JP2001067481A - Device and method for image retrieval, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image retrieval device which can retrieve a registered image fast with high precision by detecting the angle of rotation of an input image fast and correcting the angle of rotation of the input image. SOLUTION: The image retrieval device has an image input part 1 which inputs an image, a window data generation part 2 which generates window data by making a window scan of specific size on the input image, a projection data generation part 4 which generates projection data from the generated window data, a projection data comparison part 5 which compares the generated projection data with pieces of previously registered template data, an angle detection part 6 which detects the angle of rotation of the input image according to the comparison result of the projection data comparison part 5, and a matching degree decision part 8 which decides the degrees of matching between the input image and registered images by using the angle of rotation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image retrieval apparatus and an image retrieval method for retrieving whether or not a pre-registered registered image exists in an input image, and a recording medium for executing the method. is there.

[0002]

2. Description of the Related Art Conventionally, an image search method has been proposed as a method for searching for a pre-registered image in an image. Generally, an image search method is required to process a search for a registered image from input images in real time.

In searching for a registered image, matching of the registered image is performed by using all pixels in the input image as a pixel of interest.
It takes enormous processing time. Therefore, conventionally, as a method of obtaining the center point of the image to be matched with the registered image,
For example, a method of extracting an area of the size of one character and obtaining the center point from the width and height of the pixels in the extracted area, or a method in which the registered image has a circle and various patterns are only inside the circle Is drawn, various methods have been proposed, such as a method of extracting a circle based on the presence or absence of an edge at a plurality of points on the circumference and obtaining a center point of the circle.

However, when the input image is rotated, the input image is rotated by a predetermined angle, and each time the rotated image data is matched with the registered image, and the registered image is registered based on the matching ratio of all angles. Retrieval processing becomes complicated, such as determining whether an image is an image, and it becomes difficult to search registered images at high speed.

[0005]

As described above, the above-described conventional image search method has a problem that it takes an enormous amount of processing time to search for a registered image, making it difficult to perform real-time processing. Was.

In this image search apparatus, image search method, and recording medium, the number of pixels in each column is counted in the same direction, and the counted number is compared with template data for each column of each angle registered in advance. It is required that the rotation angle of the input image is detected, the rotation angle of the input image is corrected, and the search for the registered image is performed at high speed and with high accuracy.

According to the present invention, there is provided an image retrieval apparatus capable of detecting a rotation angle of an input image at a high speed, correcting the rotation angle of the input image, and performing a high-speed and high-accuracy search of a registered image.
And an image search method for detecting the rotation angle of the input image at high speed, correcting the rotation angle of the input image, and performing high-speed and high-accuracy search of the registered image, and executing the image search method. The purpose of the present invention is to provide a recording medium.

[0008]

In order to solve the above-mentioned problems, an image search apparatus according to the present invention is an image search apparatus for searching whether a registered image registered in advance exists in an input image. There, an image input unit for inputting an image, a window data generating unit for performing window scanning of a predetermined size on the input image to generate window data, a projection data generating unit for generating projection data from the generated window data, A projection data comparison unit that compares the created projection data with a plurality of template data registered in advance; an angle detection unit that detects a rotation angle of the input image based on a comparison result in the projection data comparison unit; A coincidence determining unit that determines the degree of coincidence between the input image and the registered image using the input image and the registered image.

Thus, an image retrieval apparatus capable of detecting the rotation angle of the input image at a high speed, correcting the rotation angle of the input image, and searching for the registered image at a high speed and with a high accuracy can be obtained.

In order to solve the above problem, an image search method according to the present invention is an image search method for searching whether or not a registered image registered in advance exists in an input image. A window data generating step of performing window scanning of a predetermined size to generate window data, a projection data generating step of generating projection data from the generated window data, and a plurality of template data registered in advance with the generated projection data A projection data comparison step of comparing the input image with the registered image; a detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step; and determining a degree of coincidence between the input image and the registered image using the rotation angle. And a matching degree determining step.

Thus, an image search method for detecting the rotation angle of the input image at high speed, correcting the rotation angle of the input image, and searching for the registered image at high speed and with high accuracy can be obtained.

In order to solve the above-mentioned problem, a recording medium according to the present invention is an image search method for searching whether a registered image registered in advance exists in an input image. A window data generating step of generating window data by performing window scanning of a size, a projection data generating step of generating projection data from the generated window data, and a plurality of template data registered in advance with the generated projection data. A projection data comparison step for comparing the input image, an angle detection step for detecting a rotation angle of the input image based on the comparison result in the projection data comparison step, and a match for determining the degree of coincidence between the input image and the registered image using the rotation angle. A degree determination step, and a configuration for recording a program for executing the step.

As a result, a recording medium for executing the above-described image search method is obtained.

[0014]

DETAILED DESCRIPTION OF THE INVENTION An image retrieval apparatus according to a first aspect of the present invention is an image retrieval apparatus for performing a search as to whether or not a registered image registered in advance exists in an input image. An image input unit for inputting an image, a window data generation unit for performing window scanning of a predetermined size on the input image to generate window data, a projection data generation unit for generating projection data from the generated window data, and a generated projection A projection data comparison unit for comparing data with a plurality of template data registered in advance, an angle detection unit for detecting a rotation angle of the input image based on a comparison result in the projection data comparison unit, and an input using the rotation angle And a coincidence determining unit for determining the degree of coincidence between the image and the registered image.

