JP2005309795A - Same object determination device, same object determination method, computer-readable recording medium to which program is recorded, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a same object determination device which compares target regions extracted from satellite images and determines whether the same objects exist, by standardizing and digitizing the differences in shape of the objects, and by determining the differences in shape based on the numerical values thus computed. <P>SOLUTION: Two target regions 101 and 102 are overlapped to each other, and the area of a disagreement region 104 belonging only to one of the two target regions is found. Then, the area of the disagreement region 104 is divided by the length of the outer circumference of an agreement region 103 to compute a standard shape difference parameter. When the standard shape difference parameter is not more than a threshold value, it is determined that the same objects exist, or a standard shape difference parameter, which is obtained by dividing the area of a disagreement region 104 by the area of the object or the square of the length of an outer circumference in stead of the length of the outer circumference, is used. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a same target determination apparatus that compares target areas extracted from satellite images and the like and determines whether or not they are the same target, and standardizes, digitizes, and calculates a difference in the shape of the target area The present invention relates to an apparatus for determining a difference in shape by numerical values.

  There is known a system for identifying a ship using a satellite image that is an optical sensor image composed of luminance values of pixels. Such an optical sensor image is a signal obtained by subjecting a received signal obtained by an optical sensor mounted on an artificial satellite to the reflection of sunlight from the surface of the earth through a CCD camera and further processing at the base station. , Obtained from the reflection of light.

  As a conventional method for identifying a target such as a ship using such a satellite image, a comparison of the total length and width of the ship is known as disclosed in JP-A-2001-221857.

  The comparison of the total length and width of the ship will be briefly explained. FIG. 12 is a diagram illustrating an outline of a target to be subjected to the same determination.

  For example, it is assumed that the sameness is determined between the target areas A and B extracted from the satellite images taken on the previous day and the current day. Conventionally, whether or not the target area extracted from the satellite image is the same target has been determined using the full length and the full width.

  When the total length of the target area A and the total length of the target area B match, and when the total width of the target area A and the total width of the target area B match, both target areas A and B are of the same target judge.

For example, the program is described as follows.
If (Length_A == Length_B &&
(Breadth_A == Breadth_B)
Same = truth;
else
Same = false;
However, even if the full length and the full width match, they are not necessarily the same target. In particular, if the horizontal plane contour (the contour of the shape seen from above) of the target is structurally different, it should be possible to distinguish from the target region.
JP 2001-221857 A

  The present invention was made to eliminate the above-described drawbacks of the prior art, and the object of the present invention is to standardize the difference in the shape of a target area extracted from a satellite image or the like by using a computer. It is an object to automate the determination of whether or not the targets related to the target area are the same by determining the difference in shape based on the calculated numerical values.

The same target determination device according to the present invention is
An identical target determination apparatus for determining whether or not the target object as each subject is the same for two target areas obtained by photographing the target object, and has the following elements (1) 2 Image storage unit for storing one target region (2) Standard shape difference calculation processing unit for calculating a standard shape difference parameter obtained by standardizing the shape difference between two target regions (3) The standard shape difference parameter is compared with a threshold value, and the comparison result Based on the above, a target determination processing unit that determines whether or not the target that is the subject in the two target areas is the same.

The standard shape difference calculation processing unit superimposes the two target areas, obtains a mismatch area belonging to only one of the two target areas on the superimposed coordinates, and calculates a standard shape difference parameter that standardizes the size of the mismatch area. Generate
The target determination processing unit determines that the target area is an image related to the same target object when the standard shape difference parameter is equal to or smaller than a predetermined threshold value or smaller than the predetermined threshold value.

  The standard shape difference calculation processing unit calculates a perimeter length depending on two target regions, and uses a value obtained by dividing the size of the mismatch region by the perimeter length as a standard shape difference parameter.

  The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the superimposed coordinates, obtains an outer circumference length of the coincidence area, and uses the outer circumference length of the coincidence area.

  The standard shape difference calculation processing unit obtains a coincidence area that belongs to both of the two target areas on the superimposed coordinates, obtains an overall outer peripheral length including the coincident area and the non-matching area, and uses the entire outer peripheral length. It is characterized by that.

  The standard shape difference calculation processing unit calculates a target area depending on two target areas, and uses a value obtained by dividing the size of the mismatch area by the target area as a standard shape difference parameter.

  The standard shape difference calculation processing unit obtains a matching region belonging to both of the two target regions on the superimposed coordinates, obtains an area of the matching region, and uses the area of the matching region as the target area. Features.

