摘要
为克服样方代表性狭窄、误差大、需重复采样等缺点,提出了基于地块窗口的估产模式,以行政村作为研究区域,利用QuickBird影像提取的地块区域作为窗口,TM影像为主要数据源,结合实地调查,分别以玉米和花生为研究对象,在地块、区域2个层面初步探讨了在地块窗口支持下遥感估产的关键技术。结果表明:采用以地块为窗口的分类方法在大比例尺下作物的提取能获得较好的结果;归一化植被指数NDVI与单产间呈线性关系,且为显著正相关;地块窗口支持下,以TM影像为主要数据源,结合实地调查等方法,是一种比较适合我国地块破碎、种植类型复杂的南方地区的遥感估产模式。
To overcome the drawbacks of stricture representativeness of quadrat,great error,resampling,the mode of yield estimation based on field window was proposed.Based on mechanism of remote sensing on crops using administrative village as survey region,the field extracted from QuickBird image and TM image was used as main data.Combining with the methods of field survey,a preliminary discussion about key technology of remote sensing yield estimation based on field window from two aspects of sampling field and region of corn and peanut was made.This study was attempted to extract the crops on large scale,considering that it would have better results of crops using classification supported by field window.It also showed that there was a linear relationship and significant positive correlation between NDVI and yield per unit area.The method using the TM image supported by field window and combining with field survey is considered to be a suitable mode of yield estimation for the southern China with disconnected field and complex plantation type.
出处
《华中农业大学学报》
CAS
CSCD
北大核心
2011年第2期206-209,共4页
Journal of Huazhong Agricultural University
基金
国家科技支撑计划项目(2006BAD101304)
华中农业大学人才启动基金项目
