JP2004062266A - Method of detecting edge position of imaged body - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of detecting an edge position of an imaged body, in which an edge position is detected only when the edge part is detected, without erroneously detecting parts which are not actually an edge as an edge part. <P>SOLUTION: Gradation image data are obtained by a CCD camera 1 and the data are binarized to generate binary image data, which are shrunk to generate shrunk image data from which noise mixed therein is removed. respective rates of white pixels and black pixels in the shrunk image data are calculated and only when the calculated rates of the white pixels and black pixels are larger than preset specified values respectively, the detecting process for an edge position is carried out. Consequently, only when an edge part is detected, the edge position is detected to shorten the process time. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for detecting an edge position of an object to be imaged to which an image processing technique is applied, and in particular, it is possible to detect an edge position only when an input grayscale image includes an edge portion. The present invention relates to an edge position detection method.
[0002]
[Prior art]
When detecting an edge position included in a grayscale image of an object imaged by a CCD camera or the like, differential image data is created by differentiating the grayscale image, and a binary image is generated for the differential image data. The binary image data is created by performing the binarization process, and a white pixel group in the binary image data is extracted (parted), and the largest white pixel group (maximum part) among the extracted white pixel groups Is detected as an edge portion, and an edge position is detected from differential image data in the edge portion (in the largest part). In this method, since the largest part in the binary image data is detected as an edge portion, it is important to set a threshold value for binarization. In setting this threshold value, it is necessary to detect the maximum part (edge portion) without being affected by the contrast of the image. Therefore, a preset ratio (density maximum value) with respect to the density maximum value in the differential image data (30 to 50% of the threshold). Then, in the differential image data, the binary image data is created such that the pixels above the threshold become white pixels and the pixels below the threshold become black pixels.
[0003]
However, in the case of this method, when there is no edge portion in the grayscale image, and in the case of an image having a small density value (dark image), the edge is actually not affected by the noise component included in the differential image data. There is a problem that a missing part is erroneously detected as an edge part. As a method for solving this problem, there are a method of setting a threshold value for binarization to a fixed value that is not affected by noise, and a method of determining whether an edge portion exists in a grayscale image by pattern matching. there were.
[0004]
[Problems to be solved by the invention]
However, in the former method, that is, in the method of setting the threshold value of binarization to a fixed value not influenced by noise, generally a high value is set as the threshold value of binarization not influenced by noise. For example, when the contrast is low, there is a problem that the edge position cannot be detected even if the edge portion exists in the grayscale image. The latter method, that is, the method of determining whether or not an edge portion exists in a grayscale image by pattern matching, is effective when the edge shape of the object is always imaged in the same state or condition. However, there is a problem that the edge cannot be detected when the shape of the edge changes every time.
[0005]
The present invention has been made in order to solve the problems of the related art, and does not erroneously detect a part that is not actually an edge as an edge part, and furthermore, detects an edge position only when an edge part is detected. It is an object of the present invention to provide a method for detecting an edge position of an object to be imaged, which performs a process of detecting an image.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, in the invention according to the first aspect, binary image data is created by acquiring grayscale image data with a CCD camera and performing binarization processing on the grayscale image data. By performing contraction processing on the binary image data, contracted image data is created, the respective proportions of white pixels and black pixels in the contracted image data are calculated, and each of the calculated white pixels and black pixels is calculated. The edge position detection method is characterized in that the edge position detection process is performed only when the ratio of each is equal to or greater than a preset specified value.
[0007]
When noise (noise) is mixed in the grayscale image data, a region that is originally a black pixel in the binary image data may be partially (in units of one pixel) a white pixel. Therefore, in the configuration according to claim 1, noise (noise) mixed in the grayscale image data is eliminated by performing the contraction process on the binary image data. The shrinking process referred to here is a process of converting a white pixel into a black pixel in a binarized screen when the eight directions around the pixel are not all white pixels. With regard to a pixel that has become a white pixel due to the incorporation of noise (noise), it is unlikely that all of the pixels in the eight directions around the pixel are white pixels. Will be converted to the original black pixels. Therefore, by performing the contraction processing on the binary image data, noise (noise) mixed in the grayscale image data is eliminated.
[0008]
Generally, when detecting an edge position, a window area is set, and then an edge position detection process is performed in the window area. This is performed in all areas of the grayscale image data acquired by the CCD camera. Like that. However, performing the edge position detection processing even in a window region where it is apparently determined that no edge position is present increases the overall processing time. Therefore, according to the first aspect of the present invention, when binary image data is created by performing binarization processing on grayscale image data, if there is no apparent edge position in the window area, the window Attention was paid to the fact that the whole area is either a white pixel or a black pixel. That is, with respect to the contraction processing data from which noise (noise) mixed by performing the contraction processing on the binary image data is calculated, the respective ratios of the white pixels and the black pixels in the window area are calculated, and the ratio of the white pixels is calculated. Only when both the ratio of the black pixel and the ratio of the black pixels are equal to or greater than a predetermined value, it is determined that an edge position exists in the window region, and the detection process of the edge position in the window region is performed. did. As a result, the processing for detecting the edge position is not performed in the window region where the edge position does not clearly exist, so that the overall processing time can be reduced.
