| Summary: | 碩士 === 國立高雄應用科技大學 === 電子工程系 === 97 === License plate recognition system is widely used in the present time, but most of the previous researches emphasized on static system. In this thesis, a license plate location and recognition method is proposed to shoot the motion-based vehicle by the video camera set up in KUAS’s entrances and exits. As a result of the video camera was set up in outdoor environment, therefore rainy environment will impact on the system. Hence this thesis will propose a license plate recognition method for moving vehicles applied in both environments.
The principle of license plate recognition is utilizing gradient analysis and the license plate location method of this thesis proposed to find out the precise position of the license plate. Then, the located license plate must pass through boundary re-adjusting、character segmentation and character recognition with template matching to finish the license plate recognition system.
In general environment, the result shows the accuracy of license plate location is 91.51%, and the accuracy of license plate recognition is 84.54%. In rainy environment, the accuracy of license plate location is 90.59%, the accuracy of license plate recognition is 81.81%. The processing time of this system needs only 40~60mS, and conforms to the requirement of real-time system.
|
|---|
