基于色彩和纹理特征的建设用地遥感专题信息自动提取方法

Approach of Extracting RS Thematic Information of Construction Land Based on the Features of Color and Texture

提出了一种基于航空影像的建设用地信息自动提取方法.针对航空影像光谱信息少,色彩信息和纹理信息丰富的特点,该方法首先采用HSV色彩变换和纹理分析手段充分挖掘影像中所包含的色彩信息与纹理信息,在此基础上采用多特征阈值分割技术将色彩信息与纹理信息有机结合进行建设用地信息提取,并使用邻域分析方法对提取结果进行修正.通过对研究区域航空影像的处理,结果表明,该方法提取精度较高,且易于实现.

This paper proposes an automatic method for extraction of construction land information from aerial images.Taking into consideration the fact that aerial images are characterized by inadequate spectral information and rich color information and texture information,this method adopts HSV transformation and texture analysis to sufficiently mine the color and texture information in the aerial images,and then uses the multi-feature threshold segmentation technology to combine color information and texture information for construction land information extraction.Finally it uses neighbor analysis to amend the results.Through practical projects in image processing,the result shows that the method has high extraction precision and is easy to implement.