A two-dimensional image mosaic method based on raster positioning in automatic visual inspection

Abstract

The invention belongs to the field of automatic image detection, and specifically discloses a two-dimensional image mosaic method based on raster positioning in automatic visual detection. The splicing method of the present invention corrects the coordinate values ​​of the sub-pictures in the large picture through the angle parameters between the shooting platform and the camera coordinates, calculates the position coordinates of the four corners of each sub-picture, and obtains the size of the big picture through the actual physical distance corresponding to each pixel. Bilinear interpolation calculates the pixel gray level of the sub-pixel coordinate position in the sub-image to obtain the pixel gray level in the large image, and obtains the spliced ​​large image. Through the parameter correction of the angle between the shooting platform and the camera coordinates, the error caused by the inconsistency of the shooting coordinates is eliminated. At the same time, when stitching, the coordinate positions of each sub-image are directly calculated to avoid stitching errors caused by the stitching method of image feature detection stitching. At the same time, the calculated coordinates use sub-pixel accuracy, which improves Stitched image quality.

Description

Technical field: [0001] The invention belongs to the field of automatic image detection, and specifically discloses a two-dimensional image mosaic method based on raster positioning in automatic visual detection. Background technique: [0002] In industrial production, it is necessary to inspect the surface of devices with textured features. Since the surface texture features of the inspected device are relatively small and cannot be distinguished by the naked eye, it is necessary to collect images by taking pictures, and then use the images for inspection. In order to ensure the accuracy of surface texture detection, it is necessary to take high-precision images. When the inspected device is large in size, the camera cannot capture all the device surfaces in one image. It is necessary to take multiple sub-images of the surface of the inspected device, and then form a complete large image through image stitching for inspection. When the surface of the detected device has ...

AI Technical Summary

Problems solved by technology

When the surface of the detected device has repeated texture features, the current splicing method is prone to overlap of local features, resulting in splicing errors, which affect the detection accuracy

Images