The evaluation method study of wheat-planting area supported by spectral database

The traditional evaluation method adopts maximum likelihood algorithm, minimum distance algorithm and spectrum angle match (SAM) algorithm to classify a remote sensing image, and then estimates wheat-planting area through computing pixels. This study puts forward a modified spectrum angle match (MSAM) algorithm combined with DEM to evaluate wheat-planting area. SAM algorithm only takes similarity of whole spectrum shape into account. Based on that, the authors introduces the MSAM algorithm. Because MSAM also covers the reflectance value, it adapts to identifying more detailed land surface object e.g. wheat. The effect of the mountain region is wiped off by DEM first; then two sample data groups of land surface measure are selected from spectral database, one of which is used to train the threshold of MSAM algorithm and SAM algorithm; the other is used to estimate partly the accuracy of both algorithms. The MSAM algorithm is superior to the latter one after comparing the result images and improves the accuracy of wheat-planting area evaluation