Multi-view intensive matching method for tilted images with invariant size

Abstract

The invention discloses a multi-view intensive matching method for tilted images with invariant size. Since tilted images have the problems that size differences and viewing angle differences are large and the images are shielded seriously, great difficulty is caused to intensive matching of the images. According to a provided multi-view image matching algorithm, a path for extracting candidate matching points first and then conducting verification of the candidate matching points is adopted, and a transverse bilateral matching strategy, a pyramid strategy and the like are combined to ensure the reliability of the matching points. Matching windows conduct resampling along epipolar lines, and the problem that the viewing angles of the tilted images are changed is solved. When the matching windows are sampled along the epipolar lines, the ground resolution ratios of the current images on object points are combined to conduct scaling adjustment on the sampled windows, and the problem of size differences of the tilted images is solved.

Description

technical field [0001] The invention relates to the field of surveying and mapping science and technology, in particular to a scale-invariant multi-view dense matching method for oblique images. Background technique [0002] In traditional aerial photogrammetry, the products such as DEM / DSM produced are 2.5-dimensional, which cannot express the information of the position of the side of the building. With the increasing demand for high-precision city 3D models in digital cities, smart cities and other application fields, oblique photography technology has developed rapidly. [0003] Oblique photography is to increase the camera equipment in the direction of oblique view on the basis of traditional vertical aerial photography. The most common is the five-lens camera system, which is to shoot forward, backward, left, right and downward at the same time. The side-view camera of oblique photography can obtain information such as the sides of buildings that are difficult to capt...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a scale-invariant multi-view dense matching method for oblique images, which overcomes the difficulties caused by the dense matching of oblique images due to large scale differences, large differences in viewing angles, and a large number of occluded areas, and can be applied to high-precision City 3D Modeling

Images