A Study on Yellow River Mainstream Line Feature Extraction
Mainstream line is significant for the Yellow River situation (regime) forecasting and flood control. An effective mainstream line feature extraction method is proposed in this paper. Considering the characteristics of tiny spectrum discrepancy between the area of mainstream and non-mainstream, the conversion method of using between-class scattering degree as projection index is proposed. After maximizing the between-class discrepancy to obtain effective information components for classification, skewness coefficient is utilized to describe the feature of mainstream line based on the statistical distribution discrepancy indicating in the components of maximum between-class scatter for the area of mainstream and non-mainstream. Experimental results indicate the effectivity of the proposed approach in mainstream line detection. The results of case computation indicate the proposed approach can extract the location of mainstream line automatically with higher accuracy.
remote sensing feature extraction mainstream line between-class scatter matrix skewness coefficient
Zhang Haichao Duan Feng Zhang Yanning Liu Xuegong Han Lin
Institute of Computer, Northwestern Polytechnical Univ ersity, Xian, 710072, China Info rmation Center of Yellow River Conservancy Commission, Zhengzhou, 450004, China Institute of Computer, Northwestern Polytechnical Univ ersity, Xian, 710072, China Info rmation Cen
国际会议
山东东营
英文
68-72
2007-10-16（万方平台首次上网日期，不代表论文的发表时间）