基于5×5模板的红外小目标图像检测算法

摘要

In view of the characteristic of small targets in the sky background, an algorithm used ior detection of small targets in the infrared image is presented. In the target of more than 9 points, the traditional image process- ing based on the 3 × 3 template does not reach the expected effect. Therefore, this paper uses the middle and top- hat filter to process the infrared image based on the 5 ×5 and 4 ×4 template. The design adopts the FPGA. Experi- ments on real infrared images in the sky background show that the algorithms can detect the small target in the infra- red image effectively.%针对天空背景下红外图像弱小目标的特点，提出了一种用于红外小目标检测的算法，在9个点及以上的目标中，传统图像预处理采用3×3模板达不到预期的效果，因此文中采用5×5，4×4的模板对红外图像进行中值滤波．灰度形态学等预处理，采用大规模可编程门阵列来实现，通过实测天空背景下的红外图像实验，表明了该算法能有效地检测出红外图像中的弱小目标。