在大面积农作物遥感估产中, 应用气象卫星数据估算种植面积一直是一个难点。已有的混合像元分解法要么实际操作困难, 要么不适用于AVHRR数据。该文在前人研究的基础上提出了一种新的方法──应用AVHRR混合像元神经网络分解模型估算种植面积。这种方法综合利用了TM数据与NOAA数据各自的优势, 具有较好的科学性和经济可行性。

In large area crops yield estimation from remote sensing data, crops planting area estimation with NOAA data remains to be resolved difficultly. Those mixed pixed decomposition methods given before are impractiable, or not suitable for AVHRR data. In view of points above, a new method, more scientific and economic than before, is put forward, that is to estimate crops planting area by means of AVHRR mixed pixel neural network decomposition model, which comprehensively makes use of the advantages of both TM data and NOAA data.