JPH0591326A - Picture processing device - Google Patents

Abstract

(57) [Abstract] [Purpose] To provide an image processing apparatus capable of automatically setting a threshold value for binarization for obtaining a target binary image even if ambient brightness changes. It is an object. An image input device for inputting a grayscale image, and an output of the image input device are compared with a predetermined threshold value and binarized to output coordinate information of a binary image and length data of black pixels. In an image processing apparatus including a run length encoder and a processor that receives an output of the run length encoder, a comparison data storage unit that stores a characteristic state of a target binary image as comparison data is provided in the processor. The processor is configured to sequentially change the threshold value of the run length encoder so that the value of the binarized comparison data matches the reference value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus capable of automatically determining a maximum threshold value for binarization processing.

[0002]

2. Description of the Related Art There is known an image processing apparatus for converting a grayscale image into a binary image of white and black with a predetermined threshold value. For example, in FIG.
It is assumed that there is a grayscale image as shown in. It is considered that the image is a gray-scale image having a gradation number n + 1, where white is 0 and black is n.
Here, by counting and integrating the number of pixels of n + 1 gradations, a histogram (frequency distribution curve) as shown in FIG. 9 is obtained. In FIG. 9, the horizontal axis represents gradation (gray level) and the vertical axis represents the number of pixels.

It is assumed that when the threshold for binarization is determined from this histogram by using the histogram reference method, it becomes k in the figure. When the threshold value k is determined, this threshold value is used in FIG.
The grayscale image shown in is binarized. That is, a pixel having a density higher than k becomes "1", and a pixel having a density lower than k becomes "0". In this way, all pixels can be binarized.

[0004]

In the above-mentioned conventional image processing apparatus, since the threshold value is uniquely determined, the threshold value is not the optimum threshold value for binarization when ambient brightness changes. There is a problem that a preferable binarized image cannot be obtained.

The present invention has been made in view of the above problems, and automatically sets a binarization threshold value for obtaining a target binary image even if ambient brightness changes. It is an object of the present invention to provide an image processing device capable of performing the above.

[0006]

According to the present invention for solving the above-mentioned problems, an image input device for inputting a gray-scale image and an output of the image input device is binarized by comparing it with a predetermined threshold value. An image processing device comprising a run-length encoder that outputs coordinate information of a value image and length data of black pixels, and a processor that receives an output of the run-length encoder. A comparison data storage unit that stores the characteristic state of the image as a comparison data reference value is provided, and the processor sequentially changes the threshold value of the run length encoder so that the value of the binarized comparison data matches the reference value. It is characterized by having done.

[0007]

Operation: The threshold for binarization in the run length encoder is
The comparison data after binarization is sequentially changed so as to come close to the target comparison data (reference value).
With this configuration, even if the surrounding conditions change, if the reference value of the comparison data is fixed, the threshold value is automatically changed in the direction to match the comparison data. Even if is changed, a binarization threshold value for obtaining a target binary image can be automatically set, and an optimum binary image can always be obtained.

[0008]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a configuration block diagram showing an embodiment of the present invention. In the figure, 1 is an image input device for inputting a grayscale image, and 2 is an output of the image input device 1 and is compared with a predetermined threshold value to be binarized. It is a run-length encoder that outputs and. A processor 3 receives the output of the run length encoder 2.

In the run length encoder 2, 21
Is a threshold setting unit that holds a binarization threshold whose value is updated each time by an external signal, and 22 binarizes the grayscale image output from the image input apparatus 1 with the threshold given from the threshold setting unit 21. It is also a binarization control unit that outputs the coordinate information of the binary image and the length data of the black pixels. In the processor 3, 31 is a comparison data storage unit for storing the characteristic state of a target binary image as a comparison data reference value, 3
Reference numeral 2 denotes a deviation storage unit that stores an allowable value of the deviation between the comparison data output from the run length encoder 2 and the reference value stored in the comparison data storage unit. The data in the comparison data storage unit 31 and the deviation storage unit 32 can be arbitrarily set by the operator. The operation of the apparatus configured as described above will be described below.

