基于图像的肇事逃逸车辆定位方法研究与仿真

摘要

Research accurate positioning of hit - and - run vehicle. Vehicles in image scene is complex and has many changes. This paper put forward a hit and run traffic image fast positioning method based on association mining algorithm to secondary check and filter information with including vehicle features. Then, it reduces image positioning errors and uses the correlation features for accurate positioning. The experiment results show that the improved algorithm can extract 2 - d parameters and improve the accuracy of identification.%研究肇事逃逸车辆快速准确定位问题.车辆在逃逸过程中,逃逸车辆图像所在场景变换较为快速,车辆行驶方向突变性较强,使得车辆定位特征衰退明显.传统的车辆定位算法建立的特征模型由于这种场景图像快速变换和方向突变的影响,无法准确的捕获运动车辆的关键特征,定位准确性较差.提出改进关联挖掘算法的肇事逃逸车辆图像快速定位方法,在变化场景对包含冗余车辆特征的信息进行二次校验过滤,进而减小定位图像错误分类的可能,运用关联特征准确完成定位.仿真结果证明,改进算法能够大幅提高肇事逃逸车辆图片的关键帧定位结果的准确性,为逃逸车辆准确定位提供了依据.