Ship target detection for complicated inner harbor SAR images

Ship detection of harbors is significant to defeat enemy´s firepower before forming. Realization of ship detection in SAR harbor images accurately and efficiently is a hot topic nowadays. The complexity of the inner harbor and the uncertainty of target types make it difficult to achieve effective ship target detection of harbor SAR images and the detection procedure is complicated. The paper analyzes the procedure of ship detection of the inner harbor SAR images, utilizes the gradient based level sets evolution method to extract the coastline, then detects ship target in the ROI inner harbor. The experiment certifies an effective result of ship target detection.