基于ASAR数据的水稻制图最佳时相参数提取  

Optimal Temporal Parameters for Rice Mapping Based on ASAR Data

作　　者：杨沈斌[1] 赵小艳[1] 申双和[1] 李秉柏[2] 谭炳香[3] 

机构地区：[1]南京信息工程大学江苏省气象灾害重点实验室 [2]江苏省农业科学院农业资源与环境研究所 [3]中国林业科学院资源信息研究所

出　　处：《江苏农业学报》2009年第3期518-523,共6页Jiangsu Journal of Agricultural Sciences

基　　金：2007年江苏省高校研究生科技创新计划项目(CX07B_048z)

摘　　要：针对ASAR交叉极化模式数据,选用HH/VV极化比值提取水稻制图的最佳时相参数。结合SAR图像统计特性,采用比值变化检测方法,通过理论公式计算和比较变化检测的总误差概率,分别提取单时相双极化和多时相双极化水稻制图的最佳时相参数。结果显示,单时相制图的最佳时相是水稻生长中后期,即水稻幼穗分化期至水稻收割前期内任一时相;而多时相制图的最佳时相为一组时相的组合,即应至少包含水稻移栽期和水稻生长中后期各一时相。最后,采用阈值分类算法对最佳时相参数进行了验证。结果显示,选用最佳时相影像和理论推导的分类阈值能够显著提高水稻制图的精度。Based on the statistical characteristic of ASAR images, a method of ratio change detection was adopted to retrieve optimal temporal parameters for rice mapping using HH/W. The optimal temporal parameters were obtained using both single-temporal multi-polarization ASAR data and multi-temporal and multi-polarization ASAR data, respectively. The result showed that the optimal temporal parameter for single-temporal rice mapping could be selected from the later stage of rice growth, i.e. from rice spike differentiation period to the period before rice harvest, while the optimal parameter for muhi-temporal rice mapping was a group of dates, which at least included the rice transplanting stage and the later stage. Validation was conducted using threshold classification method. The result indicated that the rice mapping accuracy could be improved significantly using the optimal temporal parameters with a derived classification threshold.

关 键 词：ASAR 水稻 比值变化检测 图像分类 

分 类 号：S511.01[农业科学—作物学]