...Buffer G, G, ...... Center of gravity (II t22 doo3
Figure ``Video RF3 Phrasal Noi 4 Chart''lrl
Diagram *r, gutter, l・1zu kiribishi* in f'iz-do*
r, -Hyo-4v-h廿I口 运h-GURU hJ〆芥芥・0h〒T口う+9Figure ★4年y-7"ILHz ward゛Branch Shijima 41 juice, Oshiri procedure correction writing 1 death July 16, 1985 1. Indication of the case Patent application filed on June 28, 1985 (2) Relationship to the case Patent applicant address 10 Hanazono Tsuchido-cho, Ukyo-ku, Kyoto-shi, 616 Name (2)
94) Cubic Electric Co., Ltd. Representative Takao Tateishi and one other person4, the agent has been amended as shown in the attached sheet. Description 1, Title of the invention: Center of gravity and main axis angle detection device 2, Claims ■ Imaging means for capturing an image of an object and outputting a video signal constituting a grayscale image, and a method for binarizing the video signal. It consists of a binary processing means that generates a binary image consisting of a plurality of vertical and horizontal pixels, and an arithmetic processing means that calculates the center of gravity and principal axis angle of the binary image based on the binary data that constitutes each pixel and the coordinates of that pixel. , the arithmetic processing means: means for calculating a zero-order moment, a first-order moment, and a second-order moment regarding a horizontal coordinate address of a binary image for each horizontal scanning line of the video signal; and for each horizontal scanning line of the video signal. means for calculating the first and second moments with respect to the vertical coordinate address of the binary image for each horizontal scan line of the video signal; means for calculating the correlation moment with respect to the horizontal and vertical pressure addresses of the binary image for each horizontal scanning line of the video signal; A centroid and principal axis angle detecting device comprising means for calculating a cumulative addition value over a vertical scanning period for each moment, and means for calculating a centroid and principal axis angle of a binary image based on the cumulative addition result. 3. Detailed Description of the Invention (Technical Field of the Invention) The present invention is applied to, for example, a visual device of a robot, which obtains an image of an object to be recognized, and determines the positional deviation of the image with respect to a reference position. The present invention relates to techniques for detecting rotational deviations and rotational deviations, and particularly relates to a center of gravity and principal axis angle detection device that calculates the center of gravity and principal axis angle of an object image in real time. <Summary of the Invention> The present invention makes it possible to detect the center of gravity and principal axis angle of an image in real time by determining various moments for each horizontal scanning line of a binary image and calculating the respective cumulative sum values. , <Background of the Invention> Generally, when detecting positional deviation or rotational deviation of an object image,
First, image 1 of the reference model (Fig. 10 (
1)), the coordinates of the center of gravity G0 of the image 1 on the XY coordinates (XG..Y,.) and the principal axis angle θ. seek. Next, image 2 (shown in Figure 10 (2)) of the object to be recognized is obtained with a television camera, and the coordinates (X, yc) of its center of gravity G and principal axis angle θ are determined in the same manner, and the amount of positional deviation is calculated. ΔX
, ΔY and rotational deviation amount Δθ are calculated using the following equations. ΔX ” A method is used to calculate the coordinates of the center of gravity and principal axis angle using the moment of the image. Figure 11 is a diagram to explain the theory of detecting the center of gravity and principal axis angle using this method. Yes, there are 25 squares in the vertical and horizontal directions in the figure.
Each pixel is made up of 6 bits, and the position of each pixel constituting the object image 3 is defined by XY coordinates. In the figure, the coordinates of any pixel 4 are (Xi. Y,) (where it j”l+ 2+ . . .
