基于灰度共生矩阵法的IKONOS影像城市草地信息提取

Extraction of urban grassland information from IDONOS image based on grey level co-coccurrence matrix

对IKONOS影像图上提取草地信息进行探索:用基于灰度共生矩阵的纹理量的分类法,通过从近红外波段提取灰度共生矩阵和灰度联合矩阵,计算并提取理想窗口的最能反映类别差异的纹理量值,实验发现取3X3窗口的CON纹理量可以较好提取出城市草地信息,通过精度评估发现具有较高的精度(平均精度达90.56%),具有较好的适应性。

The algorithm of extracting city grassland information from high resolution images is devised. The value of texture quantity for suitable window which can best reflect the classification difference is calculated and extracted using the classification method of texture quantity based on grey level co-occurrence matrix. The experiment shows that the way can better extract urban grassland information, and the precision is higher.