摘要
受空间分辨率的约束,MODIS遥感影像中存在很多混合像元,不利于水体边缘信息的提取,为提高水体边缘混合像元的检测精度,本文提出了多结构元素组合型边缘检测算法。方法选取了适合于遥感影像水体边缘检测的膨胀型和腐蚀型算子,同时针对遥感影像自身的特点,构建了多尺寸多方位的结构元素,设计了不同的方向权重和尺寸权重系数,再将基于不同检测算子的边缘检测结果组合得到最终结果。实验结果表明与传统算法相比,多结构元素组合型边缘检测算法边缘提取效果良好,提高了水体边缘的检测精度。
In the MODIS remote sensing images, mixed pixels exist due to the resolution limitation, and it is not conduciw＇, to wa- ter edge information extractiou. In orde,＂ to improve the detection accuracy of water mixed pixels, the paper proposed combined muhi- structure, edge detection algorithm. This method selected the dilation and erosion operators for water edge detection of remote sensing image. Meanwhile, considering the characteristics of remote sensing images, it proposed multi-scale and multi-directional structural el- ements, amt designed different direetiun and scale weight coefficients. The final result was the combination of edge detection results based on different detection operators. The experimental results showed that it would be superior to the traditional algorithm in the as- peel of edge extraction, and the detectiou accuracy of water edge could be improved.
出处
《测绘科学》
CSCD
北大核心
2013年第2期90-92,共3页
Science of Surveying and Mapping