  The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the superimposed coordinates, obtains an entire area including the coincidence area and the mismatch area, and determines the entire area as the target area. It is used as an area.

The standard shape difference calculation processing unit superimposes the two target regions, obtains a matching region belonging to both of the two target regions on the superimposed coordinates, and a mismatch region belonging to only one of the two target regions, Obtain the area of the matching region, further determine the total area consisting of the matching region and the non-matching region, and determine the value obtained by dividing the area of the matching region by the total area as a standard shape difference parameter,
The target determination processing unit determines that the target area is an image related to the same target object when the standard shape difference parameter is equal to or larger than a predetermined threshold value or larger than the predetermined threshold value.

  The standard shape difference calculation processing unit calculates the outer peripheral length depending on the two target areas, further calculates the square of the outer peripheral length, and calculates a value obtained by dividing the size of the mismatch area by the square of the outer peripheral length. It is characterized by being a parameter.

  The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the overlapped coordinates, obtains the outer circumference length of the coincidence area, further obtains the square of the outer circumference length of the coincidence area, It is characterized by using the square of the outer peripheral length of the region.

  The standard shape difference calculation processing unit obtains a coincidence area that belongs to both of the two target areas on the superimposed coordinates, obtains an overall outer circumference length of the coincidence area and the non-coincidence area, and further determines the overall outer circumference length. The square is obtained, and the square of the entire outer peripheral length is used.

The same target determination method according to the present invention includes:
An identical target determination method for determining whether or not a target that is a subject is the same for two target areas in which a target is photographed, and has the following elements (1) 2 Step of calculating a standard shape difference parameter obtained by standardizing the shape difference between two target areas (2) The standard shape difference parameter is compared with a threshold value, and based on the comparison result, the target that is the subject in the two target areas is the same. Determining whether or not.

A computer-readable recording medium on which a program according to the present invention is recorded,
A computer that records a program for causing a computer to be the same target determination device to determine whether or not the target as the subject is the same for two target areas in which the target is photographed, to execute the following processing (1) Processing for calculating a standard shape difference parameter obtained by standardizing the shape difference between two target areas (2) Comparing the standard shape difference parameter with a threshold value and based on the comparison result The process of determining whether or not the target object that is the subject in the two target areas is the same.

The program according to the present invention is:
It is a program for causing a computer, which is the same target determination device to determine whether or not the target object as each subject is the same, for the two target areas obtained by photographing the target object, to execute the following procedure. (1) A procedure for calculating a standard shape difference parameter obtained by standardizing a shape difference between two target areas (2) The standard shape difference parameter is compared with a threshold, and based on the comparison result, an object in the two target areas is compared. A procedure for determining whether or not a target is the same.

  In the present invention, the difference in the shape of the target area extracted from the satellite image or the like is standardized and digitized, and the difference in shape is determined based on the calculated numerical value. Therefore, it is determined whether or not the targets related to the target area are the same. Judgment can be automated.

Embodiment 1 FIG.
Hereinafter, the present invention will be described based on embodiments shown in the drawings. FIG. 1 is a diagram showing an outline of the invention.

  Reference numerals 101 and 102 denote regions (target regions) that are assumed to be target images extracted from satellite images (an example of a long-distance captured image including an aerial photograph image). However, these are target areas extracted from different satellite images, and it is a problem whether they are image areas related to the same target. In other words, it is necessary to determine whether each subject is the same (for example, including the same type of object as a sister ship).

In the present invention, the two target areas to be compared are overlapped, and the identity of the target is determined using parameters calculated from the coincidence portion and the disagreement portion of both areas. Note that the overlapping is performed so that the direction and the center of gravity of the target areas coincide. If the resolutions are different, the image of one or both target areas is corrected to be enlarged or reduced at a predetermined magnification in the X-axis direction and the Y-axis direction so that the same resolution is obtained according to the resolution ratio. Overlapping with.

  In the present invention, a parameter obtained by standardizing the difference in the shape of the target area is used for determining the identity of the target. This parameter is called a standard shape difference parameter. Details will be described later.

  Hereinafter, the configuration of the same target determination apparatus and the process of the same target determination by the apparatus will be described. FIG. 2 is a diagram illustrating a configuration of the same target determination device. The same target determination apparatus includes an image reading processing unit 201, a read image storage unit 202, an image binarization processing unit 203, a binary image storage unit 204, a principal component analysis processing unit 205, a principal component data storage unit 206, and resampling. Each element includes a processing unit 207, a resampling image storage unit 208, a standard shape difference calculation processing unit 209, a superimposed image storage area 210, a standard shape difference parameter storage unit 211, and a target determination processing unit 212. In addition, an area for storing each parameter is provided.

