JPH0261777A - Graphic recognizing device - Google Patents

Abstract

PURPOSE:To exactly execute the recognition of a circular graphic even when the density is changed in a circular graphic part or contrast with a background is small by inputting an input picture, whose outline is extracted, to an arithmetic means. CONSTITUTION:At first, a two-dimensional density value A (x,y,0) of an original picture is written to a frame memory A and 0 is written to a frame memory V. Laplacian DELTAA of a picture to be held to the frame memory A is obtained and a result, for which this Laplacian and the contents of the frame memory V are added, is written to the frame memory V again. Next, 1/8 is multiplied to the contents of the frame memory V and a result, for which such a multiplied value is added with the contents of the frame memory A, is written to the frame memory A again. The writing is executed to the frame memory A to be >= a prescribed threshold S. Then, the values of addresses (x), (y) and (n) of the frame memory A to be >= the prescribed threshold S are detected. When a counter (n) is >=R/2X2<1/2> picture elements (R is the maximum radius of a circle to be found), processing is finished.

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] This invention relates to a figure recognition device that recognizes circular figures using a two-dimensional wave equation.

The purpose of the present invention is to accurately recognize 9 circular figures even when the shading changes in the circular figure part or when the contrast with the background is small.

An input image of a circular object is input as a two-dimensional gray value to a calculation means, and a two-dimensional wave equation with the gray value as an initial value is solved, and the center of the circular object is determined by the extreme value of the amplitude in the solution of this wave equation. A figure recognition device configured to recognize a position is provided with contour extraction means for extracting the contour of an input image, and the input image whose contour has been extracted by the contour extraction means is input to a calculation means. .

[Industrial Application Field] The present invention relates to a figure recognition device that recognizes circular figures using a two-dimensional wave equation.

Circular shapes occupy a large proportion of the visual recognition objects in the mechanical industry. Most of the work done by lathes and milling machines is to form circles, and bolts and rivets are fastened to round holes. Furthermore, the wiring terminal portions of printed boards widely used for mounting electronic components are often ring-shaped. Since it is formed using printed wiring, the position of each terminal part is known, but it is inevitable that some misalignment will occur, so during automatic bonding, the center position of the terminal part is determined and accurate bonding is performed. It is desirable to do so.

In addition, in production facilities that have introduced robots, the robots pick up parts made and transported in upstream processes one by one and place them on processing machines, or move them to conveyors in downstream processes. However, in this case, the parts position,
It is necessary to accurately recognize the shape, and for this purpose, it is necessary to be able to accurately measure the center position coordinates and radius of a circle.

[Conventional technology] Various methods for recognizing circular objects have been proposed.

Most conventional methods calculate the center of gravity and circumference by extracting the contour, so when the contour line is interrupted, it may not be possible to trace it as a closed loop and the if center and circumference may not be calculated. If this discontinuity is minute, interpolation technology is known to fill in the gap, but if a large part is missing, interpolation becomes difficult.
Often elegant. This drawback occurs when processing grayscale images using reflected light.

This is particularly problematic.

Therefore, it has been pointed out that --like processing and the macro f-method are effective as recognition methods that are not affected by background, surface patterns, and noise. One example is a method in which skeletal lines are extracted by generating waves from around an image.

Using a similar idea, we compare the circumference of a circular object to waves on the water surface.
Japanese Patent Application Laid-Open No. 58-51387 discloses a figure recognition method for determining the center position by utilizing the fact that waves propagated from the outer periphery gather at the center after a certain period of time and form a peak.

This figure recognition method uses a wave equation to recognize two circular objects. Recognition of a circular drawing using the wave equation is performed as follows. For example, when a 9-round basin is filled with water and an impact is applied to the edge, waves generated from the surroundings propagate towards the end, and after a certain period of time, a peak occurs in the center. Therefore, the area around the 9-circular object is considered to be the area surrounded by the circle on the water surface.

If we simulate the movement of waves inside a calculator, we should see a similar peak at the center. If the position of the peak is known, the center position of the circle can be found, and the radius can be determined from the time it takes to reach the peak.

The strength of this recognition method is that it allows parts of one yen to be missing or multiple circles to overlap each other. The reason why a peak occurs is that the waves that occur at the same time on the circumference -E all strengthen each other in the same phase at the center point, but even if part of the two circles is missing, the remaining arc has sufficient length. If so, there is no difference that the waves generated there will be in the same phase and strengthen each other. Regarding surface patterns and noise, unless they are circular, the phase of the waves generated will be random and will not produce a large peak, and will be independent of the waves originally generated from the circumference and will not interfere with them.

