A deep learning model for container code detection and recognition applied to smart port operations

Abstract

Computer vision, a key area within artificial intelligence, has been rapidly advancing and is increasingly applied across various industrial domains. Based on the architecture of Convolutional Neural Networks (CNNs), numerous state-of-the-art models have been developed to address a range of tasks, including object detection, image segmentation, and optical character recognition (OCR), etc. Among these, YOLO (You Only Look Once) stands out for its high-speed and accurate object detection capabilities, while EasyOCR has proven to be an effective tool, offering high character recognition accuracy. The present study focuses on the detection and recognition of container codes by integrating the YOLOv11 model with EasyOCR. The research encompasses the construction of a training dataset, model training, and model performance evaluation. Output results indicate that the proposed model achieves an accuracy of over 90%, demonstrating its feasibility and strong potential for real-world applications in the smart ports.