256), the density weighting function of that pixel 4 is f (X
, Yj), the moments Mp, I (where p, q=Q, 1.2...) of image 3 are given by the following equation 0, and the coordinates of the center of gravity G of image 3 (Xa, Ye ) and each principal axis θ are given by the following 0 to 0 equations using the above moment. 2 Mgo Mow 2 NX
NyHowever, M'06. is the area of the pattern to be measured. X, , Yc are the X coordinates of the center of gravity of the pattern to be measured. This is the Y coordinate. In this way, by executing the calculation of the above equations 0 to 0, the coordinates (Xc Ye ) of the center of gravity G of the image 3 and the principal axis angle θ can be detected. However, each of the above equations is a double integral. It includes multiplication, division, and inverse trigonometric functions, and if these operations were performed using software, it would take a significant amount of time (on the order of seconds).
In particular, it is completely impossible to complete the calculation within the video signal output period for one screen, and if this was implemented using hardware, there would be problems in terms of cost and space. <Objective of the Invention> The present invention is intended to solve the above problems all at once, and aims to provide a center of gravity and principal axis angle detection device that can detect the coordinates and principal axis angle of the center of gravity of an object image in real time. do. <Structure and Effects of the Invention> In order to achieve the above object, the present invention includes an imaging means for capturing an image of an object and outputting a video signal constituting a grayscale image, and a binarization process for this video signal to generate a plurality of horizontal and vertical images. A center of gravity and principal axis angle detection device includes a binary processing means that generates a binary image consisting of pixels, and an arithmetic processing means that calculates the center of gravity and principal axis angle of an object based on the binary data constituting the corner pixels and the coordinates of the pixels. The arithmetic processing means includes means for calculating a zero-order moment, a first-order moment, and a second-order moment regarding a horizontal coordinate address of a binary image for each angular horizontal scanning line of the video signal; Means for calculating a first moment and a second moment with respect to a vertical coordinate address of a binary image for each horizontal scanning line; calculating a correlation moment with respect to a horizontal and vertical coordinate address of a binary image for each horizontal scanning line of the video signal; The image forming apparatus is provided with means for calculating the cumulative addition value over the vertical scanning period for each of the moments, and means for calculating the center of gravity and principal axis angle of the binary image based on the results of the cumulative addition. According to this invention, one
Video Make full use of the signal period without wasting it,
The center of gravity position and inclination of an object image can be detected in real time. Therefore, information on object positions and angles can be obtained at high speed, and high-speed control of robots and the like can be realized. Furthermore, the coordinates of the center of gravity and the angle of the main axis can be detected using software and a simple hardware configuration, and the invention achieves remarkable effects such as cost reduction and simplification of the circuit configuration. <Description of Embodiments> FIG. 1 shows an example of the circuit configuration of a gravity center and main axis angle detecting device according to an embodiment of the present invention, and FIG. 2 is a diagram for explaining the theory of this device. In FIG. 2, each vertical and horizontal square represents a pixel, and each pixel has a horizontal coordinate address X,,X,... XtS& and vertical coordinate addresses Yl, Y,. ...+yzsi is assigned. Image 5 is obtained by binarizing the video signal obtained by imaging the object.
In figure 0, which is a value image, for example, the object part is composed of black pixels and the background part is white pixels, G indicates the center of gravity of image 5, and its coordinates (XG Y@) and principal axis angle θ are ■〜
[Phase] As shown in the equation, it is expressed as a function equation of Σ. ΣXi ball f (X direct, Yj) Since the image 5 is a binary image, the density weighting function f (xi, yj) described above is expressed as r, 0 as shown in the following equation.