[Problems to be Solved by the Invention] For example, as shown in FIG. 8, when a circular figure to be recognized has different shades of black and white, the circular figure can be recognized using the conventional method using the wave equation. When a recognition device attempts to recognize the waves, the waves of the white part and the black part weaken each other at the center position of the circle, and the amplitude peak no longer occurs at the center of the 9th circle.

Furthermore, if the image of a circular object does not have sufficient contrast with the background, the peak value that occurs at the center position of the two circles will be small, and therefore the center of the circle may not be accurately recognized.

Therefore, an object of the present invention is to accurately recognize a 9-circular figure even when the shading of the 2-circular figure part is different or the contrast with the background is small.

[Means to solve the problem] FIG. 1 is a principle block diagram according to the present invention.

The figure recognition device according to the present invention inputs an input image of a nine-circular object to the calculation means 11 as a two-dimensional gray value, solves a two-dimensional wave equation with the gray value as an initial value, and solves a two-dimensional wave equation using the gray value as an initial value. A figure recognition device configured to recognize the center position of the circular object based on the extreme value of the amplitude,
It is provided with a contour extraction means 12 for extracting the contour of an input image, and is configured such that the input image whose contour has been extracted by the contour extraction means 12 is input to the calculation means 11.

[Operation] From an input image of a circular object, the contour of the circular object is extracted and emphasized by the contour extraction means 12, and then the image is input to the calculation means II, where a circular figure recognition process is performed. As a result, circular figures can be accurately recognized even when there is a change in the shade of a circular object or when there is a small contrast with the background.

[Example] The present invention will be described in detail below with reference to one drawing. Second
The figure is a block diagram showing a schematic configuration of an image processing processor that implements a figure recognition device as an embodiment of the present invention.

In the figure, In to In are frame memories. Each frame memory is connected to an 8-bit video bus 3 via a bidirectional bus buffer 2. 4 is a spatial filter, which includes a line buffer 41, a product-sum/Laplacian processing unit 42 with 3×3 coefficients, and an absolute value/scaling circuit 4.
Contains 3. 5 is an adder circuit, and this adder circuit 5
and the spatial filter 4 are interconnected via an internal bus as shown by the broken line. 6 is A/D conversion and I)
This is a conversion circuit that performs /Δ conversion.

The operation of the example device will be described below.

Conventional figure recognition devices use the grayscale values of the input image themselves as initial values when solving the wave equation. On the other hand 1
In the embodiment device according to the present invention, an extracted and emphasized 0 contour is used instead of the input image itself as the initial value of the wave equation.

As a method for emphasizing the outline, for example, the well-known 5obel
Arithmetic operations are available. This is typically provided in standard image processors, and allows for faster speeds (
For example, you can perform Enkyo in 33m5). In the case of the device of this embodiment as well, this function can be easily realized.

In other words, the 5obel operation can be schematically expressed as shown in FIG. To realize this with the device of this embodiment, 5. For example, first, the original image in the frame memory 1o.

instead of.

The absolute value of the obtained value is sent to the frame memory 11.

Next, the original picture in 111 and frame memory IQ.

The absolute value of the obtained value is written into the frame memory 12.

Finally, the contents of the frame memories 11 and 12 are added through the adder 5 and then written into the frame memory 10.

As a result, as shown in the density sectional view of FIG. 3B, the frame memory 1°, which originally contained nine input images as they were, is replaced with data of only the contour portions with density changes.

The contour data of the circular object obtained in this way is processed using a wave equation according to the following procedure, thereby determining the center and radius of the circular object.

First, the brightness of a grayscale image is regarded as a three-variable function of horizontal and vertical coordinates x, y, and time. This can be expressed as A (x, y,
t, ), and the two-dimensional wave equation % expression % is solved using the 1-frame image A (x, y, L.) input at time 3 To as an initial value. The boundary condition is around L(
square) and always assume △(x, y, shi) = 0. Physically, this is the same as the equation that describes the state of the elastic membrane identified around the 1E rectangle.

In order to convert equation (1) into a difference equation and solve it sequentially, new variables are introduced and a first-order simultaneous equation is created.

That is, when formula (1) is recalculated using V (x, y, u), it becomes as follows.

