基于数学形态学的遥感影像城市道路提取研究

Study on the urban road extraction from remote sensing image based on mathematical morphology

以内蒙古自治区乌海市海勃湾区为研究区,通过选取合理的结构元素,构建了基于数学形态学的遥感影像城市道路提取模型。首先,对遥感影像进行二值化处理;其次,用Hough变换检测出道路方向;再次,通过分析城市道路的几何特征,在二值图像上,基于Hough变换检测出的道路方向选定线性结构元素,构建了基于数学形态学的城市道路提取模型,提取了乌海市海勃湾区主干道的道路信息;最后,经过与乌海市海勃湾区城市道路局部遥感影像作对比,分析道路提取结果,得出结论。结果表明,道路提取精确度达到98.77%,遗漏误差1.23%,冗余误差只有0.02%,说明本文提出的模型能有效地提取直线型道路与弯曲型道路,且基本不受噪声的影响,取得了较好的效果。

Taking Haibowan District of Wuhai City,Inner Mongolia Autonomous Region as the research area,the model of urban road extraction from remote sensing image based on mathematical morphology is constructed by selecting reasonable structural elements.Firstly,the remote sensing is binary processed.Secondly,the road direction is detected by using Hough transform.Thirdly,with the analysis of the geometric characteristics of urban roads,and the selection of linear structural elements based on the road direction detected by Hough transform on the binary image,and the urban road extraction model based on mathematical morphology is constructed to extract the road information of the main road in Haibowan District of Wuhai City.Finally,the results of road extraction are analyzed and a conclusion is drawn after comparing with local remote sensing images of urban roads in Haibowan District of Wuhai City.The results show that the accuracy of road extraction is 98.77%,the omission error is 1.23%,and the redundancy error is only 0.02%,which indicates that the model proposed in this paper can extract the straight road and curved road effectively,is basically not affected by noise.