  FIG. 3 is a diagram showing an overall processing flow of the same target determination. First, an image reading process (S301) by the image reading processing unit 201 is performed. Two images to be determined are read from a CD-ROM or a hard disk into the read image storage unit 202 (a temporary storage area is sufficient). In this example, the image to be read is an image in which multivalued pixel groups are arranged in a matrix.

  Next, image binarization processing (S302) by the image binarization processing unit 203 is performed. In this process, the target area is extracted for each of the two read images. This extraction is performed by processing in consideration of distinction by luminance and continuity of regions. Then, a binarized image that distinguishes the target area from the other areas is generated as an extraction result, and stored in the binarized image storage unit 204 (a temporary storage area is sufficient). In this example, the pixel value of the target area is set to “1”, and the pixel values of the other areas are set to “0”.

  Subsequently, principal component analysis processing (S303) by the principal component analysis processing unit 205 is performed. In this processing, principal components such as principal component direction (target orientation), center of gravity, full length, full width, etc. (graphical) are respectively obtained for target regions in two binary images stored in the binarized image storage unit 204. Feature) data. The calculated two principal component data are stored in the principal component data storage unit 206.

  Next, resampling processing (S304) by the resampling processing unit 207 is performed. In this process, one or both of the target areas in the two binarized images stored in the binarized image storage unit 204 are rotated, the principal component direction of each target is normalized, and similarly enlarged or reduced. And the center of gravity and the resolution (length per pixel) are normalized by the movement and two resampled images are obtained. These resampled images have the same direction, center of gravity, and resolution. These resampled images are stored in the resampled image storage unit 208 (a temporary storage area having a size of Ship_x × Ship_y is sufficient).

  In this example, resampling for overlapping the target areas is performed using the center of gravity as a reference, but another point may be used as a reference point. For example, resampling may be performed in which the target areas are overlapped with respect to the tip or the center point of the last tail.

  Then, standard shape difference calculation processing (S305) by the standard shape difference calculation processing unit 209 is performed. In this process, a standard shape difference parameter between two target regions is calculated. The calculated standard shape difference parameter is stored in the standard shape difference parameter storage unit 211.

  Finally, target determination processing (S306) by the target determination processing unit 212 is performed. The standard shape difference parameter stored in the standard shape difference parameter storage unit 211 is compared with a predetermined threshold, and based on the comparison result, whether or not the target object that is the subject in the target area is the same (same type). Make a decision.

  Here, the standard shape difference calculation process (S305) will be described in detail. FIG. 4 is a diagram showing a standard shape difference calculation processing flow.

  First, image logic operation processing (S401) is performed. FIG. 5 is a diagram showing an outline of the image logic operation. In this process, images of two target areas normalized by resampling are overlapped, and a logical operation is performed for each pixel one by one, and either “match”, “mismatch”, or “out of area” is selected. Judgment is made. The determination result is stored in a corresponding pixel on the superimposed image storage area 210 as a superimposed image.

  FIG. 6 is a diagram illustrating determination of an image logical operation. The pixel value of one target area A corresponding to the same position on the resampled image and the pixel value of the other target area B are logically calculated to determine the determination result. When both are “0” (pixels outside the target area), it is determined as “out of area”, and “0” is stored as an output value (determination code) indicating “out of area”. If any one is “1” (pixels in the target area) and the other is 0 (pixels outside the target area), it is determined as “mismatch”, and “1” is output as an output value indicating “mismatch”. Is stored. If both are “1” (pixels in the target area), it is determined as “match”, and “2” is stored as an output value indicating “match”.

  For each pixel, the number of determination results (Count_0 as the number of pixels outside the area, Count_1 as the number of pixels that do not match, and Count_2 as the number of pixels that match) is counted. In this manner, the total number of pixels “out of the area”, the total number of “mismatch”, and the total number of “match” are finally calculated.

  Next, as shown in FIG. 4, a convex hull calculation process (S402) is performed. In this process, the outer peripheral length of the pixel area (example of ship area and target area) determined as “match” by the image logic operation process (S401) is calculated. In this example, the perimeter length is obtained as the convex hull length (Convex_Hull). This outer peripheral length is an example of an outer peripheral length depending on two target areas. Hereinafter, the convex hull calculation process will be described in detail.

