传统的车辆视频检测算法都是基于带有复杂背景的面阵CCD图像,不利于对象分割和特征提取。线阵CCD图像的背景比较单一,且帧率远高于面阵CCD,可以实现高精度的车辆检测。提出基于小波变换的线阵CCD图像的车辆检测算法。试验结果证明该算法能有效降低光照和汽车阴影等因素造成的误检率。 Traditional vehicle video detection algorithms are all based on an area array CCD image with complex background, which is not good for object segmentation and feature extraction. Linear CCD image background is relatively simple, and the frame rate is much higher than the area array CCD, you can achieve high-precision vehicle testing. A vehicle detection algorithm based on wavelet transform is proposed. Experimental results show that the algorithm can effectively reduce the false detection rate caused by factors such as light and shadow.