・　・　・(2) When formula (2) is replaced with a difference approximation formula, it becomes as follows.

V (x, y, t, + δ / 2) = V
(x, y, shi−δ shi/2)+a2 δL△Δ
(x, y, L) Δ (x, y, L + δ) Δ
(X, , Y, + δ and V (x, y, + δ and /2) The algorithm for solving equation (3) is given below.

Step 1: V=o, Δ=Δ(0) Step 2: V←V+a2δ and ΔΔ Step 3: Δ-A+δ and V Step 4 Return to step 2 In this embodiment, in the above algorithm, a2δ and = l, The calculation is performed with δt=1/8, and the procedure at that time is shown in FIG. The processing in this embodiment uses at least two frame memories. That is, frame memory A and frame memory V,
These frame memories physically correspond to the frame memories la, 11 in FIG. The frame memory Δ is made to hold the two-dimensional grayscale value of the image after a predetermined time, and the frame memory ■ is made to hold the time differential value of the two-dimensional grayscale value.

In step S1, the two-dimensional grayscale value A (x.

y, 0) is written into the frame memory 8, 0 is written into the frame memory V, and a counter n indicating the number of times of approximation is set to 1.

In step S2, the Laplacian ΔA of the image held in the frame memory A is determined, and this and the frame memory V
The result of adding the contents of is written into the frame memory V again.

Next, in step S3, the content of the frame number V is multiplied by a predetermined number 9 less than 1, for example 1/8.

The result of adding this value to the contents of frame memory △ is 11
and writes itF into frame memory 8.

In step S4, the values of addresses x, y, and n in the frame memory that are greater than or equal to a predetermined threshold value S are detected, and these values are calculated into a peak table as shown in FIG.

In step S5, the value of the counter n is incremented by one, and in step S6, it is determined whether or not the number of pixels has exceeded 172f pixels in n months. Here, 17 is the maximum "-11 diameter of the circle to be seen, and the wave advances by 27/T pixels in l approximation cycles. If the determination in step S6 is negative, steps 82 to S5 are repeated. However, in the case of a hexagonal determination, the process is terminated.

Through this processing, the center position of the circular object is determined from the x and y addresses in the frame memory with f = less than or equal to a predetermined threshold value S, and the radius of the circular object is determined from the number of repetitions.

In the flowchart of FIG. 4, step S2 and step S
3 means that calculations are performed for all pixels in each frame memory, and the time required for each step is an I frame period (for example, about 33 m5). Step S2
Assuming that the repetition of step S3 is referred to as one approximation cycle, the frame memory A holds A (x, y, t, +nΔt) at the time when the n-th approximation cycle has progressed.

In step 82 and step S3, since it is necessary to perform calculations on the entire contents of the frame memory within one frame period, a high-speed image processor is generally required to execute this processing routine. The pictable in step S4 is as shown in FIG. 5, and the coordinates x, y of the center of the circle are obtained from this peak table.
and its radius can be easily determined.

1- The calculation of the wave equation in the embodiment described above can also be implemented by configuring as follows.

That is, some types of image processors may not be able to execute the processing shown in FIG. 4 from the viewpoint of processing speed. for example,
In step S2, the Laplacian ΔA is added to the data read from the frame memory ■ and written into the same frame memory ■, but it is generally difficult to do this within one frame period. This is because it is necessary to switch between reading and writing of the memory element within the sampling period. each[
! If the frame memory is always in the read state or the write state within one frame period, it will be easier to control and the circuit will be simpler.

Furthermore, depending on the processor, it may not be possible to apply the Laplacian to one frame memory and add it to another frame memory as in step S2 of FIG. 4. In other words, 1) 1 pure f1.・-It may only have the function of sum addition5. In order to execute the above-described processing using a processor with only pure functions, Q' can be implemented by increasing the number of working memories and implementing the processing procedure shown in FIG.

That is, in this modification, in addition to frame memory A and frame memory V, two working memories w+ and W
Add 2. In step S11, the absolute value of the two-dimensional gradation value from which the outline of the original image is extracted is entered into frame number A, and 9 is substituted into frame number ().