  FIG. 7 is a diagram showing a convex hull calculation processing flow. First, outline extraction processing (S701) is performed. As described below, for the pixel region identified as “match” on the superimposed image (hereinafter referred to as “match region”, an example of a ship region), the coordinates of the outer pixel are obtained and stored in each parameter region.

  First, the X coordinate value of each pixel in the matching area is sequentially compared to determine the minimum value, and stored in the parameter Xmin as the matching area minimum X coordinate. Similarly, the X coordinate value of each pixel in the matching area is sequentially compared to determine the maximum value, and stored in the parameter Xmax as the matching area maximum X coordinate. Further, the Y coordinate values are sequentially compared for the pixel groups having the same X coordinate value in the coincidence area, the minimum value is determined, and stored in the parameter Ymin [ship_x] as the minimum Y coordinate for each X coordinate. Ship_x is each X coordinate value, and Ymin is array type data. Further, sequentially, the Y coordinate values of pixel groups having the same X coordinate value in the coincidence area are compared, the maximum value is determined, and stored in the parameter Ymax [ship_x] as the maximum Y coordinate for each X coordinate. Ship_x is each X coordinate value, and Ymax is array type data.

  An example is given for the above parameters. FIG. 8 is a diagram illustrating an example of a superimposed image that is a target of outermost contour extraction.

  The minimum value of the X coordinates of each pixel occupying the matching area (shaded area in the figure) is “2” as shown in the figure, and Xmin (matching area minimum X coordinate: example of ship area minimum X coordinate) = 2.

  On the other hand, the maximum value among the X coordinates of each pixel occupying the coincidence area is “13” as shown in the figure, and Xmax (example of coincidence area maximum X coordinate: ship area maximum X coordinate) = 13.

  The minimum value of each X coordinate occupying the matching area is “0”, “2”, “2”, “2”, “2”, “2”, “2”, “2” in the order of X = 1 to 14. ”,“ 2 ”,“ 2 ”,“ 3 ”,“ 3 ”,“ 4 ”,“ 0 ”. Therefore, Ymin [ship_x] (minimum Y coordinate for each X coordinate of the coincidence area: example of minimum Y coordinate for each X coordinate of the ship area) = (0, 2, 2, 2, 2, 2, 2, 2 , 2, 2, 3, 3, 4, 0). Note that the minimum value of the X coordinate (X = 1, 14 in this example) having no matching area is “0”.

  On the other hand, the maximum value of each X coordinate occupying the coincidence area is “0”, “6”, “6”, “6”, “6”, “6”, “6”, in the order of X = 1 to 14. “6”, “6”, “6”, “5”, “5”, “4”, “0”. Therefore, Ymax [ship_x] (maximum Y coordinate for each X coordinate of the matching area: example of maximum Y coordinate for each X coordinate of the ship area) = (0, 6, 6, 6, 6, 6, 6, 6 , 6, 6, 5, 5, 4, 0). Note that the maximum value of the X coordinate (X = 1, 14 in this example) having no matching area is “0”.

  Next, the convex hull lower part calculation process (S702) is performed as shown in FIG. In this process, the convex hull below the matching region is obtained using Ymin [ship_x] obtained by the outline extraction process (S701).

  The convex hull lower part calculation process (S702) will be described in detail. FIG. 9 is a diagram showing a convex hull lower part calculation processing flow. FIG. 10 is a diagram illustrating an example of a superimposed image that is a target for calculating the convex hull lower part. Each process will be described with reference to the example of FIG. The upper contour in the figure is the lower part of the convex hull that is the calculation target of this processing.

  First, the counter I is initialized (S901). Specifically, Xmin (matching region minimum X coordinate) is substituted into counter I. In the example of FIG. 10, “2” is assigned to the counter I.

  Next, a base point (Xbp, Ybp) as a parameter is set (S902). Specifically, the counter I is substituted into Xbp that is the X coordinate of the base point, and the array value Ymin [I] of the minimum Y coordinate with the counter I as a subscript is substituted into Ybp that is the Y coordinate of the base point. In the example of FIG. 10, the base point (2, 2) is set with the value “2” of the counter I as the X coordinate and the value “2” of the array value Ymin [I = 2] as the Y coordinate. 1001 in the figure is the base point.

