基于MRF的自适应正则化红外背景杂波抑制算法

针对复杂背景下红外弱小目标检测难题,将背景杂波抑制归结为从原始红外弱小目标图像中重建目标数据的过程,据此提出了一种基于马尔可夫随机场模型(MRF)的自适应正则化滤波算法.该算法采用MRF,建立了红外弱小目标图像的先验概率模型,并根据图像的粗糙度设计了新的势函数.在此基础上,采用MRF对背景杂波抑制过程进行正则化处理,从而实现了对红外背景杂波的自适应各向异性抑制.理论分析与实验结果表明,该算法能够随图像局部纹理特征的变化自适应地调整滤波算子结构,从而可在复杂背景下自适应地抑制杂波、增强信号,有效地提高了图像的信噪比,且该算法结构简单,更易于硬件实时实现.

INFRARED BACKGROUND CLUTTER SUPPRESSION ALGORITHM OF ADAPTIVE REGULARIZATIONBASED ON MRF

Aiming at the difficulty in detecting infrared(IR) dim small target under strong background clutter, the process of background suppression was attributed to the reconstruction of the target signal from the original IR dim small target image. Thus, a novel adaptive regularization filtering algorithm based on Markov random field(MRF) model was proposed. In our algorithm, the prior probability model of the IR dim small image was established by MRF, and a new potential function was introduced according to the roughness of the IR image. On this basis, the adaptive anisotropic filtering effect for background clutter suppression was realized by regularizing the process of background clutter suppression with MRF. Theoretical analysis and experimental results show that this algorithm can adjust the operator adaptively according to the local texture distribution character of the image. Thus, the target was enhanced and strong background clutter was eliminated. The proposed algorithm can improve the signal-to-noise ratio (SNR) of the image obviously with the advantage of its logical structure simple to be implemented in real-time system.