Takes any value of JrlJ. (In the above ■ ~ [phase] formula, ΣY, -N. and ΣYj'' In the embodiment shown in Figure 2, examples of the moments are shown on both the left and right sides of the figure.In this embodiment, a binary image 5 is generated for each horizontal scan line of the video signal. The 0th moment NJ + 1st moment ΣX and the 2nd moment ΣX! with respect to the horizontal coordinate address are determined by hardware, and the 1st moment YjN and the 2nd moment with respect to the vertical coordinate address are calculated in the next horizontal scanning line. Y”N, , and the correlation moment Y, Σ regarding the horizontal and vertical coordinate addresses
At the same time as calculating Xi using software, the effective vertical scanning period is determined for each horizontal scanning line, and in the next vertical blanking period, the calculations of formulas 1 to 2 are executed based on the cumulative addition value. , the coordinates of the center of gravity G (
xGyye) and the principal axis θ are calculated using software. FIG. 3 shows a time chart of the video signal VD in this embodiment. In the figure, the upper half diagram shows the video signal VD during the vertical scanning period, and the lower half diagram shows the video signal VD during the horizontal scanning period. A video signal VD is shown, respectively. In the figure, VD is a vertical synchronization signal, ) (D is a horizontal synchronization signal, and one vertical scanning period (16.6 milliseconds) is 20H (however, IH means one horizontal scanning period, which is 63.5 microseconds). ) and the remaining effective vertical scanning period.This effective vertical scanning period includes 2 vertical blanking periods.
It includes 42 horizontal scanning lines, and each horizontal scanning line includes 2S6 pixel data. In the figure, Yr, Yr, Yko, ...* Yz*x
is the vertical coordinate address in FIG. 2, and X 1
1 x 2. X! +..., X□Todoroki correspond to horizontal coordinate addresses, respectively. Next, to explain the circuit configuration of FIG. 1, in the figure, a television camera 6 images a stationary or moving object 7, for example, from above, and receives a video signal VD that forms a grayscale image.
Output. The synchronization separation circuit 8 receives the video signal VD! Horizontal synchronization signal HD, vertical synchronization signal VD, clock signal G
K, etc. are separated and outputted to the binarization circuit 9. The binarization circuit 9 converts the video signal VDi by setting a certain threshold level to the video signal VDi.
D, is black-and-white disulfurized to generate a binarized image. The pixel counter lO is for counting the number of pixels (black pixels in this embodiment) constituting the object portion for each horizontal scanning line of the video signal VD, and its count value N is I/
C via O (Input Output) boat 11
P U (Central Processing Un
it) is designed to be incorporated into 12. The horizontal counter 13 is for allocating the above-mentioned horizontal coordinate address, and starts counting the clock signal GK after a time corresponding to, for example, 10 pixels has elapsed since the fall of the horizontal synchronizing signal HD. Adder 4 and Buffer 5
is for calculating the cumulative addition value 4x for the horizontal coordinate address where a black pixel is located and outputting it to the CPU 12, and the content of the horizontal counter 13 X! and buffer 15
The contents are input to the adder 14 and subjected to cumulative addition processing. The table conversion ROM 16 stores the contents of the horizontal counter 13
Table TI for converting J to its square value X(''
(shown in FIG. 4) is stored, and the content of the horizontal counter 13 is X! The address of the table conversion ROM 16 is accessed, and its square value X,2 is output to the adder 17. The adder 17 and the buffer 18 are for calculating the cumulative addition value ΣX,z5 of the square value X5z for the horizontal coordinate address where the black pixel is located, and outputting it to the CPU 12. The output of X! 2 and the contents of the buffer 18 are input to the adder 17 and subjected to cumulative addition processing. The count value N of the counter 0 above, and I1 for each buffer gold and silver
0 port 11 to the CPU 12, and the CPU
12 executes predetermined arithmetic processing based on these data to determine the coordinates (XG, Yfi) of the center of gravity G of the image and the principal axis angle θ. In the figure, F ROM (Pr6grammable
ReadOnly)'Ie+wory) 19 stores a series of programs such as positional deviation detection, and R
