一种改进的Surendra运动目标伪装效果检测方法

现有的伪装检测主要是针对静态目标，并不能客观反映机动装备在作战行动中伪装的有效性。综合运动目标检测技术，结合伪装学原理，以军用卡车模型作为研究对象，基于Surendra背景更新建模算法并结合Lab三通道，提出一种机动装备迷彩伪装效果检测方法。根据所建立的实验系统，计算序列图中军用卡车伪装前后的形状特征参数集合之间的欧氏距离，得到实验序列图像的欧氏距离平均值为189.45，说明实验系统中伪装前后运动目标的形状特征参数之间极不相似，即经过伪装后的目标更具有一定的动态变形伪装特性。借助性能评价模块得出伪装前后目标图像相关系数范围为0.003 5~0.127 5，证明了该检测方法的科学性和可靠性；根据检测运动目标的算法原理，可知伪装后的运动目标图像与其中心背景之间更具有一定的融合性。

A Surendra-based Improved Detection Method of Moving Target Camouflage Effect

The existing camouflage detection is mainly for static targets, which can not objectively reflect the effectiveness of camouflage in combat operations of mobile equipment. The moving object detection technology and the principle of camouflage are synthetical used for target detection, and the military truck model is used as the research object. Based on Surendra background update model and Lab three-channel, a new method for detecting the camouflage effect of moving objects is proposed. According to the established experimental system, the Euclidean distance between the shape feature parameters of the military truck before and after camouflage is calculated, and the average Euclidean distance of the image sequences is 189.45. It showes that the shape feature parameters of the moving objects before and after camouflage are very different. The target after camouflage has a certain characteristics of dynamic deformation camouflage. At the same time, with the aid of the performance evaluation module, the correlation coefficient range of target image before and after camouflage is 0.003 5~0.127 5. The scientific and reliability of the detection method is proved. According to the principle of moving target detection algorithm, it can be known that the images of moving target after camouflage has a certain fusion with the images of central background.