基于背景学习与方向预测的车辆图像跟踪

针对移动目标检测,提出了基于变化因子参考背景学习与方向预测算子的图像跟踪方法.以高速行驶车辆为目标建立变化因子,在此基础上进行视频图像参考背景的学习,并通过互相关匹配与坐标变换实现运动目标的定位及速度确定,结合方向预测算子对车辆行驶矢量进行预测,进而实现目标车辆的图像跟踪.跟踪实验及性能比较证明本方法可获得准确稳定的运动车辆跟踪结果,为运动目标的图像监控研究提供了新的思路.

Vehicle's image tracking based on background learning and direction predicting

For the purpose of moving target's detecting, an image tracking method based on the changing factor of reference background learning and the direction predicting operator is presented. The high-speed moving vehicle is used as the detecting target for establishing a changing factor, on this basis the image sequence's reference background learning is conducted, then by cross-correlation matching and coordinate transforming the moving target-vehicle is coordinate-positioned and speed-determined, combining with the direction predicting operator the moving direction of target vehicle is predicted, thus vehicle, s image-tracking can be realized. Through tracking experiment and performance comparison, it is proved that by this new method an accurate and stabletracking result of moving vehicle can be gotten;a new idea can also be provided for the research of moving target's image-monitoring.