Review for the current performances for Machine Visions in Industrial Robotic welding and drilling tasks

Industrial robots have been more and more involved in the automation industry due to their capability to perform precise tasks, with an accuracy in sub-millimeters, tasks such as welding and drilling where a successful cooperation between the robots and the machine vision is necessary to end tasks within a demanded accuracy and in less execution time. The feedback from the machine visions is used for enhancing the efficiency of detection, tracking and control of the robot motion by utilizing their visual information. The feedback, therefore, improves the safety of the system by preventing the robots from being damaged and operators from being injured which, in turn, saves the production time. Robotic welding tasks is one of important applications for industrial robots where high temperatures can limit the ability of welding workers to monitor and control the process around the close proximity of the welding area. Here the use of vision sensors could significantly aid in terms of protecting the welded area and provide safety to the operators. Similarly for the drilling tasks where robots are widely used, visions systems has significantly improved their performance to achieve desired accuracy. This paper targets these application areas and presents a review of the state-of-the-art equipment, methodologies and practices used within the associated research areas of robotic systems in the context of vision systems. It also examines the recent contributions of the vision systems in robotic tasks and highlights on their performance, the use of algorithms for image processing and calibration procedures adopted, and their contribution towards the effectiveness of robotic positioning resolution and accuracy.