Feature Descriptor Determination Method, Device, and Computer-Readable Storage Medium

Abstract

The embodiment of the present disclosure discloses a feature descriptor determination method, device and computer-readable storage medium. The method includes: obtaining the normal vector of the feature point on the first image; determining the local coordinate system according to the normal vector of the feature point; wherein, the normal vector of the feature point represents the Z-axis direction of the local coordinate system; The first image block composed of the feature points and the neighborhood point sets of the feature points, and on the plane determined by the X axis and the Y axis of the local coordinate system, determine the real space sampling area used to simulate the first image block corresponding to The front view of the target; according to the first image, the local coordinate system and the front view of the target, the feature descriptors of the feature points are determined. The embodiments of the present disclosure can guarantee the reliability of the feature descriptors of the determined feature points, and when using the feature descriptors to perform image feature matching, the reliability of the matching results can also be effectively guaranteed.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, and in particular to a method, device and computer-readable storage medium for determining a feature descriptor. Background technique [0002] In the technical field of image processing, image feature matching is required in some cases. For example, when there is a need to identify whether the characters in two images are the same person, image feature matching needs to be performed. In this case, the feature descriptors of the feature points on the two images may be determined respectively, and then the feature descriptors of the feature points on the two images may be compared, so as to realize image feature matching based on the comparison result. [0003] At present, due to various reasons (such as different shooting angles), image features will undergo different deformations in different directions, which will affect the reliability of the feature descriptors of the determi...

AI Technical Summary

Problems solved by technology

[0003] At present, due to various reasons (such as different shooting angles), image features will undergo different deformations in different directions, which will affect the reliability of the feature descriptors of the determined feature points, thereby affecting the image feature matching. , the reliability of the matching result

Images