Gradient binaryzation based rotation-invariant and scale-invariant scene matching method

Abstract

The invention discloses a gradient binaryzation based rotation-invariant and scale-invariant scene matching method, and relates to the field of scene recognition. According to the method, on the basis of a classical binary description BRIEF algorithm in which only gray scale intensity is compared, horizontal and vertical gradient comparison is added, texture information of a sampled area is saved, and accordingly, matching error rate is reduced. Moreover, an image scale pyramid is created, image feature point detection and feature description are performed within different scales, gravity center vector directions are added during descriptor calculation, and direction and scale invariance of binary descriptors is achieved. Experiments show that binaryzation based rotation-invariant gradient sampling descriptors have high robustness, and the matching accuracy rate is 73.06% higher than that of the BRIEF algorithm in average when a scene image is rotated greatly and the scale is varied.

Description

technical field [0001] The invention relates to the field of scene recognition, in particular to a gradient binarization-based rotation scale invariant scene matching method suitable for scene matching. Background technique [0002] Scene matching is often used to search for the same content in two scenes, and has a wide range of applications in the fields of scene recognition and object recognition. [0003] As a specific implementation method of scene matching, image feature point matching realizes the matching and recognition of the current scene by matching the feature points of the current image with the historical images in the database. Therefore, image feature point matching has become the focus of current research. Feature point matching includes three parts: feature point detection, feature point description, and feature point matching. First, search for stable feature points on the two images. These feature points can still be detected by the detector after scale...

AI Technical Summary

Problems solved by technology

Although SURF has greatly improved the calculation speed based on the idea of ​​SIFT algorithm, it still cannot meet the needs of real-time systems.

In order to further reduce the complexity of the algorithm, Calonder et al. proposed the BRIEF[3] descriptor based on binary description, which randomly selects a number of (usually 128, 256 or 512) sampling point pairs in the sampling area around the feature point , use 0 or 1 to represent the gray scale relationship between the sampling point and the two points, and finally form a 128, 256 or 512-dimensional binary feature descriptor. When matching, the Hamming distance is used for calculation, and the calculation speed is greatly improved. Meets the matching requirements for real-time systems, but BRIEF is not rotation invariant due to the lack of orientation normalization for descriptors

Images