  In step S903, the parameter J is initialized with I + 1, incremented sequentially, and candidate points (J, Ymin [J]) are set based on the parameter J up to Xmax (matching region maximum X coordinate). The slope of the straight line connecting is calculated. Then, the inclinations of the candidate points are compared to determine the smallest inclination, and the candidate point having the smallest inclination is selected. When there are a plurality of candidate points having the smallest inclination, it is effective to select a candidate point having the largest X coordinate value (= parameter J). Subsequent processing is performed using the parameter J which is the X coordinate value of the candidate point selected in this way. In the example of FIG. 10, the candidate points (3, 2), (4, 2), (5, 2), (6, 2), (7, 2), (8) with respect to the base point (2, 2). , 2), (9, 2), and (10, 2) are all minimal with a slope of 0. Therefore, the candidate point (10, 2) having the largest X coordinate value is selected. Therefore, the subsequent processing is performed using “10” as the parameter J.

  The distance between the two points of the base point (Xbp, Ybp) and the selected candidate point (J, Ymin [J]) is added so as to be accumulated as the length of the convex hull (S904). Specifically, first, in order to calculate the distance between two points, the square value of the difference between the X coordinate values and the square value of the difference between the Y coordinate values are calculated, and the square root of these sums is calculated. Next, the calculated distance between the two points is added to the parameter Convex_Hull indicating the length of the convex hull. In the example of FIG. 10, the distance “8” from the base point 1001 to the candidate point 1002 is obtained, and in addition to the convex hull length parameter (initial value 0), the convex hull length parameter is “8”.

  As described above, the processing for one line segment when the convex hull is approximated to a polygon is completed. In order to process the next line segment, the parameter J, which is the X coordinate value of the candidate point, is substituted into the counter I (S905). Thereby, the candidate point in the previous process becomes the base point in the next process. In the example of FIG. 10, the value “10” of the parameter J is assigned to the counter I, and the point 1002 becomes the next base point.

  Then, the processing for the line segment (S902 to S904) is repeated as described above until the counter I reaches Xmax (matching region maximum X coordinate) (S906). Thereby, the length of the convex hull lower part from Xmin (matching region minimum X coordinate) to Xmax (matching region maximum X coordinate) is calculated. In the example of FIG. 10, the length of the lower part of the convex hull is obtained by summing the distance between the points 1001 and 1002, the distance between the points 1002 and 1003, and the distance between the points 1003 and 1004.

  Subsequently, as shown in FIG. 7, a convex hull upper part calculation process (S703) is performed. In this process, the convex hull at the top of the matching region is obtained using Ymax [ship_x] obtained by the outline extraction process (S701). In this process, the operation of the convex hull lower part calculation process (S702) described above is reversed on the image. The lower profile in FIG. 10 is the upper part of the convex hull that is the calculation target of this process.

  The convex hull upper part calculation process (S703) will be described in detail. FIG. 11 is a diagram showing a convex hull upper part calculation processing flow.

  First, similarly to S901, the counter I is initialized (S1101). Next, a base point (Xbp, Ybp) is set (S1102). Unlike S902, the array value Ymax [I] of the maximum Y coordinate is substituted into Ybp which is the Y coordinate of the base point.

  In S1103, the parameter J is initialized with I + 1, and the candidate point (J, Ymax [J]) is set based on the parameter J up to Xmax (matching region maximum X coordinate) while sequentially incrementing. The slope of the straight line connecting is calculated. Then, the inclination of each candidate point is compared to determine the largest inclination, and the candidate point having the largest inclination is selected. When there are a plurality of candidate points having the largest inclination, the candidate point having the largest X coordinate value (= parameter J) is selected. Subsequent processing is performed using the parameter J which is the X coordinate value of the candidate point selected in this way.

  Similar to S904, the distance between the two points is accumulated so as to be accumulated as the length of the convex hull (S1104). Then, the processing for the line segment (S1102 to S1104) is repeated (S1106) as described above until the counter I reaches Xmax (matching region maximum X coordinate). Thereby, the length of the convex hull upper part from Xmin (matching region minimum X coordinate) to Xmax (matching region maximum X coordinate) is calculated.

Finally, as shown in FIG. 7, a convex hull summation process (S704) is performed. The length of the lower part of the convex hull calculated in S702 and the length of the upper part of the convex hull calculated in S703 are added together to obtain the length of the entire convex hull (the outer peripheral length of the matching region). If there is a continuous pixel on Xmin (the matching region minimum X coordinate), the length of the column of the pixel is also added. Similarly, when there is a continuous pixel on Xmax (matching region maximum X coordinate), the length of the column of the pixel is also added.

  Subsequent to the above-described convex hull calculation processing (S402), standard shape difference calculation processing (S403) is performed as shown in FIG. In this process, a standard shape difference parameter (Std_Shape_Diff), which is a parameter obtained by standardizing the shape difference of the target area, is calculated.

