Affiliations: Computer Engineering Technical College, Artificial Intelligence College, Guangdong Polytechnic College of Science and Technology, Zhuhai, Guangdong, China
Correspondence:
[*]
Corresponding author: Jun Yu, Computer Engineering Technical College (Artificial Intelligence College), Guangdong Polytechnic College of Science and Technology, Zhuhai 519090, Guangdong, China. E-mail: 20374937@qq.com.
Abstract: Aiming at the problems of high feature mismatch rate and low matching efficiency in the current virtual image multi feature matching algorithm, a virtual image multi feature matching algorithm based on 3D scene reconstruction is proposed. Firstly, in the 3D scene reconstruction, the virtual image is preprocessed to eliminate the noise in the virtual image, avoid the noise interference in the compression process, and improve the signal-to-noise ratio of the image; Secondly, the de-noising virtual image is enhanced to enhance the details of the image; Finally, the corresponding information feature vector is constructed from the information features extracted from the virtual image, and the virtual image multi feature matching algorithm is completed. Experiments show that the multi feature matching rate of the designed algorithm is high, the error matching rate is low, the maximum spatial distortion rate is only 0.6%, and the compressed image quality and matching performance are good.
Keywords: 3D scene reconstruction, virtual image, multi feature matching algorithm, signal to noise ratio, spatial distortion rate