Further, a counter n indicating the recent number of times is set to 1. Next, the process proceeds to step S+2, where the Laplacian image ΔΔ is written into the working frame memory W1, and 1 working frame memory W.
The contents of the frame memory V are added to the data read from the frame memory V and written into the working frame memory W2, and the data read from the working frame memory W2 is written into the frame memory ■. In step SI2, each of the above-mentioned calculations is performed sequentially, so a time period of three frames is required. Next, in step S13, the data read from the frame memory V is multiplied by 1/8 and written into the working frame memory W+. Write the data of W2 to frame memory A. Since each calculation described in step SI3 requires one frame period, three frame periods are required in total. Step 12 performs the same calculation as step S2 in the flowchart of FIG. 4, and step SI3 performs the same calculation as step S3 in FIG.

Next, the process proceeds to step SI4, where the values of the addresses x, y, and n of the frame memory Δ, which are equal to or higher than the predetermined threshold value S, are set as the fifth value.
Calculate your fortune on a peak table like the one shown in the figure. Step S
At +5, the counter n is incremented by one, and at step 16 it is determined whether n is greater than or equal to R/27T pixels,
In the case of a negative determination, steps S12 to SI5 are repeated, and in the case of an affirmative determination, this processing routine is ended. - As mentioned above, in this modification, four frame memories are used, and the time required for one approximation is six frames and one frame period.

As a further variation, more working memory may be used to perform the operations in a shorter time than in the embodiment of FIG.

FIG. 7 shows a modification when five frame memories are used.

In FIG. 7, first, in step S21, two-dimensional grayscale values of the original image are written into the frame memory Aeven, and 0 is substituted into the frame memory Even. Further, a counter n is set to 1 without indicating the number of times of approximation. In step S22, the Laplacian image ΔAeven of the frame memory Δe Ven is written to the frame memory W, and the data read from the frame memory W and the frame memory V (data read from 3ven and the old In step S23.

71) - Multiply the data read from the frame memory Aeven by 1/8 and write it to the frame memory Veven, and add the data read from the frame memory Veven to the data read from the frame memory Aeven to erase the frame. Predict on Δodd. Step S
In step S22 and step S23, one frame cycle is required to perform the respective calculations, so step S22
In order to execute step S23 and step S23, a time of two frame periods is required. Step S22 executes the same calculation as step S2 in the flowchart of FIG.
Step S23 executes the same calculation as step S3 in FIG.

Next, proceeding to step 5A24, frame memory A
The values of addresses x, y, and n for which o dd is greater than or equal to a predetermined threshold value S are displayed in a peak table as shown in FIG. At step 322 ~ step S24, f7jri number 1 "
This means that one approximation cycle is being executed.

In step S25, the counter n is incremented by one, and the process proceeds to steps 326 to 328, where even number 1 is selected.
Execute the approximation cycle. In steps 326 to 328, exactly the same processing as in steps 322 to S24 is performed by changing even to odd and odd to even.

In step S29, the value of the counter n is further incremented by one, and then the process proceeds to step 530, where n is R/27.
It is determined whether the number of pixels is greater than or equal to T pixels, and if the determination is negative, the process from step 22 is repeatedly executed, and if the determination is positive, the present processing routine is terminated. In this embodiment, since five frame memories are used, calculations can be executed in a period of four frames.

[Effects of the Invention] According to the present invention, it is possible to accurately recognize a nine-circle figure even when the shading is different in the two-circle figure portion or the contrast with the background is small.

[Brief explanation of the drawing]

FIG. 1 is a principle block diagram according to the present invention. FIG. 2 is a schematic configuration diagram of the image processing processor of the embodiment. FIG. 3 is a schematic explanatory diagram of a method for extracting and emphasizing the outline of an image of a circular object. FIG. 4 is a flowchart of the figure recognition processing procedure using the wave equation in the embodiment device. Figure 5 is a schematic diagram of the peak table. FIG. 6 is a flowchart of a figure recognition processing procedure in a modified example. FIG. 7 is a flowchart of the figure recognition processing procedure in another modification; FIG. 8 is a diagram explaining the conventional problems. Frame memory, bidirectional path buffer video bus, spatial filter, addition circuit Δ/D conversion and D/Δ conversion circuit TE4 country, Figure 8, form 111 mia 1, the number 111, the number 6, the deformed material 11];

Claims (1)

[Claims] Computing means (
11) to solve a two-dimensional wave equation with the gradation value as an initial value, and to recognize the center position of the circular object based on the extreme value of the amplitude in the solution of the wave equation. The apparatus includes a contour extraction means (12) for extracting the contour of an input image, and is configured such that the input image whose contour has been extracted by the contour extraction means (12) is input to the calculation means (11). A figure recognition device featuring: