The Research of Building Earthquake Damage Object-Oriented Segmentation Based on Multi Feature Combination with Remote Sensing Image

• Published in: Procedia computer science 2019-01, Vol.154, p.817-823
• Main Authors: Yan, Zhao, Sheng, Cao De, Zhong, Ren Hua
• Format: Article
• Language: English
• Subjects: Fuzzy-based Segmentation Parameter optimizer; Multi-feature joint segmentation; Object-oriented; texture information
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2019.06.077

A multi feature combined remote sensing image segmentation method is proposed to solve the problems of insufficient use of the feature information of the existing remote sensing image segmentation objects, time consuming and excessive dependence on the scale parameters. Through the improved fast scanning algorithm (FSAM), the initialization of the over segmented primitives is constructed, and the texture features are introduced into the image spectrum and shape features to measure the heterogeneity of each region. A fuzzy logic analysis method is used to exercise supervised training based on the characteristic indexes of selected target segmentation samples, and the optimal segmentation parameters are calculated by automatic iteration. In order to verify the effectiveness of the segmentation algorithm, a typical disaster area in Bam area is selected for verification. The experimental results show that the segmentation results are more accurate, the contour boundary of the region is relatively smooth and compact, and the combination of texture regions is more consistent with the cognition and habit of human eyes. It can effectively improve the quality of the segmentation, and can better realize the intelligent interpretation of remote sensing image.