Abstract: Gesture interaction is the important research area in human-computer interaction. Vehicle gesture recognition system can reduce the distraction caused by operating instrument and improve the safety of driving. Influenced by illumination changes, the internal environment of the car, camera imaging quality and other factors, large amount of complex noise exists in the vehicle gesture images, which seriously affect the accuracy of gesture segmentation, feature extraction and gesture recognition. In this paper, an image processing method suitable for vehicle gesture images was proposed to solve this problem. In this method, one-dimensional nonlinear diffusion filtering was used to remove the noise in the high frequency sub band after wavelet decomposition and get the preliminary denoising image. Then, the preliminary denoising image was further denoised by multi-scale bilateral filtering. Experiment results show that the proposed method can better remove the noise and prevent the blurring of details of the vehicle gesture image than other methods.