稻棉轮作周期监测样区的最优选择方法  

Optimal selection of sample areas for monitoring cotton-rice rotation periods

作　　者：孙玲[1] 朱泽生[2] 

机构地区：[1]江苏省农业科学院农业经济与信息研究所,江苏南京210014 [2]海军指挥学院理论研究系,江苏南京210014

出　　处：《江苏农业学报》2012年第5期1189-1193,共5页Jiangsu Journal of Agricultural Sciences

基　　金：江苏省农业三新工程项目[SX(2011)392;SXC(2012)410]

摘　　要：以水稻、棉花和TM卫星遥感影像为材料,在江苏省兴化市进行稻棉轮作周期监测样区最优选择试验,在限定监测费用,并最大限度提高监测精度的条件下,研究了候选样区的水稻面积、棉花面积和地理空间位置对稻棉轮作周期监测样区最优选择的影响。结果表明:运用基于线性回归和整数规划的代表性抽样模型,在没有云层干扰的卫星遥感影像中,对由38个乡镇所构成的候选样区总体进行最优抽样,选择了其中6个乡镇或候选样区作为最优监测样区总体;估算的候选样区总体和最优监测样区总体的稻棉轮作周期分别为2.19年和2.11年;与候选样区相比较,最优监测样区总体估算精度损失了3.78%,但监测费用减少了84.21%。Field experiment in Xinhua city of Jiangsu province was conducted by using rice,cotton,and satellite remotely sensed images,under the conditions of monitoring cost constraint and monitoring error minimization,to examine the effects of rice area,cotton area,and geographic location of candidate sample areas on optimal selection of sample areas for monitoring cotton-rice rotation periods.The results showed that,with an representative sampling model based on linear regression and integer programming,six towns were selected as a population of optimal sample areas for monitoring the cotton-rice rotation period in Xinhua city out of 38 towns which served as a population of candidate sample areas on satellite remotely sensed cloudless images.The cotton-rice rotation periods of populations of candidate sample areas and optimal sample areas were respectively 2.19 and 2.11 years.In comparison with the population of candidate sample areas,the monitoring cost of the population of optimal sample areas decreased by 84.21%,while the monitoring error of the population of optimal sample areas increased by 3.78%.

关 键 词：稻棉轮作周期监测 费用约束 误差最小化 最优监测样区 线性回归 整数规划 

分 类 号：P237[天文地球—摄影测量与遥感]