A Heterogeneous Image Matching Method Based on Line Segment

Abstract

The present invention relates to the technical field of image processing, and provides a heterogeneous image matching method based on straight line segments, which extract line features in a reference image and a perception image respectively, and use them as primitives for heterogeneous image matching; construct line segment endpoints and endpoints and the neighborhood relationship between endpoints and line segments, and connect the endpoints and line segments with neighborhood relationships to obtain straight line segments with better stability, and construct a line segment set that accurately represents the scene structure; in the reference image line segment set and Construct line segment pairs between the perceptual image line segment sets, and calculate possible transformation parameters; select the transformation parameters that make the perceptual image and the reference image have the highest similarity, and realize the matching between the perceptual image and the reference image. The present invention can effectively avoid the double-sided phenomenon when extracting road straight lines from aerial images, greatly reduces the number of line segments involved in matching, and can use as few line segments as possible to represent the geometric structure of the scene; the implementation is simple, the calculation speed is fast, and it can be quantified analyze.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a heterogeneous image matching method based on straight line segments. Background technique [0002] Scene matching technology has important applications in the field of autonomous navigation of UAVs. In order to realize the all-weather autonomous navigation of UAVs, a feasible technology is to use infrared cameras or SAR radars to image the ground, and then match the captured real-time images with the pre-prepared satellite reference images to obtain the UAV’s location and orientation information. [0003] Due to the difference in shooting time, angle and imaging sensor, the same-name points on the real-time image and the reference image do not have a simple and consistent functional relationship in gray value, so the matching method for homogeneous images is generally not suitable for heterogeneous images. The current typical heterogeneous image matching meth...

AI Technical Summary

Problems solved by technology

[0004] The region-based heterogeneous image matching method is widely used in the field of medical image processing, which has the advantages of high matching accuracy, but the time efficiency is low, and often only the translation parameters can be estimated, but the rotation parameters cannot be estimated.

Images