According to this configuration, since the rotation angle of the input image is detected based on a comparison between the projection data and the plurality of template data, the detection of the rotation angle of the input image is speeded up, and based on the detected rotation angle. By performing the correction of the rotation angle of the input image, it is possible to perform a search for a registered image in the input image at high speed and with high accuracy.

According to a second aspect of the present invention, there is provided an image search method for searching whether a registered image registered in advance exists in an input image. A window data generating step of generating window data by performing scanning, a projection data generating step of generating projection data from the generated window data, and comparing the generated projection data with a plurality of template data registered in advance. Projection data comparison step to be performed, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and a coincidence determination step of determining a coincidence between the input image and the registered image using the rotation angle And

With this configuration, since the rotation angle of the input image is detected based on a comparison between the projection data and the plurality of template data, the detection of the rotation angle of the input image is speeded up, and based on the detected rotation angle. By performing the correction of the rotation angle of the input image, it is possible to perform a search for a registered image in the input image at high speed and with high accuracy.

According to a third aspect of the present invention, there is provided an image search method for searching whether or not a registered image registered in advance exists in an input image. A window data generating step of performing scanning to generate window data; a binarizing processing step of performing binarization on the generated window data to generate binarized window data;
A projection data creation step of creating projection data from the quantified window data, a projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and a comparison result in the projection data comparison step. The method includes an angle detection step of detecting a rotation angle of the input image, and a coincidence determination step of determining a coincidence between the input image and the registered image using the rotation angle.

With this configuration, since the projection data is created from the binarized window data, the rotation angle of the input image can be detected at a higher speed.

According to a fourth aspect of the present invention, there is provided an image search method for searching whether or not a registered image registered in advance exists in an input image. A window data generating step of performing scanning to generate window data, a multi-level processing step of performing multi-level processing on the generated window data to generate multi-level window data, and a generated multi-level window data A projection data creation step of creating projection data from the input image, a projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and a rotation of the input image based on a comparison result in the projection data comparison step. An angle detecting step of detecting an angle; and determining a degree of coincidence between the input image and the registered image using the rotation angle. A coincidence determination step, in which it was decided to have a.

According to this configuration, since projection data is created from the multi-valued window data, there is an effect that the detection of the rotation angle of the input image is further improved in accuracy.

According to a fifth aspect of the present invention, there is provided an image search method according to any one of the second to fourth aspects.
In the projection data comparison step, the angle of the template data is set for each angle smaller than 6 degrees.

According to this configuration, there is an effect that the accuracy of the template data is improved and the detection of the rotation angle of the input image is further improved.

According to a sixth aspect of the present invention, there is provided the image search method according to any one of the second to fifth aspects, wherein:
In the projection data comparison step, when the difference between the created projection data and the template data is equal to or larger than a predetermined value, the process shifts to comparison with the next template data.

With this configuration, the comparison of the input image with template data that is unlikely to be matched is omitted, so that the rotation angle of the input image can be detected at a higher speed.

According to a seventh aspect of the present invention, there is provided the image search method according to any one of the second to sixth aspects, wherein:
In the projection data creation step, pixel values are counted in a plurality of directions.

According to this configuration, an error that occurs only with the coefficient of the pixel value in one direction is prevented, so that the rotation angle of the input image can be detected more accurately.

According to an eighth aspect of the present invention, in the image search method according to the seventh aspect, in the projection data comparing step, a comparison result between the projection data in one direction and the template data is used in the other direction. This is to control the comparison result between the projection data and the template data.

With this configuration, it is possible to narrow the search angle in the other direction, so that the rotation angle of the input image can be detected at a higher speed.

According to a ninth aspect of the present invention, there is provided an image search method according to any one of the second to eighth aspects, wherein
In the projection data comparison step, the comparison is performed using any of the maximum number of pixels, the minimum number of pixels, the standard deviation, and the average number of pixels.

According to this configuration, it is possible to quickly determine that the image is a non-registered image without obtaining the degree of coincidence of all columns, and it is possible to narrow down the angle of the registered image at a high speed.

According to a tenth aspect of the present invention, there is provided a recording medium for executing an image search method for searching whether a registered image registered in advance exists in an input image,
A window data generating step of performing window scanning of a predetermined size on an input image to generate window data, a projection data generating step of generating projection data from the generated window data, and a plurality of pre-registered generated projection data. A projection data comparison step of comparing the input image with the template data of the input image, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and a degree of coincidence between the input image and the registered image using the rotation angle. And a program for executing the matching degree determination step to be determined.

According to this configuration, there is an effect that an image search method capable of performing a high-speed and high-accuracy search for a registered image in an input image can be performed at an arbitrary place and period.

Hereinafter, an embodiment of the present invention will be described with reference to FIGS.
This will be described with reference to FIG.

