摘要
对遥感影像中居民地提取的各种方法作了一般性的阐述,并在吸收这些算法优势的同时,提出了一种全自动实现遥感影像中居民地目标的提取方法:先求取非最大值抑制后的梯度图,而后进行高斯模糊处理,以提高影像上居民地区域纹理特征一致性;通过增强大津法和形态学的方法全方位腐蚀与膨胀,完成面状地物居民地的提取;最后进行图象矢量化处理。论述方法的同时,展示和分析了实验结果。
To remote sensing image for the extraction of residential area of various method made to elaborate generally, and absorbed these calculate ways advantage, at the same time provided a valid one to extract residential characters automatically ： Firstly, the grads map which restrained by not max value is calculated; then gauss blur disposing is applied to improve the coherence of the residential area textures on image; By using morphologic method to erode and expand the textures omni directionally to realize the extraction of the area residential and vectors track.; Carry on the portrait vector turns the processing finally. Discussed the method, displayed and analyzed the result of the experiment.
出处
《地理空间信息》
2008年第1期22-25,共4页
Geospatial Information
关键词
遥感影像
高斯模糊
形态学
矢量跟踪
remote sensing image
Gauss blur
morphologic
vectors track