Here, a case where the total number of black pixels in a binary image is used as the comparison data will be described as an example. The grayscale image to be captured is not an image in which a plurality of images are discretely scattered as shown in FIG. 8, but a monomodal image. An image output signal (grayscale image) output from the image input device 1 such as a television camera enters the run length encoder 2 and is binarized and encoded to compress information.

Specifically, the binarization control unit 22 compares the threshold value set in the threshold value setting unit 21 with the grayscale image output from the image input device 1, and the pixel having a higher density than the threshold value is compared. Is binarized to “1” (black) and pixels having a density lower than the threshold are binarized to “0” (white). FIG. 2 is a diagram showing an example of a binary image binarized in this way.

The binarization control unit 22 edits the binary image shown in FIG. 2 into a format as shown in FIG. 3 and outputs it.
In FIG. 2, the shaded area is the black pixel area. Each pixel can be specified by the coordinates in the X and Y directions. For example, the black pixel in the top row has a horizontal run length of 2, and the initial coordinates are X = 2 and Y = 3. The start position of the black pixel in the next stage is X = 3, Y = 2 and its length is 3
Is. When the same editing is performed thereafter, the format shown in FIG. 3 is obtained. In the run-length coding, since the start position of the black pixel and its length are used, the code can be compressed.

Data in the format shown in FIG. 3 enters the processor 3. The processor 3 adds the length information of the encoded data for each run to obtain the total number of black pixels in the binary image. In the case of FIG. 3, the total number of black pixels is 12. The processor 3 compares the comparison data 12 that is the output of the run length encoder 2 with the comparison data reference value stored in the comparison data storage unit 31, and the comparison data that is the output of the run length encoder 2 becomes the reference value. Run length encoder 2 to be equal
The threshold value of the threshold value setting unit 21 therein is sequentially changed.

Finally, the threshold when the comparison data output from the run length encoder 2 becomes equal to the reference value is binarized as a formal threshold. With such a configuration, it is possible to perform binarization that is not affected by changes in the surroundings even when the surroundings change. For example,
When the comparison data is set to 25 total pixels in a certain state,
It is assumed that the surrounding situation, for example, the brightness has changed. Even if the brightness changes, the total number of pixels does not change to 25. Therefore, even after the change, the threshold is automatically changed so that the total number of pixels becomes 25.

If the binarization threshold value is determined so that the output comparison data becomes equal to the reference value, the number of repetitions may increase and it may become impractical. In such a case, if the deviation between the output data and the reference value falls within a certain value, the binarization threshold value determination process may be stopped and the threshold value at that time may be set as the official threshold value. That is, the processor 3 obtains a deviation between the total number of input black pixels and the comparison data reference value, and compares the deviation with the reference deviation stored in the deviation storage unit 32. Then, when the deviation falls within the reference deviation, the processing is interrupted. By doing so, the image can be binarized in a state close to the target binary image, and the processing becomes faster.

FIGS. 4 to 6 are flow charts showing an automatic binarization decision sequence according to the present invention, and show the operation of the processor 3. First, to use that value later,
The operator sets the binarization convergence range (the ratio of the difference in the total number of black pixels for determining the convergence to the total number of black pixels in the target binary image. The deviation described above) (S1). ..
Next, it is checked whether there is an initial threshold assumption (S
2) In some cases, STEP 1 is omitted and skip to STEP 2.

The assumption of the initial threshold value means the following processing. The entire range of gray scale levels is equally divided including the maximum and minimum values, and the total number of black pixels in the binary image at those levels is stored in advance in the comparison data storage unit 31. Find the point that is closest to the total number of black pixels in the desired binary image. Then, the threshold of the level of the closest point is set as the initial threshold.

If there is no assumption of the initial threshold value, the image is taken in by the threshold value (sample level) of the evenly divided points and binarized to calculate the total number of black pixels (S3). Next, the deviation from the target total number of black pixels (reference value) is calculated (S4). Then, it is checked whether or not it is within the convergence range (permissible value) of the deviation stored in the deviation storage unit 32 (S5), and when the deviation is within the permissible value, the threshold value at that time is set as a new threshold value. , The process ends (S
6).