(Embodiment 1) FIG. 1 is a block diagram showing an image retrieval apparatus according to Embodiment 1 of the present invention.

In FIG. 1, reference numeral 1 denotes an image input unit for inputting image data, and 2 denotes window data of N pixels × N pixels centering on a certain pixel of interest with respect to the image data input by the image input unit 1. The window data generating unit 3 generates a binarization process of the window data generated by the window data generating unit 2 (for example, a binary value obtained by calculating the brightness of image data and comparing the brightness with a predetermined threshold value). Binarization processing unit 4 which performs binarization processing) projects the window data binarized by the binarization processing unit 3 by counting the number of pixels for each column in the same direction (for example, the vertical direction). A projection data creation unit 5 for creating data is a projection data creation unit for comparing the projection data created by the projection data creation unit 4 with template data of each angle registered in advance. Data comparison unit, 6 the projection angle detecting section for detecting an angle as a result of comparison by the data comparing unit 5, 7 is in accordance with the detected angle by the angle detecting section 6,
An angle correction unit 8 corrects the angle of the window data binarized by the binarization processing unit 3. The angle correction unit 8 matches the window data after the angle correction corrected by the angle correction unit 7 with a registered image registered in advance. It is a coincidence determination unit that determines the degree.

Here, the configuration of FIG. 1 will be described with reference to FIG. FIG. 3 is an image diagram showing a registered image to be detected. In FIG. 1, the window data generation unit 2 and the binarization processing unit 3 are exchanged, and the image data input by the image input unit 1 is binarized by the binarization processing unit 3 to perform binarization processing. With respect to the image data binarized by the unit 3, the window data generation unit 2 may generate window data of N pixels × N pixels around a certain pixel of interest. Further, without performing the binarization processing, the projection data creation unit 4 obtains the projection data by calculating the sum of the pixel values for each column in the same direction with respect to the multivalued window data (multivalued window data). May be created. The target pixel may be all pixels of the image data input by the image input unit 1, but here, as shown in FIG. 3, a registered image 9 having a circle and having a pattern only inside the circle.
In the following description, the center pixel of a circle is detected, and the detected pixel is set as a target pixel. Further, the angle correction unit 7 can rotate the registered image, register a registered image of each angle, and select a registered image according to the detected angle. Although the degree of coincidence between the window data and the registered image is determined here, a second window different from the size of the window data for which the projection data is created is generated, and the degree of coincidence with the registered image is determined. You may.

Next, the operation of the image retrieval apparatus configured as described above will be described with reference to FIGS. 2, 4 to 7. FIG. 2 is a flowchart showing the operation of the image search apparatus of FIG. 1, FIG. 4 is an angle template correspondence diagram showing template data of each angle with respect to a registered image, and FIG. 5 shows an image in a window of a predetermined size with respect to an input image. FIG. 6 is a data diagram showing projection data for the image of FIG. 5, and FIG. 7 is an image diagram showing an image after correcting the angle of the image of FIG.

First, image data is input by the image input unit 1 (S1), and a pixel of interest is extracted by detecting the center pixel of a circle in the image data input by the image input unit 1. The method of detecting the center pixel of the circle will be described later in detail. Next, the window data generation unit 2 generates window data of a predetermined size (for example, 25 pixels × 25 pixels) centering on the extracted pixel of interest (S
2. window data generation step), for the window data generated by the window data generation unit 2,
The binarization processing unit 3 performs binarization processing (for example, binarization processing by obtaining the brightness of a color image signal and comparing the brightness with a predetermined threshold) (S3, binarization processing step) ). Here, in the binarized window data (binarized window data) 11 (FIG. 5), first, the number of pixels in the first column in the vertical direction is counted by the projection data creation unit 4, and then the number of pixels in the vertical direction is counted. The number of pixels in the second column is counted, and similarly, the number of pixels in all columns in the vertical direction in the window is counted for each column (S4, S4).
5. Projection data creation step). The count value for each column in the vertical direction obtained as described above is used as projection data 12 (FIG. 6), and template data 1 for each angle registered in advance is used.
0 (FIG. 4). Details of the method of creating the template data 10 and the method of comparison will be described later. Here, the projection data comparison unit 5 compares the obtained projection data 12 with the template data 10B-0 of the angle of 0 degrees registered in advance (S6, projection data comparison step). Now, template data 1 of 0 degree angle
Since only 0B-0 is compared, an angle of 0 degree is detected as the rotation angle of the input image (S7). Similarly, template data 10B-0 to 10B of all angles
Compare with -350 (S8). The angle detection unit 6 determines, based on a comparison result between the projection data 12 and the template data 10 at an angle of 0 ° to 350 °, the angle of the template data having the smallest value of the comparison result (the most similar to the projection data 12). , The projection data 12 and the 70-degree template data 10B-70 are the most similar.)
Is detected (S7, angle detection step). Through the above steps S6 to S8, the inclination angle of the window data 11 (the rotation angle of the input image) is detected. Here, the window data 11 is converted by the angle correction unit 7 into step S7.
The angle is corrected by the angle (70 degrees in this case) detected by (1) (S9). Details of the angle correction method will be described later. Next, the degree of coincidence between the angle-corrected image 13 (FIG. 7) obtained by correcting the angle of the window data 11 and the registered image 9 (for example, the Matching by matching ratio of white pixels)
Is determined (S10, matching degree determination step), and it is determined whether or not the image matches the registered image 9. A series of processes from step S2 to step S10 is completed for a certain target pixel, and if there is a next target pixel (S11), the target pixel is moved (S12). The processing up to step S10 is repeated. Similarly, for all pixels, steps S2 to S1
When the processing up to 0 is performed and the processing is completed for all the target pixels (S11), the search processing ends.

