JPH0551944B2 - - Google Patents

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a binarization processing method for obtaining binarized image data optimal for recognizing the outline and shape of an observed object.

[Conventional technology]

This type of object recognition device that has been used in the past uses a camera to image an object as a standard sample in advance, and converts the obtained camera output into a binary value, which is used as a standard for determining good quality. Obtain the conversion level, binarize the output of the camera obtained by imaging the inspection object, and compare it with the above-obtained binary level as a standard for determining good quality.
Good or defective products are determined based on the number of pixels in the valued image and their distribution (histogram). However, it is only in very special cases that the distinction between good and defective products is clearly expressed as a binary level, and the level is usually ambiguous. It is difficult to uniformly determine. Furthermore, even if this is determined once, if there are changes in the testing environment, such as changes in the brightness of the room due to the weather on the day of the test, or subtle differences in the surface condition of the test object, it may be difficult to find a non-defective product. may be determined to be defective. Particularly in the case of a deep object, any of the plurality of feature points will be out of focus of the camera, making it impossible to obtain an image with good reproducibility and reducing reliability. Therefore, it is conceivable to perform determination for each feature point separately using a plurality of cameras, but this operation is extremely complicated. Further, for example, in the method disclosed in Japanese Patent Application Laid-Open No. 62-47788, a video signal corresponding to an image with blurred edges is obtained using a camera, and standard data obtained from a reference object and This is done by comparing the measured data obtained from the observed object. However, even in this conventional technology,
Standard data must always be obtained in response to changes in the environment, and making such decisions is extremely difficult. Further, the output of the image sensor is converted into binary image data using two reference levels via a binary conversion circuit, and the focus is corrected by the lens focus correction amount calculated from the difference image of the image data. The technology is
For example, JP-A-55-137784, JP-A-61-
It is disclosed in JP-A No. 20468, Japanese Patent Application Laid-open No. 129016-1980, etc.

[Problem to be solved by the invention]

However, all of these conventional techniques are designed to easily obtain a clear image by accurately focusing on one specific surface, but an accurately focused image has other parts that are extremely sharp. The image may be blurred and therefore not necessarily the best image for recognizing all parts of the shape of the observed object.
For example, in an image in which one face of a staircase shape is accurately focused, the shape of other parts that are out of focus cannot be recognized. It is better to set the focal length to a point where one specific surface cannot be precisely focused, but all other parts, including the shapes of other surfaces, can be sufficiently recognized. This may be desirable for recognizing contour shapes. Therefore, the present invention provides image processing that makes it possible to easily obtain a binarized image that is optimal for properly recognizing the contour shape of a complicated object to be observed, such as a staircase shape or a deep object. The object of the present invention is to provide a binarization processing method in an apparatus.

[Means to solve the problem]

In order to achieve the above object, the binarization processing method in the image processing apparatus of the present invention includes: imaging the observed object 1 by the imaging means 6 at a certain focal length and outputting a signal, and setting the binarization reference level to Q. The image pickup signal is outputted as binarized image data through the first binarization circuit 10, and a second binary signal whose binarization reference level is Q-Δq which is different from the above Q by a certain level is outputted as binarized image data. conversion circuit 1
1, and outputs the imaging signal as binarized image data through a circuit 1, and excludes the binarized image data from the first binarizing circuit and the binarized image data from the second binarizing circuit. The difference pixel number K is detected by calculating the logical OR, and the focal length of the imaging means is changed by a predetermined amount in accordance with the difference pixel number K, and the Execute the steps ~ to detect a new number of difference pixels K, calculate the difference Δk between the number of difference pixels K in the above step and the new number K of difference pixels in the above step, and set the above Δk to a predetermined reference value. The above steps are performed until K0 is reached, and the binarized image data at the focal length when the above Δk becomes less than or equal to a predetermined reference value K0 is determined to be appropriate.

