论文部分内容阅读
针对面向对象的遥感影像分析技术这一热点问题,通过研究滑坡的地学特征和影像特征,基于QuickBird多光谱影像,提出了一种先对Quickbird影像进行多尺度分割,然后利用改进的分割质量评价函数来选择滑坡最优分割尺度,最后在最优分割尺度上构建滑坡提取规则集并进行滑坡信息提取的方法。该方法通过对提取的滑坡信息进行精度验证与分析,能使对象内部异质性和对象之间异质性达到综合效果最好。结果表明:滑坡体的正确提取率为75.86%,本文方法具有一定的可靠性和准确性。
Aiming at the hot issue of object-oriented remote sensing image analysis technology, this paper proposes a method of multiscale segmentation of Quickbird image based on QuickBird multispectral imagery by studying geosciences and image features of landslide. Then, by using improved segmentation quality evaluation function To select the optimal segmentation scale of landslide, and finally to build the rule set of landslide extraction and the method of landslide information extraction on the optimal segmentation scale. The method can verify and analyze the extracted landslide information with accuracy, which can make the heterogeneity within the object and the heterogeneity of the object reach the best combined effect. The results show that the correct extraction rate of landslide body is 75.86%. The method of this paper has certain reliability and accuracy.