Acceso usuarios
Chen Geng, Jeffrey S. Sarmiento, Anton Louise De Ocampo, Rowell Hernandez
Due to the complex image background, diverse changes in license plate size and position, license plate localization becomes difficult. To optimize the detection effect of electric vehicle charging station occupying vehicles, this study introduces deep learning theory and conducts research on the detection method of electric vehicle charging station occupying vehicles. Firstly, perform grayscale, binarization, coarse localization, and tilt correction on the initial collected images; Then, the connected domain method is used to segment the vehicle characters occupying the space; Finally, the Faster R-CNN neural network was applied, combined with the advantages of Regional Proposal Network (RPN) and Convolutional Neural Network (CNN), to complete the detection of electric vehicle charging station occupancy vehicles. The experimental results show that the application of the proposed method significantly improves the accuracy of license plate localization, with detection accuracy higher than 95.7% and detection time lower than 13.2ms, which is better than the comparison method and has better application effects.
