基于DWT-SVD影像处理的河流骨架线提取  

Extraction scheme for river skeleton line based on DWT-SVD

作　　者：赵威成 马亚辉 郑甲伟 Zhao Weicheng;Ma Yanhui;Zheng Jiawei(College of Mining Engineering ,Heilongjiang University of Science and Technology,Harbin 150022,China;School of Forestry,Northeast Forestry University,Harbin 150022,China;China Petroleum Engineering & Construction Corp.north China company,Renqiu 062550,China;China Construction First Bureau Huajiang Construction Co.,Ltd.,Beijing 100161,China)

机构地区：[1]黑龙江科技大学矿业工程学院,黑龙江哈尔滨150022 [2]东北林业大学林学院,黑龙江哈尔滨150040 [3]中国石油工程建设有限公司华北分公司,河北任丘062550 [4]中建一局华江建设有限公司,北京100161

出　　处：《矿山测量》2019年第3期121-123,127,共4页Mine Surveying

摘　　要：河流骨架线是一种重要的基础参考数据,有着十分重要的意义。目前,提取河流骨架线的方法常常会出现断裂,植被等非水体信息也没有得到有效的抑制。本文提出利用DWT-SVD对低反差原始影像进行增强处理,提高NDWI和INDVI的特征提取精度,利用INDVI对NDWI进行非水体信息影响抑制和水体信息增益,使得水体信息提取更为准确,利用数学形态学约束河流二值图像,最终提取河流骨架线。实验表明,该方法提取的河流骨架线更符合河流的真实位置与形状,具有更高的准确性。River skeleton line was an important basic reference data,which had the great significance. At present,the method of extracting river skeleton line often appeared fracture,and the non-water information such as vegetation had not been effectively suppressed. In order to improve the accuracy of NDWI and INDVI feature extraction,DWT-SVD was used to enhance the original image with low contrast. INDVI was used to suppress the influence of non-water body information on NDWI and increased the information gain of water body,which made the extraction of water body information more accurate. Using mathematical morphology to constrain river binary images,and finally river skeleton lines was extracted. Experiments showed that the skeleton line extracted by this method was more met the true position and shape of the river and had higher accuracy.

关 键 词：DWT-SVD NDWI INDVI 数字形态学 

分 类 号：TD672[矿业工程—矿山机电]