A Novel Coastline Detection Algorithm Based on Modified Iterative Selection Method
Misjudgment often occurred in low contrast remote sensing images, because most widely used image segmentation algorithms often have a larger threshold. To overcome this problem, a novel coastline detection algorithm is proposed. A restriction function is involved into conventional iterative selection process. According to langrage multiplier, a modified iterative selection model is formulated. This modified method utilizes the gradient of images to obtain an optimal threshold. A region grouping rule is proposed to distinguish land and sea. Experimental results show superior performance of proposed method in terms of accuracy. As an application, our method has been applied to extract the coastline of the remote sensing image with promising results.