Fast license plate location and recognition using wavelet transform in android

In license plate recognition system (LPRS), there have several parts which are the key steps of the LPRS as license plate location (LPL), character segmentation (CS), and character recognition (CR). In this paper, we develop a complete LPRS in mobile smart device. The reasons are that have advantages which are easy carrying, powerful camera existence, and extensive application in potential. First at all, we define a processing range which is region of interest (ROI). In the part of LPL, we use wavelet transform to detect the horizontal axis in ROI based on texture feature of license plate (LP) and block scanning based on aspect ratio of LP to locate the LP location. In the part of CS, we mark the character edges based on color feature of LP to segment the characters. In the part of CR, we normalize the characters at first, then, compare them with character samples in system database. We use voting result to recognize. As result in the experiments, the LPRS which is building on mobile smart device have superior recognition results and fast processing performance.