Research On Scrap Metal Sorting Robot With Machine Vision

The scrap metal recycling is an important part of the development of circular economy. Sorting is an important means to realize the effective utilization of scrap metal. In this paper, an intelligent robot based on machine vision is developed for sorting non-ferrous metal block. The main work and research contents are as follows:1. A non-ferrous scrap metal sorting line is designed based on machine vision+conveyor grasping method. The open control system is setup on base of EtherCAT communication and CodeSys developing software.2. After the analysis of the theoretical knowledge of target detection and positioning in monocular vision, the calibration experiment of machine vision is made and an object position and angle measuring softeware is developed using OpenCV library, provideing the location for conveyor grasping online.3. To maximize the working efficiency, a new real-time pyramid search algorithm is presented for optimal time rendezvous point in the area of robot picking on moving conveyor belt. Experiments on a practical sausage sorting line show that it can achieve over10percent promotion in sorting efficiency. Furthermore, a dynamic conveyor picking strategy is analyzed and developed based on a changeable pickup limit line to self-regulate conveyor speed with the density of the targets on moving belt. Tests show that it is especially effective in sorting objects coming on belt at random.4. In reality it is very hard for the designed and constructed PMSM to produce the ideal sinusoidal back electromotive force (back-emf), thus resulting in higher torque ripple by using conventional vector control scheme. To cancel the torque ripple, this paper analyzes the instantaneous electromagnetic torque production with Park-like transformation and presents a cost-effective method to generate constant propulsion with optimal current commands in a lookup table, which is offline calculated on non-ideal sinusoidal back-emf. The target capture accuracy can be improved by a more stable speed of the conveyor belt.5. In this part, the method of trajectory design with electronic cam is studied. After comparing the characteristic parameters of the three kinds of curves, five order polynomial curve is selected for manipulator speed motion planning. The steps and modules are expained to realize the dynamic grasping based on electronic cam.6. After a careful study of the proportional guidance method, the technology is rebuilt for dynamic grasping operation on the conveyor along the shortest path to moving target, and Matlab simulation experiment is made for linear tracking intercept target method.