When the total black pixel number difference (deviation) is not within the convergence range, it is checked whether or not the total pixel number difference is the smallest so far (S7). If it is the minimum, the threshold value, the total number of black pixels, and the total black pixel number difference (deviation) at that time are updated and stored (S8). If it is not the minimum, S8 is skipped. Next, it is checked whether the image acquisition at each sample level is completed (S
9) If it is finished, the process proceeds to STEP2. If it is not finished, the process returns to S3 to continue the image capturing process.

In STEP 2, the level of (threshold value -1) obtained in STEP 1 is used as a new threshold value to capture an image,
Binarization and total black pixel number calculation are performed (S10). Next, the total number of black pixels obtained by the binarization process and the comparison data storage unit 3
The deviation (difference in the number of black pixels) from the reference value stored in 1 is calculated (S11). The total black pixel number difference thus obtained is compared with the allowable difference stored in the deviation storage unit 32 to check whether or not it is within the convergence range (S1).
2). If it is within the convergence range, the threshold value at that time is set as a new threshold value, and the process ends (S6).

If it is not within the convergence range, this time, contrary to S10, the level of (threshold value +1) obtained in STEP 1 is used as a new threshold value for image capture, binarization, and total black pixels. The number is calculated (S13). Next, the deviation (black pixel number difference) between the total number of black pixels obtained by the binarization process and the reference value stored in the comparison data storage unit 31 is calculated (S14). The total black pixel number difference thus obtained is compared with the tolerance stored in the deviation storage unit 32 to check whether it is within the convergence range (S15). If it is within the convergence range, the threshold value at that time is set as a new threshold value, and the process ends (S6).

If the total black pixel number difference is not within the convergence range, it is checked whether or not the total black pixel number difference obtained as a result of processing in STEP 2 is the smallest so far (S1).
7). Then, when the difference between the total number of black pixels is the smallest at the points before and after, the threshold change amount is set to +1 (S18) or -1 (S1).
9). That is, when the difference in the total number of black pixels is minimized when binarizing with (threshold value -1), the threshold change amount is set to
If the total black pixel number difference is minimized when the value is set to 1 and binarized by (threshold value + 1), the threshold value change amount is set to +1. If it does not become smaller than the difference in the number of black pixels calculated in STEP1, the threshold calculated in STEP1 is set as the official threshold, and the process ends (S6).
Although it is not within the convergence range, the threshold value at that time is determined to be the optimum threshold value. When the above processing is completed, the threshold value, the total number of black pixels, and the total black pixel number difference are updated (S20).

In step 3, the total number of black pixels is repeatedly calculated while adding the amount of change in the threshold value obtained in step 2. That is, while the level at which the total number of black pixels is minimized is found, the binary image is captured while repeatedly changing the threshold value, and the threshold value, total black pixel number, and total black pixel number difference at that time are continuously updated. If the minimum level is not found, it is determined that the level of the point before that point is the optimum binarization threshold.
During this process, if the total black pixel number difference (deviation) falls within the initially defined convergence range, the level at that point is set as the optimum binarization threshold.

Specifically, the amount of change in the threshold value obtained in STEP 2 is added to the threshold value setting section 21 (S21), the image is captured with the changed threshold value, and the total number of black pixels is calculated (S22). Next, the difference in the total number of black pixels of the target binary image is calculated (S
23). Next, it is checked whether or not the total black pixel number difference is within the convergence range (S24), and if so, the threshold value at that time is set as a formal threshold value and the process ends (S24).
6).

If the difference is not within the convergence range, it is checked whether the difference in the total number of black pixels is smaller than the smallest one so far (S25). The difference is updated (S26), and the process returns to S21. In this way, the threshold value is adjusted to the optimum value. S2
If it is not smaller than the total black pixel number difference up to now, since the previous setting value is more optimal, the threshold value at that time is set as a formal threshold value, and the process ends (S
6). By performing the above processing, the optimum binarization threshold value for obtaining the binary image closest to the target binary image state can be obtained.

