基于细胞响应模型的红外背景抑制

An Improved Algorithm for Background Suppression Based on Cell Response Model

为检测复杂背景红外图像中的小目标,解决红外弱小目标检测易受复杂背景干扰的问题,提出基于非下采样Contourlet变换(NSCT)和细胞响应模型的红外图像背景抑制算法。首先采用非下采样Contourlet变换对图像进行多尺度、多方向分解,得到各子带系数;然后根据目标、背景的不同特点以及子带的方向性,选取细胞响应模型的参数对中频各个子带系数进行处理;最后对处理后的子带系数进行重构得到背景抑制图像。实验结果表明:本文算法取得了更好的背景抑制效果,提高了单帧红外图像的处理能力,降低了后续检测与跟踪的难度。

An algorithm of infrared image background removing based on NSCT and cell response model was proposed to improve the performance of infrared dim target detection under complex background.This algorithm used the non-subsampled Contourlet transform（NSCT） to decompose infrared image into a series of subband coefficients corresponding to multi-scales and multi-directions of the original image.Then,the differences between target and background and the orientations of different subband are used to choose the parameters of the cell response model.Such models were further utilized to get the new intermediate frequency subband coefficients.Finally,the suppressed images were reconstructed by the new coefficients.Experiment showed that the proposed algorithm got a better result of background suppression,improved the ability to process single frame infrared image,reduced the difficulty of further detection and track.