一种降低红外小目标图像中云干扰的算法

摘要

To improve the detection probability, a new algorithm based on shape characters of target for IRSS was raised. Firstly, the algorithm removed background noise to enhance the SNR of infrared image by mathematical morphology filtering. And then, based on the characters of targets' head and tail, constant false alarm ratio (CFAR) detection was used to separate potential targets from clouds interference. Compared with the result of traditional algorithms, the new algorithm decrease clouds interference can more quickly, and is more helpful to increase the detection probability of infrared small targets.%为有效提高目标的检测概率,提出了一种基于目标形态特征降低云层干扰的算法.该算法首先采用形态学滤波法进行背景抑制以提高图像的信噪比,然后根据目标头尾两侧水平方向存在阶跃信号的特征,利用恒虚警技术,将潜在目标从干扰中分离出来,待后续处理.实验结果表明,与传统的目标检测算法相比,该算法能够快速地消除由云层所引起的干扰,且便于硬件实现,对提高红外小目标的检测概率具有较好的辅助作用.