Research On Lao License Plate Recognition Based On Mask RCNN

License plate recognition technology is a comprehensive application of related technologies such as computer vision,image processing and deep learning.Both traditional license plate recognition technology and deep-learning-based license plate recognition technology have been widely used in China.However,the Laos license plate recognition technology is still in the preliminary exploration stage.This thesis studies the Laos license plate recognition scheme based on Mask RCNN,and verifies through experiments that the scheme can complete the operation of Laos license plate recognition.This thesis investigates and compares the traditional license plate recognition technology and the license plate recognition technology based on deep learning.License plate recognition technology is generally divided into five steps: image input,image preprocessing,license plate location,character segmentation and character recognition.This article introduces the image processing technology involved in license plate recognition,deep learning related theories and the structure and function of common convolutional neural networks.This thesis collected Lao license plate pictures under different scenes,multiple angles and different lighting conditions.After a series of preprocessing operations on the pictures,the first Lao license plate data set was produced,and then sorted out.Based on the Mask RCNN network model,the license plate location training process,the letter recognition training process,and the Laos license plate province recognition training process were completed,achieving a relatively ideal recognition effect.This article saves the trained model and verifies the good effect of the training result.According to the technical scheme,this thesis completed the production of the Laos license plate recognition data set,and verified the license plate recognition effect,achieving a 99% correct rate.This fills the gaps in Laos license plate recognition technology,and provides theoretical and experimental basis for the application of Laos license plate recognition technology in actual scenarios.