基于混合智能的車牌識別關(guān)鍵算法研究.pdf

1、安徽大學碩士學位論文基于混合智能的車牌識別關(guān)鍵算法研究姓名：陸玉申請學位級別：碩士專業(yè)：計算機技術(shù)指導教師：羅斌2010-05基于混合智能的車牌識別關(guān)鍵算法研究 II Abstract With the rapid development of China's economy, major cities in the traffic management level is gradually increasing, and be

2、cause of this, many of the intelligent management of traffic management measures began to be large-scale applications. In this one, motor vehicle license management as an important indicator of traffic vehicles, became t

3、he focus of management practices. Based on this, the vehicle license plate recognition system (LPR, License Plate Recognition) RD and application of intelligent transportation systems began to be affected, the most impor

4、tant factor of modernization, its also become a hot issue of the development of modern transportation. In the LRP system, the pair of algorithms are of particular importance, it contains the following parts, is related t

5、o academic studies of Zhongdian, namely: the Ding Wei Ju Liang license, vehicle license plates license Fenge and character recognition. Meanwhile, in order to enhance the license plate recognition, which requires efficie

6、nt recognition algorithm used to ensure that even in ambient lighting conditions Bad filming location and vehicle licenses and other factors inherent flaws, the still greater robustness to help the system accurately, rea

7、l-time identification requirements. In this paper, my work is focused on the image processing by the emulator. I analyze three problems, which are license locating, segmentation and character recognition. Before license

8、locating, 1. this paper first analyzes the current artificial intelligence methods which had been used, then compared with the current quantum evolutionary algorithm and particle swarm algorithm, then based on these me

9、thods, a new hybrid intelligent quantum particle swarm algorithm is built to preprocess the image; 2. the light of today's most commonly used method of positioning on the vehicle license, vehicle license and then com

10、bined with some of the inherent physical characteristics, in this paper, using methods based on mathematical morphology, adaptive function by constructing a structural element to the car candidate target area to locate t

11、he license; 3. summing up the vehicle license plate character segmentation algorithm based on continued use of domestic license plate license plate number of a priori knowledge such as the interval between the character

12、s within the license plate characters within the composition, then the objectives of the digital image in car card region vertical and horizontal projection, combined with knowledge of these bright, so can be a complete

13、license plate segmentation; 4. in character recognition, taking into account the actual license plate recognition features, calculated using the theory of intelligent design an improved BP neural network, and in accorda

14、nce with the classification of character features, respectively, with the improved BP neural network license plate the character recognition. KEY WORDS: License Plate Recognition, DSP, Quantum Evolution, Particle Swarm