【Example】

FIG. 1 shows the overall configuration of an apparatus for carrying out a binarization processing method in an image processing apparatus of the present invention. The observed object 1 is positioned and fixed on a carrier 2, and is subjected to binarization processing according to the present invention while being transported from the front process 3 to the rear process 4.
Recognition of the shape of the object is performed. The sensor 5 detects that the carrier 2 has reached this recognition position. A focusing mechanism 6 is placed directly above this recognition position.
A video camera 6, which is an example of an imaging means equipped with a video camera 6, is arranged to be able to take an image of the object 1. The video camera 6 and the sensor 5 are controlled by a control circuit built into the recognition device main body 7. The final judgment after recognition is sent to the subsequent process 4 as an output signal from the terminal 8. The device main body 7 is equipped with a display 7a, which displays a binarized image based on binarized image data obtained by imaging. The video camera 6 is scanned along the line i-i' in FIG. 2, images the object 1, and outputs the output voltage after photoelectric conversion to the control circuit. This object 1 has a complicated shape, and has four holes 1 on planes 1a, 1b, 1c, and 1d at different distances from the video camera 6.
e, 1f, 1g, and 1h have been established. Therefore, as shown in FIG. 2b, the holes 1e, 1f, 1
g and 1h are feature points, and areas centered around these holes are designated as observation areas j1, j2, j3, and j4. Observation areas j1 and j2 including 1e and 1f exist on line i-i', and each observation area is sequentially observed according to the following observation method. Figure 3 focuses on these holes 1e and 1f, and shows the cross section i.
-i' is shown as the output voltage for one line.
As described above, the surfaces 1a and 1b with the holes 1e and 1f are located at different distances from the video camera 6, so the cameras cannot be brought into focus at once. In particular, when capturing images in detail rather than close-up, the depth of focus of the lens system becomes shallow, and this tendency becomes more pronounced. Therefore, when the surface 1a is focused, the hole 1e can be imaged sharply, but the hole 1f becomes blurred. Therefore, if you try to binarize these images and measure the hole diameter, you can accurately determine the diameter of the hole 1e from the two reference levels Q1 and Q1 - Δq of the observation area j1. In j2, it is difficult to set the reference level Q2 itself, and the diameter of the hole 1f cannot be determined accurately. For the hole 1f in such a state, the binarization reference level Q2 and Q2−, which differs by the amount of change Δq, are
Binarized image data in the case of Δq are shown in Figure 4 a and b.
It is shown in And in Figure 4c, these two
This shows difference image data obtained by calculating the exclusive OR of two binary image data. Note that when the focal points match as in the hole 1e, the binarized image data hardly changes with respect to the amount of change Δq, and therefore no difference image data can be obtained. In the case of blur as in hole 1f, a considerable change occurs in the binarized image data relative to the amount of change Δq, and the difference image data is
The result will be as shown in Figure c. Letting K be the number of difference pixels of this difference image data, the following relationship exists. K=f1(Δq) (1) Therefore, when considering the lens system focus correction amount Δm with respect to the number of differential pixels K, the following relationship exists. Δm=f2(K) …(2) Generally, the functions f1 and f2 in equations (1) and (2) are nonlinear, but when considering the minute change Δq, they can be considered linear as in equation (3). Good too. Δm≒α·K (3) Here, α is a proportional coefficient, and equation (3) can be used in a table. Therefore, for each observation area, the reference level Q
The lens system focus correction amount Δm can be determined from the difference in the number of pixels K obtained by setting two levels, Q-Δq, and Q-Δq, which are slightly different from this level. This Δm is determined by the focusing mechanism 6 of the video camera 6.
You can change the focal length by instructing a. Using this changed focal length, a new difference pixel number K is detected in the same manner as above, and the original difference pixel number K is detected.
Calculate the difference Δk between and the new difference pixel number K, and calculate this Δk
The binarized image data at the focal length when becomes less than or equal to a predetermined reference value K0 is determined to be appropriate. In this way, binarized image data is obtained for each observation region of the observed object 1 at a focal length where Δk is equal to or less than the reference value K0. Next, the method of the present invention will be specifically explained with reference to FIG. First, prior to observation, the focal length to each observation area is measured in advance with respect to a standard observation object. Further, when correcting the focal length in the direction of increasing or decreasing the focal length, the focal length is set shorter or longer than the appropriate focal length, and the focal length is set so that a little blurring occurs. This makes it possible to control the focal length correction in only one direction. Therefore, 6 video cameras were installed for each observation area.
The object 1 in the observation area is imaged at the focal length and a signal is output. This signal is output as binarized image data a of "1" or "0" via the first binarization circuit 10 whose binarization reference level is Q, and at the same time, the binarization reference level is is outputted as binarized image data b via the second binarization circuit 11 where Q-Δq is different from Q by a certain level. Both binary image data a and b are subjected to an exclusive OR in a logical operation circuit 12, and counted by a counter 13 to detect the difference number K of pixels. The focus correction amount Δm corresponding to the difference pixel number K is stored in a table, so from this table, the focus correction amount Δm corresponding to the difference pixel number K is calculated.
is supplied to the focusing mechanism 6a to increase or decrease the focal length by Δm. Under this changed focal length, the above steps ~ are executed to detect a new difference pixel number K. Therefore, the original number of difference pixels K and the new number of difference pixels K
Calculate the difference Δk. In the normal case, the new K<
It becomes the original K. Therefore, as a convergence condition for correction, the difference Δk=(original K)−(new K) is determined, and the above steps are executed until the above Δk reaches a predetermined reference value K0. Then, processing is performed so that the binarized image data at the focal length when Δk becomes less than or equal to a predetermined reference value K0 is determined to be suitable. This reference value K0 is set in advance to an appropriate value according to the shape of the observed object 1. For example, if the hole is shallow and all parts of the object can be recognized by focusing on one surface, then K0 may be small.
Also, if the object has large steps such as stairs,
K0 will be set large. After obtaining appropriate binarized image data for all observation areas j1 to j4 as described above, the focal length is reset to the first initial focal length and waits for the next object to be observed to be transported. Note that although the above example shows an object that has a hole in each observation area, it is an especially excellent method for binarization processing when the object has a shape that has a protrusion in the center, and can be performed quickly. Appropriate binarized image data can be obtained.

