The Optimal Path Planning Of Tripod Robot Based On Visual Processing And Ant Colony Algorithm

In industrial production,objects often need to be sorted.The action is monotonous and repetitive.Operators are easily exhausted and inefficient.However,vision robot can finish the job easily,whose accuracy and rapidity is incomparable.In this paper,a visual system was designed for Tripod Robot to make it have visual processing function.In this paper,the process of sorting operation was simplified as the precise pick and place of geometries on the Tripod platform.In order to improve the accuracy of sorting,and increase the speed of this process,visual processing and path planning were designed.Image processing module achieved image acquisition and preprocessing,location of center points and angles,recognition of shapes,color and symbols.The ant colony algorithm was used to obtain optimal path,which realized the shortest moving path and placed geometries with same color together.Image acquisition and preprocessing.A side view of platform was obtained with a camera fixed on the control cabinet.The perspective transformation was used to correct the distortion image.Gray scale was applied to remove the color information.Bilateral filtering was utilized to reduce the noise.An improved dual threshold Otsu was designed for binary processing.An modified Canny operator was adopted to extract satisfied edge contours.Location and recognition of the platform and geometries.A two-step location method was created to gain the precise positions of geometries on outer ring and inner holes.Hough transforms of angles between edge lines and coordinate axis were adopted to obtain values of rotation angles.Hough transforms of circles and lines with geometry properties were used to detect shapes.After extracting RGB values of center points on each geometry,color classification was realized in HLS color system.An algorithm of symbol recognition was created to distinguish symbols by means of judging classification and inner boundary numbers of connected domains.Optimal path module based on ant colony algorithm.The problem of optimal path planning was summarized as a traveling salesman problem with constraints.Slow speed of convergence and falling into local optimum easily were two major defects of traditional ant colony algorithm.Therefore,an self-adaptive heuristic function was designed to improve the searching speed.Andupdating mechanism of pheromone was optimized to achieve the global optimum.The optimal grasping sequence of the nine geometries was obtained eventually.TCP/IP communication module.Through establishing an TCP/IP connection between Tripod and computer,we achieved to transform the whole information of geometries to B&R PCC controller.Its control system of motion calculated the data and generated control instructions to drive the mechanical arms to move.The arms picked up the geometries on the outer ring and placed them into the corresponding holes in the inner ring according to the predetermined path.The accuracy of grasping and rapidity of process were all satisfactory.Experimental results showed the visual system implemented the specific purpose of pick and place accurately and rapidly on the Tripod platform.As a result,the research has great significance to expand the application scope of Tripod Robot.