针对成像型反舰导弹所获红外图像的特点，提出了一种将海天/海岸线检测与目标提取相结合的 舰船目标提取方法。该方法先综合运用Sobel边缘检测和局部Hough变换等数字图像处理技 术从红外图像中检测出海天/海岸线，然后通过对Kapur熵二值化图像进行投影变换实现舰船目标提取。实验结 果表明，该方法能够有效地确定舰船目标在海天/海岸线区域中的位置，可保证目标提取的准确性。

According to the features of the infrared image obtained by an imaging anti-ship missile, a method which combines sea-sky-line and coastline detection with target extraction is proposed for the extraction of ship targets. In this method, the digital image processing techniques such as Sobel edge detection and local Hough transform are used to detect the sea-sky-line and coastline in an image. Then, the entropy method introduced by Kapur is used to implement projection transform on the binary image so as to extract the ship target. The experimental result shows that the method can determine the position of a ship target in the sea-sky-line and coastline region and has a higher extraction accuracy.