  In the present embodiment, a value obtained by dividing the area of the mismatched portion (mismatched region) between the target region A and the target region B by the outer peripheral length of the matched portion (matched region) between the target region A and the target region B (matched region outer peripheral length pair The discrepancy area ratio) is used as a standard shape difference parameter. When the standard shape difference parameter is equal to or less than a certain threshold value, it is determined that the target object that is the subject in both target areas is the same. Note that the matching area outer circumference length to the mismatch area area ratio is an example of a standard shape difference parameter related to the outer circumference length to the mismatch area area ratio.

For example, the program is described as follows. In this example, sqar () is a macro function for obtaining a square value.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Convex_Hull / resolution (= periphery length of matching region));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / Resolution) / Convex_Hull;

Embodiment 2. FIG.
As a standard shape difference parameter, a mode in which a parameter (total outer circumference length to mismatch area area ratio) obtained by dividing the area of the mismatch area by the entire outer circumference length (the outer circumference length of the OR area of the target area A and the target area B) will be described. . This parameter is also an example of the perimeter length versus the mismatched area area ratio.

  In the convex hull calculation processing (S402) of the first embodiment, the outer circumference length is calculated for the matching area. In this embodiment, the outer circumference length (convex hull is set for the entire area including the matching area and the non-matching area. (Length: Convex_Hull) is calculated. That is, as the outer peripheral length of the ship region, the outer peripheral length of the entire region is used instead of the outer peripheral length of the matching region. This outer peripheral length is an example of an outer peripheral length depending on two target areas.

  Further, in the standard shape difference calculation process (S403), the standard shape difference parameter is calculated using the entire outer peripheral length described above. Similarly to the first embodiment, when the standard shape difference parameter is equal to or smaller than a certain threshold value, it is determined that the target object as the subject in both target areas is the same.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Convex_Hull / resolution (= total outer perimeter length));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / Resolution) / Convex_Hull;
The standard shape difference parameter, which is the ratio of the perimeter length to the non-matching area shown in the first and second embodiments, is an [m] dimension and is a parameter depending on the target size. However, since it has a dimension of [m] like the full length and full width, there is an advantage that the comparative examination with the full length and full width is easy.

Embodiment 3 FIG.
In the present embodiment, a value obtained by dividing the area of the mismatch region between the target region A and the target region B by the area of the match region between the target region A and the target region B (matched region vs. mismatch region area ratio) is used as the standard shape difference parameter. Use. The area of the coincidence area between the target area A and the target area B is an example of a target area that depends on the two target areas.

  In the present embodiment, since the outer peripheral length is not used, the convex hull calculation process (S402) is unnecessary.

  In the standard shape difference calculation process (S403), the standard shape difference parameter is calculated by dividing the area of the mismatch region by the area of the match region as described above. When the standard shape difference parameter is equal to or less than a certain threshold value, it is determined that the target object that is the subject in both target areas is the same. Note that the area ratio of the matching area to the mismatching area is an example of a standard shape difference parameter related to the area ratio.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Count_2 / sqar (resolution) (= area of matching region));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
Count_1 / Count_2;

Embodiment 4 FIG.
As a standard shape difference parameter, a mode in which a parameter obtained by dividing the area of the inconsistent region by the entire area (whole area / inconsistent region area ratio) will be described. The total area (the area of the OR region between the target region A and the target region B) is obtained by adding the areas of the coincidence region and the mismatch region. The overall area is an example of a target area that depends on two target areas.

  Also in this embodiment, since the outer peripheral length is not used, the convex hull calculation process (S402) is unnecessary.

  In the standard shape difference calculation process (S403), the standard shape difference parameter is calculated by dividing the area of the non-matching region by the entire area as described above. When the standard shape difference parameter is equal to or less than a certain threshold value, it is determined that the target object that is the subject in both target areas is the same. Note that the area ratio of the entire area to the inconsistent area is an example of a standard shape difference parameter related to the area ratio.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Count_1 + Count_2 / squar (resolution) (= total area));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
Count_1 / Count_1 + Count_2;

Embodiment 5 FIG.
In the present embodiment, a parameter obtained by dividing the area of the coincidence region by the entire area (whole area vs. coincidence region area ratio) is used as the standard shape difference parameter.

  Also in this embodiment, since the outer peripheral length is not used, the convex hull calculation process (S402) is unnecessary.

