Object Detection and Text Recognition for Conveyor Belt Systems with Robotic Operations

Computer Vision has become an indispensable tool in modern industrial automation, offering significant advances in efficiency, accuracy, and real-time decision making. This thesis investigates how robotic systems and computer vision can be combined to enable dynamic object manipulation and detection in industrial settings. The project focuses on a conveyor belt system that uses a deep learning model that has already been trained to detect moving objects. In the detection process, objects are first classified and then localized, and then their actual locations are calculated using depth information. Following object localization and identification, a robotic arm uses the processed data to carry out accurate pick and place tasks. The system is designed to address scenarios involving moving objects and demonstrates the potential of computer vision and robotics to increase productivity and automation in manufacturing processes. The project demonstrates how using advanced computer vision techniques for real-time object detection and text recognition with robotic manipulations can increase the efficiency and precision of industrial operations.