Multi-target car license plate detection from complex environment

Summary form only given. In recent years, there has been much research on license plate recognition (LPR), and LPR systems are used frequently. The three major parts in LPR are plate detection, character segmentation and character recognition. A smart and simple algorithm for LPR is presented. This algorithm can be applied for multiple license plate detection. There may be many cars in an image with a complex background. We study how to extract license plates effectively. We use edge detection, image dilation, block filtration and license proportion checking to extract plates accurately. Experimental results show that the algorithm is robust in extracting multiple plates in an image with a complex background. Results yield more than 86.67% correct location of license plates.