The details of the method of creating the template data 10 will be described with reference to FIG. The image 10A-0 is an image obtained when the registered image 9 to be detected is rotated at an angle of 0 degrees (an image obtained when the registered image 9 is not rotated at all), and the image 10A-70 is an image obtained by rotating the registered image 9 to be detected at an angle of 70 degrees. Similarly, images 10A-350 are images obtained by rotating registered image 9 to be detected by an angle of 350 degrees. The template data 10B-0 is projection data obtained by counting the number of pixels for each column in the vertical direction with respect to the image 10A-0.
The template data 10B-70 includes the image 10A-70
The template data 10B-350 is projection data obtained by counting the number of pixels in each column in the vertical direction with respect to the image 10A-350. Here, the template data is created at an angle of 10 degrees. However, for example, by creating the template data at an angle of 1 degree, the inclination of the input image can be detected with higher accuracy, and matching can be performed with higher accuracy. Can be.

Projection data 12 and template data 10
The method of comparison with will be described with reference to FIGS.
The value of the i-th column (i = 1, 2,..., N) of the projection data 12 is Si, and the i-th column of the template data 10B-k of the angle k degrees (k = 0, 10,..., 350) The value of Tki
Then, a comparison result Rk between the projection data 12 and the template data 10B-k of the angle k is obtained as shown in Expression (1). Similarly, a comparison result Rk between the template data of another angle and the projection data is obtained, and an angle k corresponding to the template data having the smallest comparison result Rk is obtained.

Rk = Σ | Si−Tki | (1) Actually, the projection data 12 with respect to the window data 11 of the input image in FIG. 5 and the template data of each angle in FIG. The comparison result R60 with the template data of 60 degrees is 46, the comparison result R70 with the template data of 70 degrees is 20, the comparison result R80 with the template data of 80 degrees is 30, and the other 0 to 35 degrees
The result of comparison with the template data of 0 degrees is larger than the result of comparison of 70 degrees, and 70 degrees is obtained by the equation (1). As described above, in the comparison between the template data of each angle and the input image, the angle closest to the angle of the input image can be detected even if the angle between the template data of each angle is created every 10 degrees. In order to match the relative angle between the image and the registered image and perform the matching with high accuracy, the minimum angle between the template data of each angle is 6
Degree or less.

Next, a method for detecting the center pixel of a circle will be described with reference to FIG. FIG. 8 is an explanatory diagram of a method for detecting the center pixel of a circle. 8, the circle 14 is a part of the registered image, the target pixel 15 is the center pixel of the circle 14, and the pixels 16A-0 to 16A-11 on the circumference are the pixels existing on the circumference of the circle 14. And the pixel 16B-0 on the circumference
16B-11 are pixels 16A- on the circumference outside the circle 14.
Pixels adjacent to 0-16A-11. Pixels 16A-j and 16B-j on the circumference (j = 0, 1,..., 11)
It is assumed that an edge exists at the boundary between the inside of the circle and the outside of the circle when Expression (2) is satisfied.

The value of the pixel 16A-j <the value of the pixel 16B-j (2) Further, the lower limit value of the characteristic color of the registered image is C0 (0 as an example) and the upper limit value is C1 (50 as an example). Then, when the value of the pixel 16A-j on the circumference satisfies the expression (3), it is assumed that the color of the pixel 16A-j on the circumference matches the registered image.

C0 (0) ≦ value of pixel 16A-j ≦ C1 (50) (3) In all of a plurality of positions (here, 12 positions) on the circumference, the expression (2) and (3) Satisfies the expression, the target pixel 1
5 is the center pixel of the circle 14.

In the input image, the pixel of interest is scanned over the entire image in pixel units, and the surrounding pixels 16A-
The processing of Expressions (2) and (3) is performed on j and 16B-j (j = 0, 1,..., 11), and it is determined whether or not the target pixel is the center pixel of the circle.

The angle correction method will be described. With the pixel of interest determined as the center pixel, angle correction is performed by the angle at which all black pixels in the binarized window are detected. For example, if the detected angle is 70 degrees, all black pixels are rotated around the center pixel by -70 degrees.

