改进的高斯背景建模在车辆检测中的应用

Improved Gaussian background modeling in the vehicle detection apply

研究实时环境下运动车辆的检测算法,针对存在渐变及重复性运动的车辆运动背景采用改进的高斯背景建模法,对背景进行更新,将所得的当前背景与前一帧视频图像进行相减,得到当前时刻车辆的运动图像。将所得图像进行数学形态学去噪处理得到较为理想的车辆检测效果图,并根据处理后的图像判断车辆是否运动以及其运动轨迹。实验证明该方法能去除噪声对图像产生的影响,对判断某一时刻车辆是否运动行之有效。

Research real-time environment movement vehicle detection algorithm, Based on the existing gradient and repetitive motion of the vehicle movement background of the improved gaussian background modeling method, To update the background,the income of the current background and the previous frame video image subtraction, get the current moment of the vehicle motion image . Will the mathematical morphology image denoising processing get more ideal of vehicle detection rendering, And according to the processed images judgment whether the vehicle motion and its trajectory and the processed image according to judge whether the vehicle motion and its trajectory. Experiments show this method can remove noise on the effects of image,to judge whether a moment vehicle movement effective.