A New Method for License Plate Location Based on Top-Hat Transform and Wavelet Transform

In this paper, a novel method of the number license plate location based on top-hat transform and wavelet transform is proposed. First, top-hat transform is used to suppress the background of image and remove the nonuniform illumination. Second, vertical edge detection and morphology filter is applied to the image, coarsely location of the license plate is completed by the structural characteristics. And then the candidate regions of the license plate are binarized and their VPMs (vertical projection map) are worked out, and one-dimensional discrete wavelet transform is applied to the VPMs, and the wavelet coefficients are set as zero and then the VPMs are reconstructed. Then, the statistics of the reconstructed VPMs and scale coefficients are combined to create the texture feature vectors. At last, BP neural network is used to recognize the real vehicle license plate. Experimental results showed the great robustness and efficiency of the proposed method.