In this embodiment, the projection data 12
Is obtained by counting the number of pixels in the vertical direction, but may be horizontal or oblique as long as they are in the same direction. Also, the directions may be plural. This will be described with reference to FIGS. 9 is an image diagram showing a registered image of right and left targets, FIG. 10 is an image diagram showing an image at an angle of 0 degree, and FIG.
1 is a data diagram showing template data for the image of FIG. 10, FIG. 12 is an image diagram showing an image at an angle of 180 degrees, and FIG.
FIG. 3 is a data diagram showing template data for the image of FIG. Here, for example, when projection data in only one direction is used, and when a symmetric image 17 as shown in FIG. 9 is registered as a registered image, template data 19 for an image 18 having an angle of 0 degrees as shown in FIG. And the template data 21 for the image 20 at an angle of 180 degrees as shown in FIG. 12 becomes equal, making it difficult to detect the angle correctly. Therefore, if projection data is created by counting the number of pixels for each column in a plurality of directions (for example, both the horizontal direction and the vertical direction), even if the image is symmetrical, the inclination of the input image can be accurately detected. can do. At this time, the projection data in the horizontal direction is equal to the projection data in the vertical direction when the input image is rotated by 90 degrees. Therefore, when comparing the projection data in the vertical direction with the template data at the angle K degrees, Projection data is angle K
What is necessary is just to compare with +90 degree template data.

As described above, according to the present embodiment, the image input unit 1 for inputting an image, the window data generating unit 2 for performing window scanning of a predetermined size on the input image to generate window data, and Binarize window data to generate binarized window data 2
A value processing unit 3, a projection data creation unit 4 for creating projection data from the generated binarized window data, and a projection data comparison for comparing the created projection data with a plurality of template data registered in advance. Unit 5, an angle detection unit 6 for detecting the rotation angle of the input image based on the comparison result in the projection data comparison unit 5, an angle correction unit 7 for correcting the rotation angle of the input image using the detected rotation angle, and angle correction Providing the coincidence determining unit 8 for determining the degree of coincidence between the input image and the registered image to detect the rotation angle of the input image based on a comparison between the projection data 12 and the plurality of template data 10 The rotation angle of the input image can be detected at a high speed, and a search for a registered image in the input image whose angle has been corrected based on the detected rotation angle can be performed. It can be performed quickly and accurately.
Further, if the binarization processing unit is a multi-level processing unit, projection data can be created from the multi-level window data, so that the rotation angle of the input image can be detected with higher accuracy. Furthermore, by setting the angle of the template data for each angle smaller than 6 degrees,
Since the accuracy of the template data is improved, the rotation angle of the input image can be detected with higher accuracy. Furthermore, since the pixel value is counted in a plurality of directions, it is possible to prevent an error that may occur only with the coefficient of the pixel value in one direction, so that the rotation angle of the input image can be more accurately determined. Can be detected. Further, the comparison result between the projection data in one direction and the template data is controlled based on the comparison result between the projection data in one direction and the template data.
It is possible to prevent an error that may occur only with the coefficient of the pixel value in the direction, so that the rotation angle of the input image can be detected more accurately. Furthermore, the comparison is performed using any of the maximum number of pixels, the minimum number of pixels, the standard deviation, and the average number of pixels. It is possible to make a determination and to narrow down the angle of the registered image at high speed.

(Embodiment 2) Hereinafter, Embodiment 2 of the present invention will be described with reference to FIGS.
FIGS. 14 and 15 are flowcharts showing an image search method according to the second embodiment of the present invention. The configuration of the image search device according to the present embodiment is the same as that of FIG. 1, and the operation is different from that of the first embodiment.

Referring to FIGS. 14 and 15, first, image data is input (S21).
A predetermined size (for example, 25 pixels ×
Window data (25 pixels) is created (S11, window data generation step), and the created window data is binarized (for example, the brightness of a color image signal is obtained, and the brightness is compared with a predetermined threshold value). (S23, binarization processing step). Next, the template data to be read is moved to the template data of the first angle (S24). Next, it is determined whether all the template data have been read and compared (S25). Since not all the template data has been read out at first, the first column of the template data and the binarized window data (binary window data) is set as a target column (S2
6) The number of pixels "S" in the first column of the binarized window data is counted (S27). At this time, the value "T" in the first column of the template data is read (S28). Here, the number of pixels “S” in the first column of the window data is compared with the value “T” in the first column of the template data, and the difference is within a certain threshold (here, the threshold is 5) (S29), it is determined that S and T are similar, then it is determined whether or not the processing has been performed for all columns (S30), and at first, all the columns in the window data are determined. Since the columns are not compared, the process moves to the next column (S31). Similarly, the number of pixels "S" in the i-th column of the window data is counted (S27), and the value "T" in the i-th column of the template data is read (S27).
28), if S and T are similar (S29), the process for the next column is further performed, all columns in the window are compared (S30), and S and T are similar in all columns. If so, the angle corresponding to the template data at that time and the sum of the comparison results of all the columns are extracted (S32). Further, S and T are compared for each column, and S
When the difference between T and T exceeds the threshold value (5) (S29), it is determined that the template data is completely different from the window data, and the template data at the next angle is compared (S33). Similarly, the comparison between the template data of all angles and the window data is performed by the number of pixels for each column, and the angles corresponding to the template data similar in all columns and the sum of the comparison results of all columns are obtained (S32). ).
When the comparison between the template data of all angles and the window data is completed (S25, projection data comparison step), the value becomes the smallest in the sum of all the comparison results obtained in step S32 (similar to the window data). An angle corresponding to the template data is extracted (S34, angle detection step). The angle of the window data is corrected in accordance with the angle obtained in step S34 (S35), the data after the angle correction is matched with the registered image (S36, matching degree determination step), and the input image is compared with the registered image. A judgment is made as to whether or not they have been performed. Step S22 to step S36 for a certain pixel of interest
When a series of processes up to and including the next pixel of interest is present (S37), the process moves to the pixel of interest (S38), and the processes from step S22 to step S36 are repeated for the next pixel of interest. Step S is performed for all pixels in the same manner.
The processing from 22 to step S36 is performed, and when the processing is completed for all the target pixels (S37), the search processing ends. Steps S24 to S31 constitute a projection data creation step.

