JPH0728998A - Expansion processing method for object in picture - Google Patents

Abstract

PURPOSE:To expand an object 2 without changing the density state of the presence area of the object before an expansion processing of the object by performing the expansion processing only on the picture elements of a part judged that the object is not present by a threshold value for discrimination the object from a background. CONSTITUTION:In pictures to be a processing object obtained by picking up the images of parts, the density data of the respective picture elements are generated, a density value for the picture element of interest is compared with the threshold value for discriminating the object 80 from the background 85 and partial pictures 10 are set only in the case of being equal to or less than the threshold value, that is, only in the case of the picture element where the object 80 is not present. Then, in a selective expansion processing process, in the case of the partial picture 10, '3' of the maximum value among the density values '1', '3', '3' and '3' of the respective picture elements is selected and an output density value '3' for an output picture element corresponding to the picture element 11 of interest in the output picture elements is set as output density data. Then, whether or not is is a final picture element of interest is judged, the picture element of interest is moved when it is not the final picture element under consideration, similar processings are repeated until the final picture element of interest and the expansion processing is performed only for the outside of the object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for subjecting an object existing in an image to expansion processing based on density data representing the degree of shading of each pixel in the image to be processed. The present invention relates to a method for expanding an object in an image, in which the gradation image obtained by performing the expansion process is processed.

[0002]

2. Description of the Related Art Image expansion processing is often used in combination with contraction processing to smooth a binary image. Further, it has been conventionally performed that a grayscale image is processed and the density data of the image is used to perform an expansion process on an object existing in the image.

The expansion process of the grayscale image is sequentially performed while comparing the density value of the pixel of interest in the image to be processed with the density values of the peripheral pixels.

This will be described below. First, the density data in each pixel of the image to be processed is obtained from a video signal output from an image input device such as a video camera. Figure 7
Shows a concrete example of this, and pixel density data a, ..., E, ..., I, ... Corresponding to each pixel of the image to be processed.
・ I am getting.

Next, the following specific methods have been proposed for performing the expansion treatment.

For example, as shown in FIG. 7, a 3 × 3 partial image 90 consisting of a target pixel e and nine surrounding pixels is set, and the densities of these pixels a to i as shown in FIG. 8 are set. Maximum value filtering is performed based on the data. This filtering process is a process of extracting the maximum density value out of the density values of the nine pixels and using that density value as the density value of the pixel of interest. Of the density values of the nine pixels of the partial image 90, the maximum density value of 30 is selected by the maximum value filter processing,
When the density data of the target pixel e is 25, it is set to 30, and the expansion processing is performed.

When the processing on the partial image 90 is completed, the pixel is shifted horizontally by one pixel so that the pixels a, d and g disappear, and instead, the partial image 91 in which the pixels j, k and l are added to the right side is newly set. The same expansion process as described above is performed. When this is completed, the same process is repeated again by shifting by one pixel, and the processes are performed one after another.

[0008]

By the way, the conventional expansion processing method for images has the following problems. Figure 9
An example will be described in which the object 95, which is an image of the object captured in the target image, exists in the background 96 as shown in FIG. When the maximum value filtering process described above is performed on the entire image to expand the object 95 based on the density data of each pixel of the partial image as shown in FIG. 8, the expanded part 97 shown in FIG.
As described above, the outer circumference of the object 95 can be expanded.

However, the density state of the object 95 changes. This is often against the wish. In this case, if there is a large change in density in the background 96,
Alternatively, even if there is, it can be sufficiently distinguished from the object 95.

Explaining the change in the density state of the object 95, when the dilation processing is applied to the binary image, the object becomes a binary level "1" or "0", so that the change in the density state of the object is It won't happen.

However, in the case of a grayscale image, the density of each pixel of the object 95 is not constant as compared with the binary image. For this reason,
If the pixel of interest exists in the object and the other pixels in the partial image including this pixel have the maximum density, this maximum density is set as the density of the pixel of interest, so the density state of the object 95 changes. .

However, before performing the expansion process, the object 95
If the contour extraction is performed and only the outer peripheral portion of the object 95 is expanded, the demand that the density state of the object 95 is not changed can be satisfied. However, the processing algorithm in this case is very complicated, and the expansion processing takes too much time, which is not practical.

In order to solve the above-mentioned problems, the present invention does not require a complicated processing algorithm, and detects the outer circumference of an object from the grayscale image of the object without changing the density state of the object in the image. It is an object of the present invention to provide a method for expanding an object in an image, which can perform expansion processing only.

