基于数学形态学的SPOT影像桥梁提取

利用数学形态学的方法对SPOT遥感影像桥梁提取进行了研究。首先运用二维特征模糊闺值分割，分别得到包含水体信息和道路桥梁信息的二值影像；进一步使用形态学开闭运算去除细小噪声，使用像素区域标记法提取出面积较大的独立水体，并通过膨胀与腐蚀运算得到连通水体；最后辅以相关知识，采用Hough变换提取出桥梁的边界。实验结果表明，该方法快速有效，识别准确率较高。

Bridge Extraction from SPOT Image Based on Mathematic Morphology

An approach to extract bridge from the SPOT images based on mathematic morphology is presented in this paper. Water area and road area binary images are gained according to the 2-dimensional fuzzy segmentation algorithm. Then tiny noises are removed with morphological opening and closing. Meanwhile, important separate water areas are extracted and linked by region-labeling, dilating and eroding methods. Finally, the boundaries of bridges are recognized with related knowledge and Hough transform. The experiment results prove the high efficiency and recognizing rate of this algorithm.