[0009]
In the invention according to claim 2, in the invention according to claim 1, when each of the calculated ratios of the white pixels and the black pixels is equal to or more than a predetermined value, the grayscale image data is differentiated. Differential image data is created, and a binary process is performed on the differential image data to create second binary image data, and a white pixel group in the second binary image data is extracted. An edge position is detected for the differential image data in the largest white pixel group among the extracted white pixel groups. With this configuration, the edge position detection process, that is, the differentiation process, the binarization process, and the white pixel group (parts) extraction process are performed only in the window area determined to have the edge position. Become.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the present embodiment, the inspection table is transparent, illumination is provided below the inspection table, and the object to be imaged is photographed by the CCD camera 1 provided above the inspection table. FIG. 2 is a block diagram illustrating an example of an image processing system to which the edge position detection method according to the present invention is applied. In FIG. 2, 1 is a CCD camera, 2 is a computer main body, 3 is a mouse, 4 is a keyboard, and 5 is a CRT screen. The CCD camera 1 captures an image of an object placed on an inspection table (not shown). The grayscale image data captured by the CCD camera 1 is stored in a grayscale image memory 9 via an interface (hereinafter, referred to as “I / F”) 6. The grayscale image data stored in the grayscale image memory 9 is displayed on the CRT screen 5 via the display control unit 10. Position information input from the mouse 3 and the keyboard 4 as data input devices is input to the CPU 31 via the I / F 7 and the I / F 8 respectively.
[0011]
The CPU 31 creates data for displaying a rectangular area specified by the mouse 3 or the keyboard 4 as a window in accordance with the program stored in the program memory 32, and extracts grayscale image data inside the window from the grayscale image memory 9. Then, a series of processes relating to detection of an edge position included in the window, such as creation of a binary image, creation of a contracted image, and calculation of the area ratio of each of white pixels and black pixels, which will be described later, are sequentially executed. The work memory 33 is a memory for a work area for various processes in the CPU 31.
[0012]
Next, the flow of processing of the edge position detection method according to an embodiment of the present invention will be described with reference to the flowchart shown in FIG. When the window for performing the image processing is designated by the mouse 3 or the keyboard 4, the CPU 31 acquires the gray image data in the window area from the gray image memory 9 (step 11). As described above, in the present embodiment, the inspection table is transparent, illumination is provided below the inspection table, and the object to be imaged is photographed by the CCD camera 1 provided above the inspection table. As the grayscale image data acquired in step 11, the portion where the object is present is black or data close thereto, while the portion where the image is not present is white or data close thereto.
[0013]
Next, the first binary image data is created by performing a binarization process on each pixel of the grayscale image data acquired in Step 11 based on a preset threshold value (Step 12). As described above, in the present embodiment, as the grayscale image data, a portion where the object is present is black or data close to black, while a portion where the object is not present is white or data close thereto. Therefore, as the first binary image data created in step 12, the portion where the object is present is generally a black pixel, while the portion where the object is not present is a white pixel.
[0014]
By the way, in the image processing system according to the present embodiment, it is inevitable that noise (noise) generated at random is mixed. Then, due to the influence of the noise (noise), in the first binary image data created in step 12, an area that is originally a black pixel may partially (in units of one pixel) become a white pixel. Therefore, with respect to the first binary image data created in step 12, all pixels near black pixels are blackened by shrinkage processing, so that shrinkage image data excluding the influence of noise (noise) is created. (Step 13). Note that the shrinking process referred to here is a process of converting a white pixel into a black pixel when all eight directions around the pixel are not white pixels in the binarized screen. By the processing in step 13, even if a pixel that originally becomes a black pixel due to noise or the like generated at random is a white pixel, the white pixel is converted to an original black pixel. It is possible to eliminate the influence of noise or the like generated at random.
[0015]
Subsequently, for the contracted image data created in step 13, the respective numbers of white pixels and black pixels in the window area are counted, and thereby the respective areas of the white pixel area and the black pixel area in the window area are calculated. (Step 14). Next, the ratio of the area of each of the white pixel region and the black pixel region calculated in step 14 to the area of the entire window region is calculated (step 15).