In the present embodiment, the projection data is obtained by counting the number of pixels in the vertical direction. However, the projection data may be in the horizontal direction or in the oblique direction as long as they are in the same direction. Also, the directions may be plural. For example, when projection data in only one direction is used, and when a symmetric image 17 as shown in FIG. 9 is registered as a registered image, the template data 19 for the image 18 having an angle of 0 degrees in FIG. Template data 21 for 180 degree image 20
Are equal, and it is difficult to detect the angle correctly. Therefore, if projection data is created by counting the number of pixels for each column in a plurality of directions (for example, both the horizontal direction and the vertical direction), even if the image is symmetrical, the inclination of the input image can be accurately detected. can do. At this time, the horizontal projection data is equal to the vertical projection data when the input image is rotated by an angle of 90 degrees. Therefore, when comparing the vertical projection data with the template data at an angle of K degrees, The direction projection data is angle K + 9
What is necessary is just to compare with the template data of 0 degree.

Further, in the present embodiment, the comparison between the projection data and the template data is performed for each column. However, the comparison is performed for each of a plurality of columns, and the template data most similar to the projection data is obtained. Is also good. If projection data is created by counting pixel values in a plurality of directions, the projection data in one direction is compared with the projection data in a certain direction by comparing the projection data in one direction with the template data by performing a comparison for each angle. And the template data can be controlled. For example, the angle is narrowed down by extracting the top five template data with a smaller comparison result using the vertical projection data, and then using the horizontal projection data, By performing the comparison processing for the angle of, the processing can be sped up.

The standard deviation, the total number of pixels, and the average number of pixels per column (= the total number of pixels / N columns) in the window data do not depend on the rotation angle. The number of pixels is obtained and compared with the total number of pixels of the registered image. When Expression (4) is satisfied, it is assumed that the window data matches the total number of pixels of the registered image, and the projection data comparing unit 5 and the angle detecting unit 6 The respective processes of the angle correction unit 7 and the coincidence determination unit 8 are performed to determine whether or not they match the registered image.

H-TH <U <H + TH (4) In the equation (4), H is the total number of pixels of the registered image.
Is the total number of pixels in the window data, and TH is a threshold value (here, 50). When the comparison of the number of pixels does not satisfy the expression (4), the processing of the projection data comparison unit 5, the angle detection unit 6, the angle correction unit 7, and the coincidence determination unit 8 is not performed, and it is determined that the image does not match the registered image. judge.

Although the comparison of the above equation (4) uses the total number of pixels, any of the standard deviation and the average number of pixels may be used as described above. As described above, when an image different from the registered image is input, the processing can be speeded up without performing angle correction or matching.

Further, in the projection data comparing section, a column in which the maximum number of pixels (or the minimum number of pixels) of the binarized window data exists is determined. Speeding up the template data comparison process by extracting template data having a number (or the minimum number of pixels) and comparing only the extracted template data with all the columns of projection data and the template data; Can be achieved.

As described above, in the present embodiment, by using the maximum number of pixels (or the minimum number of pixels), the angle can be detected at high speed without comparison with the template data of all angles. .

(Embodiment 3) FIG. 16 is a block diagram showing an image retrieval apparatus according to Embodiment 3 of the present invention. In FIG. 16, 31 is a CPU, 32 is a display, 3
3 is a memory, 34 is a scanner, 35 is a hard disk,
36 is a floppy disk drive, 37 is an MO disk drive, and 38 is a CD-ROM disk drive.

The computer-readable recording medium such as the floppy disk drive 36, MO disk drive 37 or CD-ROM disk drive 38 realizes the image search method using the projection data described in the first and second embodiments. A program is recorded, and each configuration 1 to 8 in FIG. 1 is realized by the CPU 31 executing the program. That is, an image is input from the scanner 34 or the like, and is temporarily stored in the hard disk 35 or the like. Here, when the program is executed, the stored image data is read, an image search process using the projection data is executed, and the result is output to the display 32 or the like.

