Feature Extraction of Driver in Traffic Image Based on Wavelet Critical Threshold Denoising Method
Qian Xiao, Yaxin Hou · 2019
With the increasing number of vehicles, traffic accidents continue to occur. Accurate extraction of driver's image information has become an important means to improve traffic safety. At present, the driver's face recognition system mainly obtains the recognized image through the grid label image, but sometimes this method can not extract the driver's characteristic information accurately, which brings errors to the driver's recognition information. In this paper, the main method of driver characteristics in traffic image based on wavelet transform is presented. Based on the traditional method, the principle of wavelet threshold denoising is introduced, and the denoised traffic image is extracted by gray processing, grid marking and binarization. The simulation results show that the resulting image is more accurate than the original method, and can accurately extract the driver's characteristic information, which provides a favorable condition for investigating the driver's information.