采用图像特征的激光干扰跟踪效果评估  

Assessment of laser-dazzling effects on tracking through image feature measurements

作　　者：钱方[1,2] 孙涛[1] 郭劲[1] 王挺峰[1] 王岩[1,2] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所激光与物质相互作用国家重点实验室,吉林长春130033 [2]中国科学院大学,北京100049

出　　处：《液晶与显示》2014年第5期805-811,共7页Chinese Journal of Liquid Crystals and Displays

基　　金：国家重点实验室自主基础研究(SKLLIM1203-01)

摘　　要：对于广泛应用于图像识别和跟踪的激光主动成像系统,激光干扰是影响该类系统性能的一个重要因素。激光干扰使系统中的光电成像元件达到饱和或损伤,在探测视场中出现光斑,从而使跟踪算法无法准确分辨目标和光斑。因此,分析并评估激光对主动成像系统的干扰效果是避免干扰的前提,同样也是重要的环节。提出一种基于图像特征的边缘相似度(ESIM)评价算法,该算法通过模板匹配确定图像的目标区域,并比较原始图像和干扰图像目标区域的亮度、对比度和边缘清晰度差异,得到归一化的评价指标,作为对激光干扰效果的评估标准。实验中分别采集了不同干扰功率和不同光斑位置的激光干扰图像,并利用相关跟踪和形心跟踪方法对目标进行跟踪,运用本文提出的ESIM算法从图像特征角度对干扰跟踪效果进行评估,结果证明ESIM能够合理评估干扰图像的失真程度,并可以较好地反映激光对主动成像系统跟踪性能的影响。Laser active imaging systems are widespread observation tools used to fulfill recognition and video-tracking. Laser-dazzling on imaging systems aims to saturate or damage the full camera image. If there is a laser spot in the field of view, the tracking system cannot distinguish between the target and the laser spot. In order to evaluate the dazzling effects of the laser and avoid such a disturbance, the assessment of laser-dazzling effects must be better understood. This paper proposes a new algo- rithm composed of edge similarity （ESIM）. Luminance comparison, contrast comparison and edge definition comparison are computed for the original and disturbed images. Based on the results of the above, we could construct a modified assessment. In terms of experiment results,ESIM is efficacious to reflect the quality of tracking image and it can evaluate the laser dazzling effect on tracking successfully.

关 键 词：激光干扰 对比度跟踪 相关跟踪 边缘相似度 

分 类 号：TN977[电子电信—信号与信息处理]