【Effect of the invention】

According to the binarization processing method of the present invention as described above, it is possible to easily and appropriately obtain binarized image data that is optimal for recognizing the shape and outline of an object to be observed that has a stepped shape or depth. Since this data can be used to immediately determine whether the observed object is good or not, highly accurate determination results can be obtained and reliability is improved.

[Brief explanation of drawings]

The drawings are explanatory diagrams for explaining the binarization processing method of the present invention. Fig. 1 is a perspective view showing the overall configuration of an apparatus for carrying out the method of the present invention, and Fig. 2 a and b show the observation area. A plan view for explanation, Fig. 3 is an output voltage waveform diagram obtained by capturing images of two observation areas with a camera of the same focal length, and Fig. 4 a shows the output voltage waveform at the standard level Q for binarization in the case of out of focus. The digitized waveform diagram, Figure 4b, is 2 at the reference level Q-Δq in the same as above.
FIG. 4c is a binary output waveform diagram of the exclusive OR of the two binary outputs, and FIG. 5 is a basic configuration diagram for calculating the lens focus correction amount. 1... Observation object, 6... Imaging means (video camera), 10, 11... Binarization circuit, Q, Q-Δq...2
Value standard level, Δm...Focus correction amount.

Claims (1)

[Claims] 1. The imaging means 6 images the object 1 to be observed at a certain focal length and outputs a signal, and the above-mentioned image is taken through the first binarization circuit 10 whose binarization reference level is Q. A second binarization circuit 1 which outputs a signal as binarized image data and whose binarization reference level is Q-Δq which is different from the above Q by a certain level.
1, and outputs the imaging signal as binarized image data through a circuit 1, and excludes the binarized image data from the first binarizing circuit and the binarized image data from the second binarizing circuit. The difference pixel number K is detected by calculating the logical OR, and the focal length of the imaging means is changed by a predetermined amount in accordance with the difference pixel number K, and the Execute the steps ~ to detect a new number of difference pixels K, calculate the difference Δk between the number of difference pixels K in the above step and the new number K of difference pixels in the above step, and set the above Δk to a predetermined reference value. Binary image data in an image processing device characterized in that the above steps are executed until K0 is reached, and the binarized image data at the focal length when the Δk becomes equal to or less than a predetermined reference value K0 is appropriate. processing method.