运用面向对象的分类方法对实验区高分辨率影像进行最优分割尺度的影像分割。首先利用ESP算法确定影像的最佳分割尺度,再运用CART决策树分类对土地利用信息进行提取,并对分类结果进行后处理,进一步提高分类精度。最终分类结果精度达到80.38%。

We used the object-oriented classification method to segment experimental high resolution images by the optimal segmentation scale in this paper. We used ESP algorithm to determine the image optimal segmentation scale at first. And then, we used CART decision tree classification to extract land use information, and carried on the post-processing of CART decision tree classification results to improve the classification accuracy. The final classification result accuracy is 80.38%.