A M (Randos Access Me+5or
y) has a work area for storing various data and for executing processing. Further, the communication control unit 21 includes a CPU 12
The calculation results are sent to the higher-level controller. FIG. 5 shows another embodiment of the present invention, in which the calculator 22 and two multiplexers 23 and 24 are used to perform time-sharing processing. Note that the other configurations in this embodiment are the same as those in the embodiment shown in FIG. 1, and the corresponding components are given the same reference numerals and their explanations will be omitted here. However, in the circuit shown in FIG. 1, if the mode is set to learning mode and the reference model is damaged with the television camera 6, the synchronization signal etc. will be separated in the synchronization separation circuit 8, and the
The digitization circuit 9 performs binarization processing on the video signal VD to generate a binary image. This binary image output is sent to the pixel counter 10 and each adder 14.17, and in this pixel counter 10 and each adder 14.17, the above-described predetermined counting or calculation is executed. Then, for each horizontal scanning period, an interrupt INT is generated to the CPU 12 by the horizontal synchronizing signal HD, and each time,
The contents of the pixel counter 10 and the addition results by each adder 14, 17 are taken into the CPU 12. FIG. 6 shows the interrupt control operation in the CPU 12. In the figure, Yl indicates the count value of a vertical counter (corresponding to the vertical coordinate address) that the CPU 12 has internally, and this vertical counter is incremented every time the interrupt INT occurs. Yes, vertical synchronization signal VD
Counting starts after a time corresponding to 20H has elapsed since the fall of . Now, assuming that the pixel counter O has completed the black pixel counting operation for the Yj-th horizontal scanning line,
First, the CPU 12 reads the content N of the pixel counter 0 in step 1 (indicated by rsTIJ in the figure), and then calculates the cumulative addition value N (corresponding to the aforementioned ΣN) in the next step 2, and stores it in the RAM 20. Store it in Furthermore, in step 3, the content Y of the vertical counter in CPUI Z,
is read, and in the next step 4, the CPU 12 calculates the cumulative addition value NT, (corresponding to the above-mentioned ΣN, Y), and stores the calculation result in the RAM 20. Next step 5
Now, in order to convert the content Y of the vertical counter into its square value Yj'', we will use the conversion table (
(not shown) is referred to to obtain the square value Yjt. Then, in the next step 6, the cumulative addition value NY of Yj''N (corresponding to the aforementioned ΣYj''Nj) is calculated,
The calculation result is stored in the RAM 20. Furthermore, in step 7, the content ΣXi of the buffer 15 is loaded into the CPU 12, and in the next step 8, the cumulative addition value N
Tt (corresponding to the 8 houses Xi) is calculated, and in the next step 9, Tt (corresponding to 8 houses Xi) is calculated, and the respective calculation results are stored in the RAM 20. Furthermore, in step lO, the content Σx of the buffer 18
i2 is taken into the CPU 12, and in the following step 11,
The CPU 12 calculates the cumulative addition value N, (corresponding to the aforementioned ΣΣX, z) and stores the result in the RAM 20. In the next step 12, it is determined whether the content Y of the vertical counter has reached r242J, that is, whether the processing of steps 1 to 11 for the final 242nd horizontal scanning line has been completed. If the judgment is “No”
At this time, the process waits for the next interrupt INT, and the above-mentioned processes for the following horizontal scanning lines are repeatedly executed. Thus, steps 1~ for the final horizontal scan line.
When the process in step 11 is completed, the determination in step 12 is “YE”.
In the next step 13.14, the coordinates of the center of gravity G0 (XG...Yr,o) are calculated using the cumulative addition values N.NT, NTt, and the results are stored in a predetermined area of the RAM 20 The next step 15.1
6.17, the cumulative addition values Nx, Nv, N
xv, respectively N, =Nv NYco”・N
xv= Nxv N XGOYG@・N
Calculate X N X co to find the principal axis angle θ. is calculated, and the result is stored in a predetermined area of the RAM 20 (
Step 20). In this case, in this embodiment, first convert the following 0 formula to find the value of Z, and then multiply this Z by 10 or 1.
