基于Level Set的海岸线检测

摘要

Choosing coastline as landmark is an important automatic image-based method for satellite navigation .The method selects a specific area in the remote sensing image ,detects the coastline in it ,and matches it with the corresponding region of the standard global coastline template library .The calculated offset determines how to adjust the posture of the satellite .Considering the characteristics of the automatic image navigation process ,we propose a method to detect land and sea boundary in infrared channel of remote sensing images .The method ,which is based on level set ,is initialized by the standard template library and is equipped with region-scalable fitting energy model .The experimental results indicate that our algorithm has advantages in efficiency ,stability and accuracy .%采用海岸线做为地标进行静止卫星导航是一种重要的遥感图像自动导航方法，此方法在静止卫星遥感图像中选取特定区域，检测其中的海陆分界线，与全球海岸线标准模板库的相应区域进行匹配，依据计算获得的偏移量调整卫星姿态。根据静止卫星图像自动导航过程的特点，本文提出了一种适用于红外通道的基于水平集算法的海陆分界线检测方法，本方法依据地标模板库进行初始化，采用一种可扩展区域能量泛函推动曲线进行演化，直至收敛。实验结果表明，本方法具有迭代次数少，稳定性好，精度高的特点。