In the above-described embodiment, the case where the total number of black pixels of the binary image is used as the comparison data has been described as an example, but the present invention is not limited to this. Instead, the sum of the number of mismatched pixels of the binary image may be used (however, in this case, only when the relative position of the target object on the screen of the camera is almost the same at each image capture). Therefore, it cannot be applied to object recognition, etc., nor can it be applied to a collection of dispersed images).

FIG. 7 is an explanatory diagram for obtaining the total number of mismatched pixels of a binary image. The number of non-matching pixels between the image A and the image B is two (X, Y) = (3, 5), (4,5). When the image A is a target binary image, the total sum "2" of the number of mismatched pixels is used as comparison data. When this method is used, if the relative positions of the two binary images to be compared with each other in the screen are substantially the same, compared to the case where the total number of pixels of the binary image is used as the comparison data, Accurate automatic binarization is possible.

In step 1 of FIG. 4, step 1
If you input the initial value of the threshold for performing the process of 2,
Since the processing of STEP1 is completely unnecessary, the time until the convergence to the optimum threshold value (the total processing time of STEP2 and STEP3) is shortened, which is more practical.

[0029]

As described above in detail, according to the present invention, the comparison data for indicating the characteristics of the binary image is set in advance as the reference value even when the surrounding situation changes. To provide an image processing device capable of automatically setting an optimum threshold value by varying a binarization threshold value so that comparison data of a binary image obtained by the binarization processing approaches a reference value. You can Further, according to the present invention, since the processing function can be realized by software,
The cost can be reduced.

Further, by comparing the state of the binarized image of the target image for automatic binarization as the comparison data, a histogram obtained by integrating the number of pixels for each gray level of the conventional gray level is used. Automatic 2 which was impossible by the method
The reference image state for binarization can be set arbitrarily. Furthermore, by using the present invention, it is possible to binarize an image without being affected by changes in the environment such as ambient brightness, so that after that, image measurement using a binary image, image processing, etc. It is convenient for performing various operations.

[Brief description of drawings]

FIG. 1 is a configuration block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a binarized binary image.

FIG. 3 is a diagram showing an output format example of a binary image.

FIG. 4 is a flowchart showing an automatic binarization determination sequence according to the present invention.

FIG. 5 is a flowchart showing an automatic binarization determination sequence according to the present invention.

FIG. 6 is a flowchart showing an automatic binarization determination sequence according to the present invention.

FIG. 7 is an explanatory diagram for obtaining the total number of mismatched pixels in a binary image.

FIG. 8 is a diagram showing an example of a grayscale image.

FIG. 9 is a diagram showing an example of a histogram of a gray level of a gray image.

[Explanation of symbols]

 1 Image Input Device 2 Run Length Encoder 3 Processor 21 Threshold Setting Unit 22 Binarization Control Unit 31 Comparison Data Storage Unit 32 Deviation Storage Unit

Claims (3)

[Claims] 1. An image input device for inputting a gray-scale image, and binarizes the output of the image input device by comparing it with a predetermined threshold value and binarizes coordinate information of the binary image and length data of black pixels. In an image processing apparatus including an output run-length encoder and a processor receiving the output of the run-length encoder, comparison data storing a characteristic state of a target binary image as a comparison data reference value in the processor. An image processing apparatus comprising a storage unit, wherein the processor sequentially changes the threshold value of the run-length encoder so that the value of the binarized comparison data matches a reference value. 2. The processor is provided with a deviation storage section for storing an allowable value of a deviation between the output comparison data of the run length encoder output and the reference value stored in the comparison data storage section, and binarized. The image processing apparatus according to claim 1, wherein when the deviation between the comparison data and the reference value is within the allowable value, the threshold value at that time is encoded as a formal threshold value. 3. The image processing apparatus according to claim 1, wherein the comparison data is a total number of black pixels or a total number of black pixels that do not match a target binary image.