南方植被区遥感线性构造自动提取——以湖南1∶5万官庄幅遥感影像为例

Automatic extraction of remote sensing linear structure in South vegetation area: An example of 1∶50 000 Guanzhuang remote sensing image in Hunan

为提高遥感线性构造提取效率,针对植被覆盖会对线性构造提取造成很大干扰的因素,提出了一种自动提取线性构造的方法。首先,利用改进型强迫不变技术进行植被抑制处理;在此基础上,选取最优的波段组合进行主成分分析得到最能反映该地区地质构造空间展布特征的影像波段;基于定向滤波技术和Canny算子边缘检测技术,对此波段自动提取遥感线性构造;最后,采用地质统计学理论和目视解译对所提取线性构造进行分析。结果表明,在提取遥感线性构造特征时,运用自动提取线性构造方法所处理获得的遥感影像结合对比1∶5万官庄幅遥感影像目视线性构造解译图成果,对相关线性构造信息进行综合分析、筛选,有利于提高线性构造解译的精确度。

In order to improve the extraction efficiency of remote sensing linear structures, a method for automatic extraction of linear structures is proposed for the factor that vegetation cover will cause a great interference to the extraction of linear structures.Firstly, the improved forced invariance technology is used to suppress the vegetation.On this basis, the optimal band combination is selected for principal component analysis to obtain the image band that can best reflect the spatial distribution characteristics of the geological structure in the region.The directional filtering technology and Canny operator edge detection technology are used to automatically extract the remote sensing linear structure in this band.The extracted linear structures are analyzed by geostatistics theory and visual interpretation.The result shows that in the extraction of the features of remote sensing linear structures, the remote sensing images processed by the automatic linear structure extraction method are combined with the visual linear structure interpretation map of 1∶50 000 Guanzhuang remote sensing image, and the comprehensive analysis and screening of relevant linear structure information are conducive to improving the accuracy of linear structure interpretation.