SURF and image processing techniques applied to an autonomous overhead crane

This work presents the use of an autonomous overhead crane that detects a moving object. The crane matches the velocity of the object while its grabbers extend to reach its location. The position and velocity of the object are detected and tracked once the object is in range of the crane. This is achieved using image processing and image enhancement techniques. The SURF (speeded-up robust features) algorithm is used to extract the object's features and then detect its location. The centroid of the object and its location are calculated continuously to obtain the targeted position and velocity. A digital PID (proportional integral derivative) controller is used to control the crane's three DC (direct current) motors in order to acquire the target with a desired performance, such as a fast response and less than 2% overshoot. The proposed mechanism reduces the processing time of an industrial application which increases the productivity rate. The mechanism was built for experimentation and the algorithm and the controller were experimentally verified and validated.