基于PDE的前视红外成像舰船目标检测

Ship-Target Detection from Forward-Looking Infrared Image Based on the PDE

针对复杂环境下的海面目标提取问题,提出一种基于偏微分方程(PDE)的红外舰船检测算法。首先采用基于PDE理论的滤波模型对初始图像滤噪并进行背景估计,然后结合邻域差分进行背景抑制后分割提取目标。实验表明,该算法能够有效检测强杂波背景中的目标,方法适应性强。

Due to the sea light,atmospheric transmission and detector noise,the ship images taken by forward-looking infrared sensors are low contrast between target and image background.Furthermore,islands,sea clutter,clouds and other complex background in the field of view may interfere with the detection.Without background suppression,image segmentation result can be less of detail information,and the structural information of target is difficult to extract.All of these factors reduce the system performance.In the paper,an infrared ship target extraction algorithm based on PDE is proposed.It mainly includes a background suppression algorithm based on PDE and a neighborhood differencial method between original image and background image.Experimental results demonstrate that the algorithm effectively inhibits the background clutter in infrared images,and lower the difficulties for the target segmentation.