Image edge detection method for splicing aerial correction images

Abstract

The invention relates to an image edge detection method for splicing aerial correction images, which belongs to the field of computer vision. Non-ideal information such as pitch angle and roll angle of the unmanned plane can cause nonlinear distortion of the aerial image, and after geometric correction, the edge of the corrected image has a large amount of irregular noise. When a plurality of aerial images are stitched to form a seamless large-scale scene image, if each aerial correction image is directly cropped, a large amount of effective information is cut off, and if not cropped, obvioussplice noise can be generated. The method performs edge detection twice for an aerial correction image with edge noise, distinguishes invalid pixels from effective pixels by a certain rule, and determines the edge noise band. When correcting image stitching, stitching noise can be avoided by only clipping this noise band and at the same time, the effective scene of the image is maintained to the greatest extent.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to an image edge detection method for aerial correction image splicing. Background technique [0002] In recent years, with the continuous maturity of UAV technology, and the advantages of UAVs such as flexible maneuverability and high image resolution, UAVs have been widely used in aerial photography. Due to the limited scene corresponding to a single aerial image, multiple images need to be stitched together through image stitching technology to obtain a high-resolution scene map, which can be used for post-disaster reconstruction measurement, natural environment monitoring, and wartime real-time map construction , Local area high-definition scene map drawing, etc. [0003] Specifically, image stitching is to match two or more images first to obtain the corresponding relationship between each other in space, and then use image fusion strategy to stitch these images into...

AI Technical Summary

Problems solved by technology

During the image correction process, the image needs to be reconstructed using an interpolation algorithm, which can cause noise at the edges of the corrected image

If you do not process the edge of the corrected image and stitch directly, there will be a lot of noise on the stitched image, which will affect the image quality

And if you simply crop the edge of the image, a lot of useful information will be cropped

Images