基于贝叶斯估计的环境车辆感知

Environmental Vehicle Perception Based on Bayesian Estimation

为了改善传统背景差分法受场景变化影响较大的缺点,文章提出了一种使用贝叶斯估计对差分后的图像像素点进行是否为车辆的概率估计方法的实验。实验结果表示,抑制场景无用的像素点变化达到12.5%。基于贝叶斯估计的环境车辆感知能够明显提高车辆的检测准确度,降低误检率。

In order to improve the traditional background subtraction method which is greatly affected by scene changes,this paper proposes a Bayesian estimation method to estimate whether the pixels in the difference image are vehicles.The experimental results show that the useless pixels in the scene can be suppressed by 12.5%from the distribution histogram.Environmental vehicle perception based on Bayesian estimation can significantly improve the accuracy of vehicle detection and reduce the false detection rate.