As described above, according to the present embodiment, a window data generating step of performing window scanning of a predetermined size on an input image to generate window data, and a projection data generating step of generating projection data from the generated window data Step, a projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and an angle detection step of detecting the rotation angle of the input image based on the comparison result in the projection data comparison step, A degree of coincidence determining step of determining a degree of coincidence between the input image and the registered image using the rotation angle; and a program for executing the same. It is possible to perform an image search method that can be performed with high accuracy at any place and period. .

[0062]

As described above, according to the image search apparatus of the first aspect of the present invention, an image search for searching whether or not a registered image registered in advance exists in an input image. An image input unit for inputting an image, a window data generating unit for performing window scanning of a predetermined size on the input image to generate window data, and a projection data generating unit for generating projection data from the generated window data A projection data comparison unit that compares the created projection data with a plurality of template data registered in advance; an angle detection unit that detects a rotation angle of the input image based on a comparison result in the projection data comparison unit; A coincidence determination unit that determines the degree of coincidence between the input image and the registered image using the angle, and the projection data and the plurality of template data Since the rotation angle of the input image is detected based on the comparison of the input image, the rotation angle of the input image can be detected at a high speed, and the registered image is included in the input image whose rotation angle is corrected based on the detected rotation angle. An advantageous effect is obtained that the search for presence or absence can be performed at high speed and with high accuracy.

According to the image search method of the second aspect,
An image search method for searching whether a registered image registered in advance exists in an input image, wherein a window data generating step of performing window scanning of a predetermined size on the input image to generate window data. A projection data creation step of creating projection data from the generated window data, a projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and a comparison result in the projection data comparison step An angle detection step of detecting a rotation angle of the input image, and a coincidence degree determination step of determining a degree of coincidence between the input image and the registered image using the rotation angle. The rotation angle of the input image is detected based on the comparison with It is advantageous that the speed of the degree detection can be increased, and the search for the registered image in the input image whose rotation angle has been corrected based on the detected rotation angle can be performed at high speed and with high accuracy. The effect is obtained.

According to the image search method of the third aspect,
An image search method for searching whether a registered image registered in advance exists in an input image, wherein a window data generating step of performing window scanning of a predetermined size on the input image to generate window data. A binarization processing step of performing binarization on the generated window data to generate binarized window data; a projection data generating step of generating projection data from the generated binarized window data; A projection data comparison step of comparing the projection data with a plurality of template data registered in advance, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and using the rotation angle. A matching degree determining step of determining a matching degree between the input image and the registered image, Since creating a projection data from the binarized window data, is advantageous effect that the detection of the rotation angle of the input image can be further performed at a high speed is obtained.

According to the image search method of the fourth aspect,
An image search method for searching whether a registered image registered in advance exists in an input image, wherein a window data generating step of performing window scanning of a predetermined size on the input image to generate window data. A multi-level processing step of performing multi-level processing on the generated window data to generate multi-level window data; a projection data generating step of generating projection data from the generated multi-level window data; A projection data comparison step of comparing the projection data with a plurality of template data registered in advance, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and using the rotation angle. A matching degree determining step of determining a matching degree between the input image and the registered image, Since creating a projection data from the binarized window data, is advantageous effect that it is possible to perform more accurate detection of the rotational angle of the input image is obtained.

According to the image search method of the fifth aspect,
In the image search method according to any one of claims 2 to 4, in the projection data comparison step, the template data is set at an angle smaller than 6 degrees, thereby improving the accuracy of the template data. Therefore, an advantageous effect that the rotation angle of the input image can be detected with higher accuracy can be obtained.

According to the image search method of the sixth aspect,
In the image search method according to any one of claims 2 to 5, in the projection data comparison step, when a difference between the created projection data and the template data is equal to or larger than a predetermined value, the comparison with the next template data is performed. By shifting, the comparison with the template data that is unlikely to match the input image is omitted, and the advantageous effect that the detection of the rotation angle of the input image can be further accelerated is obtained.

According to the image search method of the present invention,
In the image search method according to any one of claims 2 to 6, in the projection data creation step, by counting pixel values in a plurality of directions, there is a possibility that only the coefficient of the pixel value in one direction will occur. Since a certain error is prevented, the advantageous effect that the rotation angle of the input image can be detected more accurately is obtained.

According to the image search method of the eighth aspect,
8. The image search method according to claim 7, wherein in the projection data comparison step, the comparison result between the projection data in one direction and the template data is controlled based on the comparison result between the projection data in one direction and the template data. Since the search angle in the other direction can be narrowed down, an advantageous effect that the rotation angle of the input image can be searched at higher speed can be obtained.

According to the image search method of the ninth aspect,
In the image search method according to any one of claims 2 to 8, in the projection data comparison step, the comparison is performed by using any of the maximum number of pixels, the minimum number of pixels, the standard deviation, and the average number of pixels. It is possible to determine the non-registered image at high speed without obtaining the degree of coincidence of all columns, and it is possible to obtain an advantageous effect that the angle of the registered image can be narrowed down at high speed.