  In the standard shape difference calculation process (S403), the standard shape difference parameter is calculated by dividing the area of the coincidence region by the total area as described above. When the standard shape difference parameter is greater than or equal to a certain threshold value, it is determined that the target object that is the subject in both target areas is the same.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_2 / sqar (resolution) (= area of matching region)) /
(Count_1 + Count_2 / squar (resolution) (= total area));
if (Std_Shape_Diff> = threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
Count_2 / Count_1 + Count_2;
As in the third to fifth embodiments, the standard shape difference parameter, which is the ratio between areas, is a dimensionless amount [square m / square m] and does not depend on the size of the target. In other words, since it is a parameter that standardizes only the difference in shape that does not include the size, the same degree of reliability evaluation results can be obtained both when comparing large target areas and when comparing small target areas. There is an advantage.

Embodiment 6 FIG.
It is also effective to replace the outer peripheral length of the coincidence area, which is the denominator of the standard shape difference parameter shown in the first embodiment, with its square.

  Therefore, in the standard shape difference calculation process (S403), the area of the mismatched portion (mismatched region) between the target region A and the target region B is divided by the square of the outer peripheral length of the matched portion (matched region) between the target region A and the target region B. Is used as the standard shape difference parameter. When the standard shape difference parameter is equal to or less than a certain threshold value, it is determined that the target object that is the subject in both target areas is the same. The matched area outer circumference squared to mismatch area area ratio is an example of a standard shape difference parameter related to the outer circumference squared to mismatch area area ratio.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Sqar (Convex_Hull / resolution));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
Count_1 / squar (Convex_Hull);
Embodiment 7 FIG.
It is also effective to replace the entire outer peripheral length, which is the denominator of the standard shape difference parameter shown in the second embodiment, with its square.

  Therefore, in the standard shape difference calculation process (S403), a value obtained by dividing the area of the mismatched portion (mismatched region) between the target region A and the target region B by the square of the outer peripheral length of the entire region (total outer peripheral length squared vs. mismatched region area ratio). ) As the standard shape difference parameter. When the standard shape difference parameter is equal to or less than a certain threshold value, it is determined that the target object that is the subject in both target areas is the same. Note that the overall outer circumference squared to mismatch area area ratio is an example of a standard shape difference parameter related to the outer circumference square to mismatch area area ratio.

For example, the program is described as follows.
Std_Shape_Diff (= standard shape difference parameter) =
(Count_1 / squar (resolution) (= area of mismatch region)) /
(Sqar (Convex_Hull / resolution));
if (Std_Shape_Diff <= threshold) Determination result = same;
else judgment result = different;
The calculation formula of Std_Shape_Diff (= standard shape difference parameter) described above can be reduced and modified as follows.
Std_Shape_Diff (= standard shape difference parameter) =
Count_1 / squar (Convex_Hull);
The above-described same target determination apparatus is a computer, and each element can execute processing by a program. Further, the program can be stored in a storage medium so that the computer can read the program from the storage medium.

  In the target determination process (S306) of each embodiment, the determination that is based on the condition that the threshold value is “below” can be made the condition that the threshold value is “smaller”. The determination that is made on condition that the threshold is “greater than or equal to” may be made on the condition that the threshold is “greater than”.

It is a figure which shows the outline | summary of invention. It is a figure which shows the structure of the same target determination apparatus. It is a figure which shows the whole process flow of the same target determination. It is a figure which shows a standard shape difference calculation process flow. It is a figure which shows the outline of an image logic operation. It is a figure which shows determination of an image logic operation. It is a figure which shows a convex hull calculation processing flow. It is a figure which shows the example of the superimposed image used as the object of outermost contour extraction. It is a figure which shows a convex hull lower part calculation process flow. It is a figure which shows the example of the overlay image used as the object of convex hull lower part calculation. It is a figure which shows the convex hull upper part calculation processing flow. It is a figure which shows the outline of the target used as the object of the same determination.

Explanation of symbols

201 image reading processing unit, 202 reading image storage unit, 203 image binarization processing unit, 204 binarized image storage unit, 205 principal component analysis processing unit, 206 principal component data storage unit, 20
7 resampling processing unit, 208 resampling image storage unit, 209 standard shape difference calculation processing unit, 210 superimposed image storage region, 211 standard shape difference parameter storage unit, 212 target determination processing unit.

