为解决目前方法在动态复杂环境下不能有效区分车灯和瞄准镜及“猫眼”效应目标识别率低的难题, 提出了一种基于纹理特征的“猫眼”效应目标识别方法.首先采用列均值相减、指数高通滤波和中值滤波对差分图像进行滤波去除车灯；然后依据“猫眼”效应目标的形状特性进行判别；最后根据“猫眼”效应目标的纹理特征, 构造隶属度向量, 按照最大隶属度原则, 采用模糊综合评判对“猫眼”效应目标进行识别.试验结果表明, 与基于形状和频率双重判据的“猫眼”效应目标识别方法相比, 该方法有更高的识别率和较低的虚警率.

In order to overcome the low recognition rate of “cat-eye” effect target and ineffective distinction of car lights and telescopic sights, a recognition method of “cat-eye” effect target based on texture character was proposed. Firstly, car lights were removed by filtering the difference images with the column-mean -subtraction, EHPF and median filter. Secondly, the discrimination was carried out in terms of the features of the shape of the “cat-eye” effect target. Finally, according to the texture features of the “cat-eye” effect target, membership vectors were constructed, and the fuzzy comprehensive evaluation was used to recognize the “cat-eye” effect target under the maximum membership degree principle. Experimental results indicate that the proposed method reduces false alarm rate and improves recognition rate compared with the shape-frequency dual criterions (SFDC) method.