Automatic Coil Delivery for Uncoiling Machine Based on Image Processing
Gang Zheng, Zewen Wang, Haowen Li, Siyu Liu · 2020
In order to automate the coil delivery for the cold rolling process and reduce manual operation error, this paper presents a method to locate and track the steel coil using monocular camera and image processing, and the core circle of coil is used as the tracking target. This proposed method firstly determines the initial target location information by RHT, then configure the qualification conditions of detection for the next time step based on the target parameters from the last time step. Canny edge detection is used to detect initial contour candidates, limited domain method, contour separation method, quadrant method are adopted to separate contour candidates and filter interference, then fit the target, and track the target by dynamic window method. Fitting the theoretical motion trajectory of the target by process a large number of video data, accurately mark the falling point and alignment point of the coil delivery process. It has been verified in the industrial field that this method has high accuracy, good tracking stability and real-time performance in the case of sufficient light. Compare with the traditional key-points detection methods based on sensor array, this system requires only a monocular camera, and realize the real-time location tracking, with higher accuracy and controllability, convenient for later maintenance.