中文摘要：

摘要：静止卫星采用遥感图像中的地标进行自动导航，海岸线的检测是图像导航的重要一步。将检测结果与全球标准模板库中相应区域的海岸线匹配，依据获得的偏移量调整卫星姿态。根据静止卫星图像导航过程的特点，提出了一种适用于静止卫星红外通道的海陆分界线检测方法，进行云检测并依据地标模板库初始化后，采用C-V模型推动曲线进行演化。实验结果表明，此方法针对静止卫星红外通道的特征，不仅具有计算量小、精度高的特点，而且具有较好的稳定性和普适性。

英文摘要：

Meteorological satellite is using landmarks in remote sensing images for automatic navigation. Coastline detection is an important step in the process. Through matching the detection result with the corresponding region o{ global standard landmark library, the offset is obtained to adjust the posture of satellite. Considering the characteristics of meteorological satellite image navigation process, a new method is proposed to detect coastline in the infrared channel. Cloud detection is made first, and a C-V model which is initialized by information in standard landmark library is used to promote the evolution of the curve. Experimental results show that the method fits meteorological infrared channel well and has advantages in efficiency, stability and accuracy.