[0014]

In order to achieve the above-mentioned object, the method for processing object expansion in an image according to the present invention uses density data representing the degree of shading of each pixel in the image to be processed. Based on the density of the pixel of interest based on a predetermined density value such as the maximum density value of the pixel in the partial image including this pixel, the expansion processing of the object in the image for expanding the object existing in the image In the method, an object presence determination step of determining the presence or absence of an object in the background based on a threshold value that distinguishes the object from the background, and a partial image in a partial image including only the pixels of the portion determined to have no object The present invention is characterized by having a selective expansion processing step of performing expansion processing based on a predetermined density value of each pixel.

[0015]

In the above structure of the object expansion processing method for an image of the present invention, in the object existence determining step, the existence area of the object in the image is judged by the threshold value for distinguishing the preset object from the background. Every time, only for the area portion where it is determined that the object does not exist, that is, only for the pixels of the background portion located around the outside of the object, the predetermined density value in each pixel in the partial image including this pixel is sequentially set, Since the expansion process is performed only on the outer circumference, the object can be expanded without changing the density state of the existing region of the object before the expansion process.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to FIGS.
It will be explained based on. Of course, the present invention is not limited to the following embodiments.

FIG. 2 shows an expansion processing apparatus used in the method for processing the expansion of an object in an image as an embodiment of the present invention.

As shown in the figure, this device irradiates the surface of the component 1 as the object with the light of the light source 16 through the half mirror 15, and the surface of the component 1 by the image pickup device 2 through the half mirror 15. Image. This imaging device 2 is CC
A D sensor is provided, and a grayscale signal for each pixel in the captured image is obtained by this.

This gradation signal is sent to the arithmetic unit 3 in a digitized form. In the arithmetic unit 3, for example, 8-bit density data is stored in the storage unit in a form corresponding to the rows and columns of the pixels. The arithmetic unit 3 is set with a program that executes the following processing.

FIG. 3 shows an example of density data of an image to be processed. An object 80, which is an image obtained by picking up an image of the component 1, exists in a background 85. Here, it is assumed that the density data of the object 80 is significantly larger than the density data of the background 85, and the density threshold value gth for distinguishing the existence of the object 80 from the background 85 is set in advance, and In that case, the value is 100.

FIG. 1 is a flow chart showing the specific processing of the method for processing object expansion in an image according to the present invention.

This will be described. First, step #
In the image pickup process of the component 1 of No. 1, density data representing the degree of shading of each pixel is created in the image to be processed obtained by imaging the component 1.

In step # 2, the density value of the pixel of interest is compared with the threshold value gth, and the threshold value (gth = 1
00) or less, that is, only in the case where the object 80 is a pixel that does not exist, step # 3 for expansion processing is performed.
Proceed to. If not, it is a pixel in the object 80, so the process proceeds to step 5 so as not to perform expansion processing.

Target pixel 1 having a density value of 1 in this embodiment
In the case of 1, the value is less than the threshold value, and therefore the process proceeds to step # 3 to perform the partial image setting process, for example, to set the partial image 10 as shown in FIG. The partial image 10 has a 3 × 3 9-pixel configuration, and 8 pixels are set around the target pixel 11. When the density value of the pixel does not exist, expansion processing is performed only with the density value of the existing pixel.

In the next step # 4, a selective expansion processing step is performed. In this embodiment, a case will be considered in which the maximum value filtering process for obtaining the maximum density value of a partial image is used. In the case of the partial image 10, the maximum value of 3 is selected from the density values 1, 3, 3 and 3 of each pixel, and the pixel of interest 11 in the output image shown in FIG.
The output density value 3 for the output pixel corresponding to is set as the output density data 12.

Subsequently, in step # 6, a final pixel of interest determination step is performed to determine whether or not the pixel is the final pixel of interest.
If it is not the final target pixel, the target pixel is moved in step # 7.

In this way, steps # 2 to # 7 are repeated until the final target pixel is reached.

In such a process, the case of the pixel of interest 21 having a density value of 3 shown in FIG. 3 will be described. Since the density value 3 of the target pixel 21 is also equal to or less than the threshold value, the process proceeds from step # 2 to step # 3 to set the partial image 20. Then, in step # 4, the density values 1, 3, 3, 3, 3, 3, 3, 3, of each pixel in the partial image 20.
The maximum value 105 is selected from 105, and the output density value 105 for the output pixel corresponding to the target pixel 21 in the output image shown in FIG. 4 is set as the output density data 22.

Next, the case of the target pixel 31 whose density value is 111 shown in FIG. 3 will be described.

Since the density value 111 of the target pixel 31 is equal to or larger than the threshold value, the process proceeds from step # 2 to step # 5 and corresponds to the target pixel 31 in the output image shown in FIG. 4 in the target pixel setting step. As the output density value of the output pixel, the density value 111 of the target pixel is set as the output density data 32 as it is.