Access the address of the conversion table (Figures 8 and 9) stored in PROM 19 with the value multiplied by /10,
Principal axis angle θ. is being calculated. N yt −N y Fig. 8 shows the conversion table T2 and the conversion curve of tan θ when the main axis angle θ0 is 0 to 22.5”, and Fig. 9 shows the conversion curve of tan θ when the main axis angle θ is 22.5 to 45”. '', the conversion table T3 and cotθ. , respectively, and this conversion table T2. T3 is the principal axis angle θ. It is used depending on the size of the Coordinates of the center of gravity G0 (XGO, Ye.) and principal axis angle θ
. When the calculation is completed, the RAM 20 is finally
Each cumulative addition value N, NT, : NTz , NX , Nv
, Nxv are cleared, and the series of processing for the reference pattern ends. Next, after setting the mode to measurement mode, the television camera 6 images the object to be recognized. In this case as well, the same image processing as in the case of the above reference model is performed, and furthermore, the pixel counter 1
0 and each adder 14, 17 counts or operations are performed. Then, for each horizontal scan, an interrupt INT is generated by the horizontal synchronization signal HD to the CPU 12, and the pixel counter 1
The contents of 0 and the addition results by each adder 14, 17 are read each time. FIG. 7 shows the interrupt control operation in the CPU 12 in this case. In the figure, steps 21 to 31 are cumulative addition values N, NT+, Nv, and NTt. a step of calculating N,v,N,, which is the sixth
This is exactly the same as steps 1 to 11 in the figure. and step 21 for the final horizontal scan line
When the process in step 31 is completed, the determination in step 32 becomes ``YES'', and in steps 33 to 40, the ° coordinates of the center of gravity G (
XG, Ye) and the principal axis angle θ are calculated. Thereafter, the CPU 12 uses these data and the data regarding the reference model obtained in the learning mode to execute the calculation of the formula 0 to 0, and calculates the positional deviation amount ΔX of the image. ΔY and the amount of rotational deviation are calculated. (Steps 41-43). Then, in the next step 44, this calculation result is output from the communication control unit 21 to the higher-level controller, and then in step 45, the cumulative addition values N, NT, are stored in the RAM 20. NTz, Nx, Nv, and Nov are cleared, and the entire series of processing ends. 4. Brief Description of the Drawings Fig. 1 is a block diagram showing an example of the circuit configuration of a center of gravity and main axis angle detecting device according to an embodiment of the present invention, and Fig. 2 is a block diagram for explaining the operating theory of the device of the present invention. Figure 3 is a time chart of the video signal, Figure 4 is a diagram showing the contents of the conversion table, Figure 5 is a block diagram showing the circuit configuration of another embodiment, and Figure 6 is interrupt control in learning mode. Flowchart showing the operation, Figure 7 is a flowchart showing the interrupt control operation in measurement mode, Figures 8 and 9 are diagrams showing the conversion table and conversion curve, and Figure 10 is the amount of positional deviation and rotational deviation of the image. FIG. 11 is a diagram for explaining the theory of detecting the center of gravity and principal axis angle. 6...TV camera 9...Binarization circuit 10...Pixel counter 12...
・CPU13...Horizontal counter 14.17
... Adders 15, 18 ... Buffer G, G, ... Center of gravity

Claims (1)

[Claims] (1) an imaging means for capturing an image of an object and outputting a video signal constituting a grayscale image; a binary processing means for binarizing the video signal to generate a binary image consisting of a plurality of vertical and horizontal pixels; It consists of an arithmetic processing means for calculating the center of gravity and principal axis angle of the binary image based on the binary data constituting each pixel and the coordinates of the pixel, and the arithmetic processing means is configured to perform two steps for each horizontal scanning line of the video signal. means for calculating a zero-order moment, a first-order moment, and a second-order moment with respect to a horizontal coordinate address of a value image; means for calculating a first-order moment and a second-order moment with respect to a vertical coordinate address of a binary image for each horizontal scanning line of the video signal; means for calculating, for each horizontal scanning line of the video signal, a correlation moment regarding the horizontal and vertical addresses of the binary image; means for calculating a cumulative sum over a vertical scanning period for each of the moments; A center of gravity and principal axis angle detection device comprising means for calculating the center of gravity and principal axis angle of a binary image based on the cumulative addition result.