An automatic selection method of minimum image set considering topological geometric multi-element constraints

Abstract

This patent discloses an automatic optimization method for the smallest image set that takes into account the constraints of topological and geometric multi-factors: first, the coverage of each image is determined by using the auxiliary data (including position, attitude, etc.) , establish the image adjacency matrix; secondly, based on the image adjacency matrix, extract the topological geometric elements related to the stereo image matching stability, such as base-height ratio, overlapping degree, image intersection angle, etc., and establish the topological geometric elements and the stereo image matching stability Third, extract image factors that can reflect image quality, such as blurriness, texture complexity, etc., and establish the functional relationship between each image factor and the number of feature extraction and its spatial distribution; then, combined with the above research results, establish Taking into account the comprehensive functional relationship between multi-topological geometric elements, image factors and the number of adjacent images, the importance weight of the image is judged; finally, with the minimum N-eye vision as the constraint condition, the necessity of each image participating in the matching is sequentially judged, and the redundancy is eliminated Image, the minimum image set that meets the stability requirements of multi-view image matching is selected.

Description

technical field [0001] The invention belongs to the field of image matching, and in particular relates to an adaptive optimization method for a minimum image set taking into account topological geometry multi-element constraints. Background technique [0002] The rapid development of mobile remote sensing platforms has made it easier and cheaper for people to obtain remote sensing images. Mount sensors such as digital cameras, video cameras, and tilt cameras on mobile platforms such as drones and vehicles, and indirectly restore the three-dimensional structure of objects in the survey area through multi-view image matching. And other fields have a very broad application prospect, which has greatly promoted the development of geographic national conditions monitoring technology. However, considering the existence of errors such as POS system error and camera shake of the mobile remote sensing platform, in order to obtain complete range and detailed ground object information,...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to design a minimum image set that takes into account the constraints of topological geometry and multi-element constraints to solve the problem of low image matching efficiency due to the large amount of data redundancy in the massive short-baseline, high-overlapping, and multi-angle image sets acquired by the mobile remote sensing platform. Adapt to the preferred method, optimize the data structure

Images