According to the recording medium of the tenth aspect, there is provided a recording medium for executing an image search method for searching whether or not a registered image registered in advance exists in an input image. A window data generating step of performing window scanning of a predetermined size on an image to generate window data; a projection data generating step of generating projection data from the generated window data; and a plurality of pre-registered projection data and the generated projection data. A projection data comparison step for comparing with the template data; an angle detection step for detecting a rotation angle of the input image based on a comparison result in the projection data comparison step; and determining a degree of coincidence between the input image and the registered image using the rotation angle And a program for executing the matching degree determination step. Advantageous effect that the search of whether there is a registration image can perform image retrieval method capable of performing fast and accurate anywhere-periods are obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an image search device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the image search device in FIG. 1;

FIG. 3 is an image diagram showing a registered image to be detected.

FIG. 4 is an angle template correspondence diagram showing template data of each angle with respect to a registered image.

FIG. 5 is an image diagram showing an image in a window of a predetermined size with respect to the input image.

FIG. 6 is a data diagram showing projection data for the image of FIG. 5;

FIG. 7 is an image diagram showing an image after correcting the angle of the image in FIG. 5;

FIG. 8 is an explanatory diagram of a method of detecting a center pixel of a circle.

FIG. 9 is an image diagram showing a registered image of right and left targets.

FIG. 10 is an image diagram showing an image at an angle of 0 degree.

FIG. 11 is a data diagram showing template data for the image of FIG. 10;

FIG. 12 is an image diagram showing an image at an angle of 180 degrees.

FIG. 13 is a data diagram showing template data for the image of FIG. 12;

FIG. 14 is a flowchart showing an image search method according to the second embodiment of the present invention.

FIG. 15 is a flowchart showing an image search method according to the second embodiment of the present invention.

FIG. 16 is a block diagram showing an image search device according to a third embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input part 2 Window data generation part 3 Binarization processing part 4 Projection data creation part 5 Projection data comparison part 6 Angle detection part 7 Angle correction part 8 Matching degree judgment part 31 CPU 32 Display 33 Memory 34 Scanner 35 Hard disk 36 Floppy Disk drive 37 MO disk drive 38 CD-ROM disk drive

Claims (10)

[Claims] 1. An image search apparatus for searching whether a registered image registered in advance exists in an input image,
An image input unit that inputs an image, a window data generation unit that performs window scanning of a predetermined size on the input image to generate window data, a projection data creation unit that creates projection data from the generated window data, A projection data comparison unit that compares the created projection data with a plurality of template data registered in advance, an angle detection unit that detects a rotation angle of the input image based on a comparison result in the projection data comparison unit, An image search device, comprising: a coincidence determining unit that determines a degree of coincidence between the input image and the registered image using a rotation angle. 2. An image search method for searching whether a registered image registered in advance exists in an input image,
A window data generating step of performing window scanning of a predetermined size on the input image to generate window data, a projection data generating step of generating projection data from the generated window data, and registering the generated projection data in advance. A projection data comparison step of comparing the input image with a plurality of template data, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and the input image using the rotation angle. A matching degree determining step of determining a matching degree with the registered image. 3. An image search method for searching whether a registered image registered in advance exists in an input image,
A window data generating step of performing window scanning of a predetermined size on the input image to generate window data;
A binarization processing step of performing binarization to generate binarized window data; a projection data creating step of creating projection data from the generated binarized window data;
A projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, A matching degree determining step of determining a matching degree between the input image and the registered image using the rotation angle. 4. An image search method for searching whether a registered image registered in advance exists in an input image,
A window data generating step of performing window scanning of a predetermined size on the input image to generate window data; and a multi-level processing step of performing multi-level processing on the generated window data to generate multi-level window data Projecting data creating step of creating projection data from the generated multi-valued window data,
A projection data comparison step of comparing the created projection data with a plurality of template data registered in advance, and an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, A matching degree determining step of determining a matching degree between the input image and the registered image using the rotation angle. 5. The projection data comparing step, wherein an angle of the template data is set for each angle smaller than 6 degrees.
Image search method described in. 6. In the projection data comparison step, if the difference between the created projection data and the template data is equal to or greater than a predetermined value, the processing shifts to comparison with the next template data. Any one of 5
Image search method described in. 7. The image search method according to claim 2, wherein pixel values are counted in a plurality of directions in the projection data creation step. 8. The comparison of the projection data in one direction with the template data based on the comparison result of the projection data in one direction with the template data in the projection data comparison step. Image search method described in. 9. The method according to claim 2, wherein in the projection data comparison step, the comparison is performed using any one of a maximum number of pixels, a minimum number of pixels, a standard deviation, and an average number of pixels.
The image search method according to any one of the above. 10. A recording medium for executing an image search method for searching whether or not a registered image registered in advance exists in an input image, wherein a window scan of a predetermined size is performed on the input image. Window data generating step of generating window data by using the above-mentioned generated window data, and projecting data generating step of generating projection data from the generated window data, and comparing the generated projection data with a plurality of template data registered in advance. A projection data comparison step, an angle detection step of detecting a rotation angle of the input image based on a comparison result in the projection data comparison step, and a match for determining a degree of coincidence between the input image and the registered image using the rotation angle Recording step, wherein a program for executing the degree determination step is recorded. .