Multi-object Recognition Algorithm based on Euclidean Feature Match in Fuzzy Remote Sensing Images

2018 2nd annual International Conference on Cloud Technology and Communication Engineering

计算机科学;无线电电子学

Huang, Fan^1

College of Marine Sciences, Shanghai Ocean University, No.999, Huchenghuan Road, Lingang New City Shanghai

201306, China^1

关键词: Feature matching;    Non-contact sensing;    Parallel recognition;    Recognition algorithm;    Remote sensing image processing;    Remote sensing images;    Segmentation precision;    Target recognition;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/466/1/012100/pdf
DOI  :  10.1088/1757-899X/466/1/012100

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

Remote sensing image target recognition is the use of artificial intelligence to distance non-contact sensing and remote sensing image recognition technology for surface covering, realizes the human of large scale surface macroscopic observation and monitoring for special purposes, to promote effective sense the state of human beings on the earth. Current remote sensing image target recognition technology mainly face the problem of low image segmentation precision and mixed primitive mass, this paper presents Multi-object recognition algorithm based on Euclidean Feature Match in fuzzy remote sensing images, firstly studied the relevant technology of the remote sensing image acquisition and image processing, analysis the current problems existing in the remote sensing image processing. By innovating the multi-objective recognition algorithm of fuzzy remote sensing image, the design of Euclidean feature matching operator is designed, and the manifold clustering processing is carried out for the feature of fuzzy image, so as to realize the optimization of multi-objective parallel recognition. The algorithm in this paper is validated by ERDAS simulation for target recognition accuracy and target recognition complexity performance.