Stereo-mapping satellite-image-matching suspicious-region automatic detection method

Abstract

The invention discloses a stereo-mapping satellite-image-matching suspicious-region automatic detection method. The method comprises the following steps of step1, generating an epipolar-line stereo image pair of a stereo image, wherein the epipolar-line stereo image pair includes a left epipolar-line image and a right epipolar-line image; step2, reading the epipolar-line stereo image pair, and carrying out wavelet transform on the left epipolar-line image and the right epipolar-line image of the epipolar-line stereo image pair respectively so as to establish image pyramids of the left epipolar-line image and the right epipolar-line image; and step3, determining suspicious regions in the left epipolar-line image and the right epipolar-line image of each layer of the image pyramids below a top layer in a layer-by-layer mode. In the invention, high-efficient thick-cloud automatic detection and rejecting can be realized; high-efficient shielding area detection, large-area water area detection and non-texture shadow area detection can be realized; and a good actual application value is possessed.

Description

technical field [0001] The invention relates to the technical field of image processing, and more specifically, to a method for automatically detecting suspicious areas in stereoscopic surveying and mapping satellite image matching. Background technique [0002] In the process of satellite and aerial stereo image matching to obtain digital surface models (digital surface model, DSM), due to the existence of factors such as large terrain slopes or occlusions, large areas of shadows, waters, thick clouds, and overexposure, some areas are matched. mistake. These error regions are often patchable in the presence of redundant observations. If these error areas can be detected automatically, the automation degree of DSM extraction under the condition of large area and multiple image redundancy will be significantly improved. [0003] Consistency detection of left and right matching results is a traditional suspicious region detection method, which has a high detection success ra...

AI Technical Summary

Problems solved by technology

However, for the error matching caused by factors such as large-area shadows, water areas, thick clouds, and overexposure, it is difficult to remove the error area by only relying on the consistency detection of the left and right matching results, and because the left and right matching consistency threshold is set too low. Some correct match region removal

Images