An Efficient Industrial System for Vehicle Tyre (Tire) Detection and Text Recognition Using Deep Learning

This paper addresses the challenge of reading low contrast text on tyre sidewall images vehicles in motion. It presents first its kind, a full scale industrial system which can read codes when installed along driveways such as at gas stations or parking lots with driving under 10 mph. Tyre circularity is detected using circular Hough transform dynamic radius detection. The arches are then unwarped into rectangular patches. A cascade convolutional neural network (CNN) classifiers applied for recognition. Firstly, novel proposal generator code localization introduced by integrating layers producing HOG-like (Histogram Oriented Gradients) features CNN. proposals filtered deep network. After localized, character detection and recognition carried out two separate CNNs. results (accuracy, repeatability efficiency) impressive show promise intended application.