Aviation wide baseline stereopair line segment feature matching method based on global and local fusion constraints

Abstract

The invention provides an aviation wide baseline stereopair line segment feature matching method based on global and local fusion constraints, and aims at solving the problem that line segment featureextraction and matching work is difficult due to the fact that the positions and postures of image viewpoints are changed greatly in the wide baseline photography process. A stereo matching homonymous image point and line segment feature detection algorithm is taken as a research basis, line segment feature extraction, description and stereo matching are taken as a main research line, and globaland local multi-level complementary information fusion constraints are taken as research ideas. The method comprises three steps of extracting global and local high-quality line segment features, constructing global and local complementary descriptors and matching measures, and performing global and local information fusion constraint-based line segment feature multi-level iterative matching, so as to improve the accurate line segment feature matching rate. And the perfection and development of a wide baseline photogrammetry theoretical system are promoted.

Description

technical field [0001] The invention relates to the relevant technical field of digital image processing in photogrammetry and remote sensing, in particular to a method for line segment feature matching of aerial wide baseline stereo image pairs constrained by global and local fusion. Background technique [0002] In recent years, with the acceleration of the global informatization process, the construction of digital regions and even smart regions is in the ascendant. At the same time, people's demand for fine, realistic and current 3D modeling is increasing, which promotes the development and application of high-inclination aerial photogrammetry technology. . Aerial oblique photogrammetry can obtain high-resolution images of ground scenes from different angles. It can not only provide rich texture, spectrum, shape, and context information for true 3D reconstruction, but also has the advantages of good spatial geometric configuration conditions and reconstruction accuracy. ...

AI Technical Summary

Problems solved by technology

[0006] (1) A line segment feature matching method based on sparse representation and geometric configuration constraints. This method uses SIFT and Pyramid Histogram of Oriented Gradients (PHOG) to jointly describe the line segment to be matched and its neighborhood, and then according to the correlation The coefficient peak value is used to determine the line segment with the same name, which improves the success rate of line segment feature matching, but the effect on sparse texture images is not good

[0007] (2) For high-resolution sparse texture images, some scholars proposed an affine invariant line segment feature matching algorithm. This method adopts the idea of ​​line classification and matching in stages, that is, firstly uses affine invariance to realize the matching of salient line segment features. On this basis, the remaining line features are divided into matched features according to the principle of spatial proximity, and finally the matching discrimination of all line segment features is completed, but the reliability of the algorithm depends on the initial feature matching

[0008] (3)

Images