License plate detection based on 2-level 2D Haar wavelet transform and edge density verification

Against vertical Sobel operator that most license plate detection (LPD) algorithms adopt, this paper presents a robust and real-time preprocessing method to enhance both edge density and intensity of license plates under various outdoor and indoor environments. The proposed method applies HL subband feature of 2D Discrete Wavelet Transform (DWT) twice to significantly highlight the vertical edges of license plates and suppress the surrounding background noise. Then, several promising candidates of license plates can easily be extracted by first-order local recursive Otsu segmentation and orthogonal projection histogram analysis. Finally, the most probable candidate is selected out by edge density verification and aspect ratio constraint. Experimental results show LPD system based on the proposed 2-level 2D Haar DWT is better than that based on conventional vertical Sobel operator or 1-level 2D Haar DWT in terms of the detection rate.