Generation method of rotation-invariant multi-scale annular feature descriptor

• Main Authors: LONG MENGPING, LEI CHENG, HUANG JIN, MEI LIYE, HU TAOBO, WANG DU, LI XIAOXIAO
• Format: Patent
• Language: Chinese; English
• Published: 03.02.2023
• Subjects: CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a method for generating a rotation-invariant multi-scale annular feature descriptor, which is used in the field of image recognition to overcome the challenge that a feature recognition technology is difficult to accurately describe key features due to the problem of image rotation. Firstly, a multi-scale annular domain is constructed and used for segmenting local features around feature points, and then multi-scale annular features are obtained; calculating the feature value of the annular feature by taking the annular feature as a basic unit; and finally, sorting the feature values obtained from the plurality of scales to form a feature vector with rotation invariance. According to the feature description scheme provided by the invention, the condition that the feature vector with rotation invariance is obtained by estimating the dominant direction of the feature points is avoided. Compared with a scheme based on the dominant direction, the feature vector provided by the invention is