Claims (15)

The same target determination apparatus for determining whether or not the target object as each subject is the same for two target areas obtained by photographing the target object, and having the following elements: (1) Image storage unit for storing two target regions (2) Standard shape difference calculation processing unit for calculating a standard shape difference parameter obtained by standardizing the shape difference between two target regions (3) Compare the standard shape difference parameter with a threshold value And a target determination processing unit that determines whether or not the target object that is the subject in the two target areas is the same based on the comparison result. The standard shape difference calculation processing unit superimposes the two target areas, obtains a mismatch area belonging to only one of the two target areas on the superimposed coordinates, and calculates a standard shape difference parameter that standardizes the size of the mismatch area. Generate
The target determination processing unit determines that the target region is an image related to the same target object when the standard shape difference parameter is equal to or smaller than a predetermined threshold value or smaller than the predetermined threshold value. 1. The same target determination apparatus according to 1.   3. The standard shape difference calculation processing unit calculates a perimeter length depending on two target areas, and uses a value obtained by dividing the size of the mismatch area by the perimeter length as a standard shape difference parameter. The same target judging device as described.   The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the superimposed coordinates, obtains an outer circumference length of the coincidence area, and uses the outer circumference length of the coincidence area. Item 4. The same target determination device according to Item 3.   The standard shape difference calculation processing unit obtains a coincidence area that belongs to both of the two target areas on the superimposed coordinates, obtains an overall outer peripheral length including the coincident area and the non-matching area, and uses the entire outer peripheral length. The same target determination apparatus according to claim 3.   The standard shape difference calculation processing unit calculates a target area depending on two target regions, and uses a value obtained by dividing the size of the mismatch region by the target area as a standard shape difference parameter. Item 3. The same target determination device according to Item 2.   The standard shape difference calculation processing unit obtains a matching region belonging to both of the two target regions on the superimposed coordinates, obtains an area of the matching region, and uses the area of the matching region as the target area. The same target determination device according to claim 6, wherein   The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the superimposed coordinates, obtains an entire area including the coincidence area and the mismatch area, and determines the entire area as the target area. The same target determination apparatus according to claim 6, wherein the same target determination apparatus is used as an area of The standard shape difference calculation processing unit superimposes the two target regions, obtains a matching region belonging to both of the two target regions on the superimposed coordinates, and a mismatch region belonging to only one of the two target regions, Find the area of the matching region, further determine the overall area consisting of the matching region and the mismatch region, and determine the value obtained by dividing the area of the matching region by the total area as a standard shape difference parameter,
The target determination processing unit determines that the target region is an image related to the same target object when the standard shape difference parameter is equal to or larger than a predetermined threshold value or larger than the predetermined threshold value. 1. The same target determination apparatus according to 1.   The standard shape difference calculation processing unit calculates the outer peripheral length depending on the two target areas, further calculates the square of the outer peripheral length, and calculates a value obtained by dividing the size of the mismatch area by the square of the outer peripheral length. The same target determination apparatus according to claim 2, wherein the same target determination apparatus is used as a parameter.   The standard shape difference calculation processing unit obtains a coincidence area belonging to both of the two target areas on the overlapped coordinates, obtains the outer circumference length of the coincidence area, further obtains the square of the outer circumference length of the coincidence area, 11. The same target determination apparatus according to claim 10, wherein the square of the outer peripheral length of the region is used.   The standard shape difference calculation processing unit obtains a coincidence area that belongs to both of the two target areas on the superimposed coordinates, obtains an overall outer circumference length of the coincidence area and the non-coincidence area, and further determines the overall outer circumference length. 11. The same target determination device according to claim 10, wherein the square is obtained and the square of the entire outer peripheral length is used. The same target determination method for determining whether or not the target object as each subject is the same for the two target areas obtained by photographing the target object, the method having the following elements: (1) A step of calculating a standard shape difference parameter obtained by standardizing a shape difference between two target areas. (2) The standard shape difference parameter is compared with a threshold value, and a target that is a subject in the two target areas based on the comparison result. Determining whether or not are identical. A computer that records a program for causing a computer to be the same target determination device to determine whether or not the target as the subject is the same for two target areas in which the target is photographed, to execute the following processing A readable recording medium (1) A process of calculating a standard shape difference parameter obtained by standardizing a shape difference between two target areas (2) A standard shape difference parameter is compared with a threshold value, and two target areas are calculated based on the comparison result. Processing for determining whether or not the target object as the subject is the same. A program (1) 2 for causing a computer, which is the same target determination device to determine whether or not the target as a subject is the same, for two target areas obtained by photographing the target, to execute the following procedure Procedure for calculating a standard shape difference parameter that standardizes the shape difference between two target regions (2) The standard shape difference parameter is compared with a threshold value, and based on the comparison result, the target that is the subject in the two target regions is the same A procedure for determining whether or not.

Images