摘要
遥感分类的精度决定了遥感数据处理后的可信度,低分辨率的遥感数据连续性强、价格低廉,但在目视解译方面较为困难,进行野外验证时成本大、周期长、分类精度难以确定,所以低分辨率植被类型的分类结果的精度估测问题并没有很好地得到解决.本文提出了利用高分辨率的遥感数据验证低分辨率遥感数据的方法,在对低分辨率NOAA数据进行分类之后,寻找到同一地区同一时期的30 m分辨率的TM数据,利用目视解译方法对TM数据进行了同类型的植被分类.经过尺度变换后,利用TM分类结果对NOAA数据进行了精度验证,在一定程度上解决了研究区域的空间尺度较大时分类精度难以确定,精度检验效率低下的问题.同时指出,高分辨率的遥感数据不仅可以进行大尺度植被观测的精度检验,同时可以进行大尺度植被观测的分类校正.
It is a normol method for us by using remote sensing data to observe the change of vegetation. The precision of classification of remote sensing data affects reliability, the advantage of low spatial resolution data is continuous period of time and free money, but that it is difficult to classify remotely sensed data with low spatial resolution by mam-eye . After classifying these data, and these data is often used to research the change of regional scale,. So the estimate of classification precision of low spatial resolution is still difficult problem. In this paper, we use high spatial resolution data to validate the low date. Finding NOAA（National Oceanic and Atmospheric Administra-tion） image and landsat-TM（Thematic Mapper） image at the same area and same year, classify vegetation types from the NOAA image by dividing area mothed and classify same vegetation types from the TM images by man-eye, after transforming scale for TM data as NOAA data resolution, compare the results of NOAA classification data with TM, precision of one sample area is 0.804348, another sample is 0.793103, the classification precision is near 80%, make sure the precision of classification NOAA data. This method solves the problem of classification precision when we research large region scale.
出处
《新疆大学学报（自然科学版）》
CAS
2013年第2期238-242,共5页
Journal of Xinjiang University(Natural Science Edition)
基金
国家自然科学基金(31160114)
关键词
精度分析
高空间分辨率数据
低空间分辨率数据
多尺度遥感数据
validate classification precision, low spatial resolution data, high spatial resolution data multi-scale remote sensing data
