摘要
研究了在GIS支持下 ,RadarsatSNB -SAR数据用于广东省大范围水稻种植面积调查的方法与应用。首先 ,对SAR数据预处理 ,包括辐射校正、Frost滤波和几何精校正等 ;然后对预处理影像的后向散射值进行信息定量判断 ,结合GIS的辅助分析 ,剔除不可能是水稻信息的像元 ;最后对剔除非水稻信息后的影像进行最大似然法的有监分类 ,获取各区域水稻分布的影像并统计面积。结果表明 ,该方法对于大范围的水稻种植面积调查具有实用、快速的特点 ,其精度在平原地区具有较高的可靠性 。
A practical method of rice mapping in large area using Radarsat SNB SAR data in Guangdong Province, China is advanced. The SAR data pre processing includes radiometric correction, Frost filter and precise geometric correction, and so on. A quantilative estimation is made on the backscatters of SNB SAR images. With GIS , interfering the information is marked and rejected by the thematic layers for extracting rice information. Finally, Maximum Likelihood Supervised Classification is applied to the images for rice mapping, and the rice area is calculated. The result shows that the method is efficient, rapid and practical in large area. The accuracy is high and credible for plain area. But for mountainous area the method needs to be improved because of the complicated terrain conditions.
出处
《热带地理》
2001年第4期346-349,359,共5页
Tropical Geography
基金
广东省"百项"工程资助项目 (编号 :99B0 5 70 1G)
