Research Of License Plate Recognition System That Applied To The Intelligent Parking Lot

With the number of vehicles increased in urban road, parking is an urgent problem to be solved. Especially in shopping malls, hospital, office building, and other high population concentration place, parking management have heavy influence on people’s travel and work efficiency. IC card parking management system is commonly used at present. By using IC card, the driver must stop to take the card when entering the parking lot, and stop to give it back and pay the parking fee when going out. This would easily cause traffic jam. And if the driver carelessly lost the IC card, it would take time and energy to find out the car’s parking information. Bring license plate recognition system in parking management would save the time of take and give back the IC card, and can reduce the staff to charge the parking fee, remove some facilities. License plate recognition enables the parking system to recognize the license plate as the car’s ID, and with the cooperating of mobile internet, drivers can pay parking fee through internet automatically. Given the economic and practical condition, and based on the investigation of the parking management system in a big office building in Hangchow, this article designed a license plate recognition system used especially in parking lot, counting on this to improve the management of our parking system and save time of driver.The license plate recognition system separate recognition process into three steps: license plate location, license plate correction and character segmentation, character recognition. The research on each step as follows:1. License plate location. Start with the size and feature of license plate in China, I analysis the real-time performance and application character of current license plate location technical. In order to satisfy the precise and real-time recognition requirement of intelligence parking lot, The paper propose a quick license plate location method based on Grayscale image vertical projection. Compared with other license location method, this method combines the edge feature with the projection feature in Grayscale image of license plate more effectively, and guarantee a high accuracy rate of 98.4 in the premise of less calculated quantity.2. License plate correction and character segmentation. After the binarization of the license plate image, depend on the centroid of character connected domain and rotating projection algorithm, the inclined license plate image can be adjusted. On the basis of character connected component analysis, propose a self-adaption binaryzation license plate character segmentation method. Different from other segmentation method, this method estimate the rationality of characters according to the valid number of character connected domain after binarization. In some less calculated quantity binarization image, the system can segment valid character connected domain quickly, while achieve a high accuracy rate of 97.1%.3. Character recognition. Based on the technology and principle of character recognition, this article analyses the particularity and difficulty of license plate character recognition, and propose the method of English and digital character recognition based on contour feature. By using innovative idea of mark concave-convex profile of a character, the recognition can be much quick and precise. Now it only needs 20 ms to recognize a character, and the recognition rate is 98.1%. Specific to low-resolution Chinese character Grayscale image, this article raise that the application of probability fuzzy template in license plate Chinese character recognition can solve the problem of low quality of Chinese character image, and have a better recognition performance.The license plate recognition system presented in this article can hoist recognizing speed by a large margin, is particularly suitable for spot license plate recognition facilities installed embedded system. It has certain economic value and practical significance in improving parking management, establishing a quick, effective, online monitoring system in intelligence parking lot.