Detection and Recognition of Vehicle Components Using Image Processing and Deep Learning Techniques
Mahasak Ketcham, Thittaporn Ganokratanaa, Patiyuth Pramkeaw, Phumraphi Boonthai, Narumol Chumuang
Abstract
This research study presents an algorithmic approach for the detection and identification of key components on license plates, including numerical digits, province codes, vehicle colors, and vehicle makes and types. The methodology integrates image processing techniques and deep learning methodologies to achieve comprehensive image analysis. The algorithm incorporates image preprocessing techniques to optimize the input images, followed by a deep neural network trained on a large annotated dataset. The algorithm utilizes image segmentation and character recognition techniques to detect and classify numerical digits on license plates. Similarly, it employs segmentation and classification processes for identifying province codes, vehicle colors, makes, and types. Extensive training ensures the algorithm's accuracy and robustness under varying conditions. The proposed algorithm demonstrates the potential for accurate license plate component detection and identification, with applications in law enforcement and traffic monitoring systems.