机载跟踪器的逃逸目标图像快速识别模型仿真  

Fast Identification of Dynamic Target Image Tracking Model of Airborne Simulation

作　　者：卫朝霞[1] 

机构地区：[1]电子科技大学成都学院,四川成都611731

出　　处：《计算机仿真》2015年第5期75-78,共4页Computer Simulation

摘　　要：机载跟踪器进行逃逸目标识别跟踪的过程中,目标往往是高速移动的,为了逃逸跟踪,运动的速度和方向具有较大的突变性,与跟踪器之间会形成较大的跟踪距离差。传统的视觉方法为了弥补这种距离差,通过减少采集图像的样本,以提高跟踪的准确性,但是一旦样本数量过少,将会导致识别跟踪的准确性降低。提出基于高斯分布的动态逃逸目标识别跟踪方法。对跟踪目标图像进行归一化处理,获取与数据库中的样本图像具有同样的尺寸和分辨率的图像。计算动态目标图像的形心,根据高斯分布原理建立动态目标的识别跟踪模型,并对模型中的相关参数进行及时更新,保证了动态目标识别跟踪的及时性和准确性。实验结果表明,利用改进算法能够有效提高动态目标识别跟踪的准确性,缩短了识别跟踪的时间。In the process of onboard tracker identifying and tracking the escape target track, targets often move fast. In order to track escaping targets, the speed and direction of its movement have great mutability, and it will form a larger tracking range difference between the trackers. An identification tracking method for dynamic escape target based on Gaussian distribution was proposed in the paper. Tracking target image was normalized to obtain the image which has the same size and resolution with sample image in database. The centroid of dynamic target image was cal- culated, and according to the principle of Gaussian distribution, the identification tracking model of dynamic target was established, and the related parameters in the model was timely updated, to ensure the timeliness and accuracy of the dynamic target identification and tracking. The experimental results show that using the improved algorithm can effectively improve the accuracy of dynamic target tracking, and reduce the tracking time.

关 键 词：机载跟踪器 动态目标 识别跟踪 高斯分布 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]