基于背景粗糙度估计的红外目标检测算法

在分析海天背景下目标红外图像特征的基础上,提出了一种基于背景粗糙度估计的红外目标自适应检测算法.该算法利用LOG算子进行目标基本轮廓检测,将目标的中心点作为区域生长的种子点,通过背景粗糙度估计自适应确定LOG算子参量,完成目标图像分割.与传统检测方法比较,采用该处理方法进行小目标检测及识别过程的预处理,可以有效地减少运算量,提高检测速度,抑制对不必要的种子点进行区域生长,提高了目标的检测概率.

A Fast Background Roughness-based Infrared Navel Vessel Target Detection Algorithm

Based on the characteristic analysis of infrared naval vessel target under the sea-sky background, a new type of adaptive infrared small target detection method is developed,which uses the LOG operator to detect the outline of the target and then decide its center point as the seed point in the region growing procedure. In the proposed target image segmentation approach, the parameters of the LOG operator are adaptively determined using the roughness estimation of the background. Compared with the traditional detection method,this procedure is the effective method for small target detection and recognition,and it can reduce the computation requirement ,improve the detection speed, restrain the unnecessary seed points in region growing process, and increase the possibility of target detection.