基于波谱分析技术的遥感作物分类方法

Crop classification by remote sensing based on spectral analysis

摘要: 为获取东北三省作物类型分布信息，精确地进行粮食估产，该文以250 m MODIS时间序列NDVI数据为主要数据源，以东北三省主要粮食作物水稻、玉米、大豆、小麦为研究对象，利用波谱分析方法对东北三省作物类型的空间分布进行研究。研究结果表明，大豆的遥感反演面积和统计面积的相关性最好(R2=0.770)，其次是玉米(R2=0.710)，水稻(R2=0.686)。该文使用的作物分类方法适用于试验条件有限，实测数据较难获得并以遥感数据为主要数据源且研究区域较大、作物类型单一、种植面积广的情况。

Abstract: In order to acquire the information of crop classification to estimate crop yield accurately in Northeast China, four kinds of crops (rice, wheat, maize, soybean) were taken as study objects and 250 m MODIS time-series NDVI data was used to analyze the crop distribution patterns based on spectral analysis method. The area derived from the crop classification result was compared with the planted area from statistical data，and the results showed that the correlation of soybeans was better than maize and rice, with R2=0.770, 0.710, 0.686 respectively. The crop classification method used in the study is suitable for the situation with limited experimental conditions, difficult for obtaining measurements (remote sensing data as the main data source) and large planting area with single crop.