Remote Sensing Image Texture Feature Extraction And Quantitative Analysis

With China's sustained rapid economic development and the process of urbanization continues to accelerate, dynamic detecting and automatic updating technology in land changes become increasingly important. Remote sensing technology is intergrited with computer technology, satellite technology, and earth knowledge to carry out observation and monitoring of the earth. The development of computer technology, high spatial resolution images in remote sensing make detecting dynamic changes in land and automatic updates possible.Traditional pixel-based remote sensing image processing methods is not suitable for richly detailed high-resolution remote sensing image processing technology. Therefore an object-oriented remote sensing image processing method has been carried out. In the object-oriented approach, from high-resolution remote sensing images of space objects from the analysis of texture feature is a key factor of intelligence of remote sensing processing technology. Remote sensing research Group at Shanghai Jiaotong University designed and Implementationed an object-oriented remote sensing image processing system - "Translation landing ELU", and updated to version 2.0. The software can process high spatial resolution remote sensing image segmentation, feature extraction, classification, monitoring changes in lands well, buiding a new breakthroughsengage in remote sensing image processing. In this paper, in conjunction with the project, feature extraction methods and analysis in high-resolution remote sensing images are introduced, and desigened and implementated Feature Extraction Module in the "translation landing ELU" platform.Combining segmentation and classification of remote sensing image processing applications background, the design of the high spatial resolution remote sensing image feature extraction module, a more in-depth discussion and analysis of a variety of feature extraction method is the main contribution of this paperFull support for remote sensing image automatic Features selection, multi-classifier automatically classified integration capabilities, enhance remote sensing processing platform automation, intelligent level is the mainly direction of research.