Local invariant gray feature-based image registration method and image processing system

Abstract

The invention belongs to the data recognition and data representation technical field and discloses a local invariant gray feature-based image registration method and an image processing system. According to the local invariant gray feature-based image registration method, feature extraction descriptors are constructed; feature points between registration images are searched, the nearest neighborprinciple is used to find matched key points; and an affine transformation matrix H between the registration images is calculated, and six parameters of the affine transformation matrix H are obtainedthrough singular value decomposition. The descriptors are constructed; sampling points are divided into an odd part and an even part, and therefore, dimensionality during the construction of the descriptors is significantly lowered, operating time is reduced, the accuracy and accuracy of registration are improved; when being constructed, descriptor vectors are sequenced according to gray values,and therefore, rotation invariance can be realized. The method of the invention has high detection precision, good noise robustness and low computational complexity, which mainly benefits from the great reduction of the dimensionality of the original descriptors and insensitiveness to illumination transformation.

Description

technical field [0001] The invention belongs to the technical field of data identification and data representation, and in particular relates to an image registration method and an image processing system based on local invariant gray features. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: Image local features are widely used in computer vision and pattern recognition, which mainly includes two aspects: detecting points of interest or regions of interest from images of interest, and calculating their invariant features. Many methods for detecting points of interest have been proposed, such as SIFT, GLOH, IWCS-LTP and so on. They construct descriptors by constructing gradient direction and location information histograms, which can achieve satisfactory results. However, these methods cannot cope with more complex orientation changes and lighting transformations. The existing technology is a local gray-scal...

AI Technical Summary

Problems solved by technology

[0004] (1) When the SIFT algorithm is looking for invariant features, it needs to establish a coordinate system and assign a main direction to each key point. The direction evaluation is easily disturbed by noise, which affects the matching accuracy.

[0005] (2) The dimension of the local grayscale sorting mode will increase with the increase of sampling points, which will affect the calculation efficiency

[0006] (3) After using RANSAC to purify matching point pairs, there may still be mismatching points, which need to be refined to eliminate errors. At the same time, when calculating the affine transformation matrix H, use three pairs of matching points to calculate all the analog points at once. It is easy to introduce parameter estimation error

Images