基于海岸线数据库的遥感图像海陆分割

Sea-land segmentation for remote sensing imagery based on coastline database

为了提高海面上舰船目标的检测效率,使用基于海岸线数据库和扫描线多边形填充的海陆分割算法对全色遥感图像的海洋区域和陆地区域进行分割。算法分为三步展开,使用海岸线数据库在图像上标出粗略的海岸线;使用边缘检测算法sobel算子进一步精细海岸线边缘;使用扫描线多边形填充算法将封闭陆地填充。实验结果表明,与单纯的图像处理算法相比,该算法的分割结果精度更高,弥补了传统图像处理算法中分割结果易受图像噪声影响的缺陷,算法更加简单高效。

To improve the detection efficiency of ship targets on the sea, propose an algorithm of sea-land segmentation based on coastline database and scan-line polygon-filling to segment the sea area and land area of panchromatic remote sensing images. The algorithm is divided into three steps, using coastline database to mark coastline on the image, using sobel operator of edge detection algorithm to further accurate coastline edge, using scan line polygon filling algorithm to fill closed land.The experimental results show that, compared with the simple image processing algorithm, the segmentation result of this algorithm has higher accuracy, which makes up for the defect that the segmentation result of traditional image processing algorithm is easily affected by image noise, and the algorithm is more simple and efficient.