According to the expansion processing as described above, only the outer circumference of the unprocessed object 80 is replaced with the maximum density value of each pixel of the partial image, and the expansion processing is performed.
The object 80 can be inflated as indicated by 81 in FIG. 4 without changing the density state before the inflating process of 0.

In the above embodiment, the pixels are moved one by one to set the partial image, and very fine scanning is performed to perform expansion processing. At the time of setting, pixels may be moved one by one (or two by two) to perform interlaced scanning. When such interlaced scanning is performed, the processing time is short.

As the partial image to be set, 8 pixels are set around the target pixel, but the width or the number is not limited. For example, the second embodiment of the present invention shown in FIG.
As in the embodiment of FIG.
The expansion processing may be performed such that the partial image areas 72 are set to be separated by 1. In addition, as in the third embodiment of the present invention shown in FIG. 6, the dead zone region 76 around the pixel of interest 75.
It is also possible to set the partial image area 77 by separating them. By setting such dead zone regions 71 and 76, the processing time can be shortened. 5 and 6, L1 indicates the width of the dead zone area, and L2 indicates the width of the partial image setting area.

As the partial image, a rectangular image composed of the pixel of interest and the surrounding pixels is usually cited, but the partial image is not limited to this, and may be of different sizes and shapes.

For example, it may have a circular shape. The pixel of interest is usually located at the center of the partial image, but in some cases, for example, when the partial image having the number of pixels of 2 × 6 is set, the pixel of interest does not completely coincide with the center of the partial image. In the case of the present invention, such an incomplete match may be made.

Further, in each of the above-described embodiments, in the expansion process for the image of interest, the maximum density for each pixel in various partial images including this image of interest is selected as a predetermined density value, and the predetermined density value is replaced with the maximum density value for calculation. Was adopted,
Any of the maximum value, the average value, and the mode value can be set as the predetermined density value for the expansion process, and the calculation method is not limited. Usually, the maximum value is calculated. However, when obtaining the maximum value or when comparing with the threshold value,
Depending on whether the image is a positive image or a negative image, the magnitude relation of comparison is reversed, so that the corresponding predetermined density value is selected accordingly. When the density data of the object is smaller than that of the background, unlike the above-described embodiment, the minimum value may be obtained and used as the predetermined density value for the expansion process.

Further, if the threshold for distinguishing the object from the background is set from the histogram of the density data based on the ratio of the area of the object to the area of the non-object, the threshold can be set easily. In addition, the rationality of the threshold is improved, the convenience is increased, and the certainty is improved.

[0038]

According to the object expansion processing method for an image of the present invention, by performing the expansion processing only on the portion of the partial image which is not considered as an object, the density state of the object before the expansion processing does not change significantly. It is possible to achieve accurate expansion processing of the outer circumference of the object.

Moreover, since the processing is simple and no complicated processing algorithm is required, the object in the grayscale image can be expanded, and the processing time can be shortened and the cost can be reduced.

[Brief description of drawings]

FIG. 1 is a flowchart showing a concrete method of an expansion processing method according to a first embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of an apparatus used in the first embodiment.

FIG. 3 is a configuration diagram of a processing target image before expansion processing according to the first embodiment.

FIG. 4 is a configuration diagram of an output image after expansion processing according to the first embodiment.

FIG. 5 is a configuration diagram of a partial image according to a second embodiment of the present invention.

FIG. 6 is a configuration diagram of a partial image according to a third embodiment of the present invention.

FIG. 7 is a configuration diagram of a partial image of a conventional example.

FIG. 8 is a configuration diagram of a partial image for explaining a conventional expansion processing method.

FIG. 9 is an explanatory diagram of an image state before a dilation process of a conventional example.

FIG. 10 is an explanatory diagram of an image state after expansion processing of a conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Target 2 Imaging device 3 Computing device 10, 20 Partial image 11, 21 Target pixel 12, 22 Output density data 15 Half mirror 16 Illuminating means 80 Object 81 Object after expansion 85 Background

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location 8420-5L G06F 15/66 415

Claims (1)

[Claims] 1. The density of a pixel of interest is based on density data representing the degree of shading of each pixel in an image to be processed, and a predetermined density such as the maximum density value of a pixel in a partial image including this pixel. In the expansion processing method of the object in the image, which expands the object existing in the image, the existence of the object in the background is judged based on the threshold for distinguishing the object from the background. An image characterized by having a determination step and a selective expansion processing step of performing expansion processing based on a predetermined density value of each pixel in a partial image including only a pixel of a part determined to have no object The method of expanding the inside body.