[0016]
The two area ratios calculated in step 15, that is, whether both the area ratio of the white pixel area to the entire area of the window area and the area ratio of the black pixel area to the entire area of the window area are equal to or greater than a predetermined ratio are determined. Then, it is determined whether or not the edge of the object is valid (step 16). That is, if any one of the two area ratios calculated in step 15 is less than the preset ratio, it is determined that there is no edge portion in the window area (step 16N), and the window area is determined. The processing in the area ends. On the other hand, if both of the two area ratios calculated in step 15 are equal to or larger than the preset ratio, it is determined that an edge portion exists in the window area (step 16Y), and the following steps 17 and subsequent steps are performed. Is performed. Note that, for the above-described preset ratio, a ratio that can accurately and stably detect an edge position with an effective area as an edge portion of the object to be imaged when the edge position is detected is set. To do.
[0017]
That is, differentiated image data is created by differentiating each pixel of the grayscale image data acquired in step 11 (step 17). Next, a threshold value is set for the differential image data created in step 17 based on the maximum density value of the differential image data, and binarization processing is performed on the differential image data based on the threshold value. 2, binary image data is created (step 18). In addition, as a method of setting the threshold value in step 18, a threshold value that can accurately and stably detect an edge position without erroneously detecting an edge portion according to an imaging state of an edge portion of an object to be imaged. (Density value). More specifically, although it depends on the shape of the object to be imaged, the measurement conditions, and the like, it is preferable to set the density within a range of 30 to 50% of the maximum density value of the differential image data.
[0018]
A white pixel group is extracted from the second binary image data created in step 18, and each extracted white pixel group is converted into one part (step 19). Next, the largest part is detected from all the parts in the window area that has been made into parts in step 19 (step 20). Finally, in the maximum part detected in step 20, an edge position is detected with respect to the differential image data created in step 17 in the part (step 21).
[0019]
Hereinabove, one embodiment of the present invention has been described. In the present embodiment, as described above, the inspection table is transparent, illumination is provided below the inspection table, and the object to be imaged is photographed by the CCD camera provided above the inspection table. The invention is not limited to this embodiment. For example, the inspection table may have a mirror surface, an illumination provided above the inspection table, and an image of an object to be imaged taken by a CCD camera provided above the inspection table. In such an embodiment, as in the above-described embodiment, as the grayscale image data, a portion where the object is present is black or data close thereto, while a portion where the object is not present is white or white. It becomes close data. Further, as another form, when the surface is an object to be imaged in a mirror surface state, the surface of the inspection table is assumed to have less reflection of light, and illumination is provided on the inspection table, and provided on the inspection table. The object to be imaged may be photographed by the CCD camera. In such an embodiment, unlike the above-described embodiment, the location where the object is present is white or data close thereto, while the location where the image is not present is black or data close thereto, The flow of the process in detecting the edge position is the same as in the above-described embodiment.
[0020]
【The invention's effect】
According to the present invention, gray-scale image data is obtained by a CCD camera, and binary image processing is performed on the gray-scale image data to generate binary image data. By doing so, contracted image data from which noise (noise) mixed in is removed is created, the respective ratios of white pixels and black pixels in the contracted image data are calculated, and the calculated ratios of white pixels and black pixels are calculated. , The edge position detection processing is performed only when the value is equal to or greater than a predetermined value, so that a part that is not actually an edge due to noise (noise) is not erroneously detected as an edge part. Further, the processing for detecting the edge position is performed only when the edge portion is detected, thereby shortening the processing time.
[Brief description of the drawings]
FIG. 1 is a flowchart showing a processing flow of an edge position detection method according to the present invention.
FIG. 2 is a block diagram illustrating an example of an image processing system to which an edge position detection method according to the present invention is applied.
[Explanation of symbols]
1 CCD camera 2 Computer body 3 Mouse 4 Keyboard 5 CRT screen 6, 7, 8 I / F
9 Contrast image memory 10 Display control unit 31 CPU
32 Program memory 33 Work memory

Claims (2)

Obtain grayscale image data with a CCD camera,
By performing binarization processing on the grayscale image data, binary image data is created,
Shrinking image data is created by performing shrinking processing on the binary image data,
Calculate each ratio of white pixels and black pixels in the contracted image data,
An edge position detection method for an image pickup object, wherein the edge position detection process is performed only when the calculated ratio of each of the white pixels and the black pixels is equal to or greater than a predetermined value. . Obtain grayscale image data with a CCD camera,
By performing binarization processing on the grayscale image data, first binary image data is created,
Shrinking image data is created by performing shrinking processing on the first binary image data,
Calculate each ratio of white pixels and black pixels in the contracted image data,
If each of the calculated ratios of the white pixels and the black pixels is equal to or greater than a preset specified value,
Create differential image data by differentiating the gray image data,
By performing binarization processing on the differential image data, second binary image data is created,
Extracting a white pixel group in the second binary image data;
An edge position detection method for an object to be imaged, wherein an edge position detection process is performed on the differential image data in the largest white pixel group among the extracted white pixel groups.

Images