Aircraft tire text recognition system based on machine vision

In this paper, we present a machine vision-based system for the recognition of aircraft tire text. Our approach employs both 2D and 3D cameras to capture detailed images of aircraft tires, thereby ensuring comprehensive data acquisition. By leveraging the capabilities of the PaddlePaddle deep learning framework, we utilize the YOLOv3 model to accurately identify the locations of characters on tire surfaces. Following this, we implement the SVTR model for precise character recognition. The integration of these technologies culminates in an automated character detection and recognition system that facilitates efficient and reliable reading of tire text information, which is critical for tire manufacturing industry.