Object detection methods for Image-based Visual Servoing of 6-DOF Industrial robot

The employment of robots in manufacturing industries is on the rise due to the need for automated industrial operations. Visual servoing is a well-advanced method of controlling the robot movement based on visual sensor feedback. In image-based visual servoing (IBVS), the image features at the final pose implicitly define the desired camera pose relative to the target. The major issue associated with image-based visual servoing of industrial manipulators is the identification and tracking of object to be manipulated from the image sequence. In this paper, two methods are employed to detect the object for providing image feature vector in the visual servoing feedback loop. One method employs a robust and fast algorithm called SpeededUp Robust Features (SURF) for image comparison. Another method utilizes the method of comparing measuring the properties of the image region and comparing with the given image to detect the object. These algorithms are applied in the image-based visual servoing system for obtaining the image feature error to drive the system. Real-time experiments are conducted in IBVS system where IRB 1200 industrial robot manipulator is active with eye-in-hand camera. The results illustrate the improved performance of the object identification algorithms.