Design of Industrial Robot Sorting System Based on A * Search Algorithm

Abstract

Robotic sorting system based on Machine Vision can quickly place the workpiece to the specified location. Gaussian transform, contour detection, binarization processing, image processing and other polygonal fitting methods  were used to identify the shape of the workpiece; Histogram technology was used to build a small text signature database method for image character recognition;  A * search algorithm was used for path planning. Experimental results show that using A * search algorithm to determine the grab order of geometry, and the optimality pruning and binary state compression method to optimize the algorithm, it is faster than the local search method and the best first  search method by 1 to 2 seconds, thus reducing the computation time.