초록 ▼

An image matching method capable of matching images with a high precision and a program and an image matching system for the same, providing a conversion unit for performing image processing based on a registered image and a match image for converting points in each image to patterns of curves based

An image matching method capable of matching images with a high precision and a program and an image matching system for the same, providing a conversion unit for performing image processing based on a registered image and a match image for converting points in each image to patterns of curves based on a distance from a reference position to the closest point on a straight line passing through each point in the image from the reference position and an angle formed by a straight line passing through the reference position and the closest point and an x-axis as a reference axis including the reference position, converting linear components in the images to patterns of a plurality of overlapped curves, and generating converted images, a correlation value generation unit for performing correlation processing based on the converted images and generating a correlation value, and a matching unit for performing the matching based on a signal indicating the correlation value generated by the correlation value generation unit.

대표청구항 ▼

The invention claimed is: 1. An image matching method implemented on a computer for matching a first image and a second image, comprising: performing image conversion processing based on a distance from a single reference position in each of the first image and second image and an angle formed by a

The invention claimed is: 1. An image matching method implemented on a computer for matching a first image and a second image, comprising: performing image conversion processing based on a distance from a single reference position in each of the first image and second image and an angle formed by a straight line passing through the reference position and a reference axis including the reference position and generating a first converted image and a second converted image in a two-dimensional space defined by the distance and the angle; performing matching processing of the first image against the second image based on results of correlation processing at a plurality of different relative locations in a first direction and a second direction orthogonal to the first direction in the first converted image and second converted image generated during image conversion; and judging whether or not image conversion processing is performed, before performing image conversion processing: an image conversion processing is carried out to the first and second images, on the basis of a distance from reference positions in the first and second images, and an angle between lines passing through the reference positions and reference axis including the reference positions; a generation is carried out for third and fourth conversion images in a two dimensional space defined by the distance and the angles of a lower resolution than that of the first and second converted images; and a judgment is carried out on the basis of a correlation processing result in a different plurality of relative positions in a first direction and a second direction perpendicular to the first direction on the third and fourth converted images. 2. An image matching method as set forth in claim 1, wherein, during image conversion, the processing is carried out of image conversion processing of converting points in each image to patterns of curves based on the distance from the reference position to the closest point on a straight line passing through the points in the image and an angle formed by the straight line passing through the reference position and the closest point and the reference axis including the reference position and converting the linear components in each image to patterns of a plurality of overlapped curves to generate the first converted image and second converted image. 3. An image matching method as set forth in claim 2, further comprising performing a position correction on the basis of patterns in the first and second converted images generated during image conversion, wherein, the image matching of the first and second image is carried out on a degree of overlapping of patterns in the first and second converted images processed during the position correction, and a coincidence or non-coincidence of patterns in the first and second converted images. 4. An image matching method as set forth in claim 3, further comprising extracting a region of the respective first and second converted images, the region having a degree of overlapping of the pattern of the curves in one converted image higher than a predetermined threshold value, wherein, the position correction is carried out on the basis of the pattern in the extracted region of the first and second converted images, and wherein, the image matching is carried out on the basis of a coincidence or non-coincidence of the pattern in the extracted region of the first and second converted images, corrected position. 5. An image matching method as set forth in claim 3, wherein during matching processing, a comparison is carried out to a different plurality of positional relationships in the first and second converted images; a degree of similarity as the correlation value is generated from the comparison result; and the image matching is carried out on the basis of the generated degree of similarity. 6. An image matching method as set forth in claim 1, wherein, during matching processing, the first and second converted images generated during image conversion are processed by a Fourier transformation in the first and second directions, and the first and second image are matched on the basis of correlated values responsive to phase components of the resultant Fourier transformation processed values. 7. Program instructions encoded on a computer readable medium to be run by an information processing apparatus for matching a first image and a second image, comprising: a first routine for performing image conversion processing based on a distance from a single reference position in each of the first image and second image and an angle formed by a straight line passing through the reference position and a reference axis including the reference position and generating a first converted image and a second converted image in a two-dimensional space defined by the distance and the angle; a second routine of performing matching processing of the first image against the second image based on results of correlation processing at a plurality of different relative locations in a first direction and a second direction orthogonal to the first direction in the first converted image and second converted image generated at the first routine; and an additional routine for judging whether or not the process of the first routine in performed, before performing the first routine, in the additional routine: an image conversion processing is carried out to the first and second images, on the basis of a distance from reference positions in the first and second images, and an angle between lines passing through the reference positions and reference axis including the reference positions; a generation is carried out for third and fourth conversion images in a two dimensional space defined by the distance and the angles of a low resolution than that of the first and second converted images; and a judgment is carried out on the basis of a correlation processing result in different plurality of relative positions in a first direction and a second direction perpendicular to the first direction on the third and fourth converted images. 8. Program instructions as set forth in claim 7, wherein including a routine for performing, in the first routine, image conversion processing of converting points in each image to patterns of curves based on the distance from the reference position to the closest point on a straight line passing through the points in the image and an angle formed by the straight line passing through the reference position and the closest point and the reference axis including the reference position and converting the linear components in each image to patterns of a plurality of overlapped curves to generate the first converted image and second converted image. 9. Program instructions as set forth in claim 8, further comprising a third routine for performing a position correction on the basis of patterns in the first and second converted images generated in the first routine, wherein, in the second routine, the image matching of the first and second image is carried out on a degree of the overlapping of patterns in the first and second converted images processed during the position correction in the third routine, and a coincidence or non-coincidence of patterns in the first and second converted images. 10. Program instructions as set forth in claim 9, further comprising a fourth routine for extracting a region of the respective first and second converted images, the region having a degree of overlapping of the pattern of the curves in one converted image higher than a predetermined threshold value, wherein in the third routine, the position correction is carried out on the basis of the pattern in the extracted region of the first and second converted images in the fourth step, and wherein in the second routine, the image matching is carried out on the basis of a coincidence or non-coincidence of the pattern in the extracted region of the first and second converted images, correction position in the third routine. 11. Program instructions as set forth in claim 10, wherein in the second routine: a comparison is carried out to a different plurality of positional relationships in the first and second converted images, generated in the first routine; a degree of similarity as the correlation value is generated from the comparison result; and the image matching is carried out on the basis of the generated degree of similarity. 12. Program instructions as set forth in claim 7, wherein, in the second routine, the first and second converted images generated in the first routine are processed by a Fourier transformation in the first and second directions, and the first and second image are matched on the basis of correlated values responsive to phase components of the resultant Fourier transformation processed values.