Autonomous Drive-Through System Using EasyOCR and Dobot

This research introduces an innovative approach to revolutionize drive-through operations in the retail sector through autonomous technologies, focusing on the critical role of optical character recognition (OCR) for recognizing Thai license plates. The integration of DoBot technology with OCR addresses common issues such as order accuracy, wait times, and operational inefficiencies inherent in traditional drive-through models. Our license plate detection model exhibits impressive performance, achieving an accuracy of 97.47%, precision of 98.10%, recall of 99.35%, and an F1 score of 98.73%, underscoring its ability to accurately identify license plates with minimal false positives. However, the OCR system's character recognition rate of 74.04% indicates moderate accuracy, suggesting areas for enhancement. This system enables precise vehicle identification, personalized order processing, and secure payment, streamlining the entire ordering process. By leveraging DoBot technology for autonomous order delivery, we enhance overall operational efficiency and reduce the workload and number of workers required, thereby lowering labor costs and improving productivity. Future efforts will focus on improving OCR preprocessing techniques and algorithms to better handle challenging Thai characters, further enhancing the system's reliability and effectiveness.