RV 2AJ-Classify and pick Objects using 3D pointcloud
With the rapid development of modern industry that requires increased productivity, the application of other technologies to the line with the robot is indispensable in that computer vision is a growing array of “eye”for the robot to perform the task. Currently bin-picking is divided into two small arrays, 2D and 3D:
Bin-picking 2D: Objects usually have the same shape and height. Often used for picking objects that run on conveyor belts. The system consists of a fixed camera with fixed height and a robot that is essentially a parallel robot or scara. This direction is now being used a lot in theindustry.
Bin-picking 3D: Objects often have the same size, shape is not too complex randomly placed in a certain space, objects overlap each other, so the need to determine the depth of the object. The system includes a depth sensor that returns the point cloud, a 5DOF or 6DOF tripod, and an additional camera that can pinpoint the location after picking to place the object exactly where it is needed.
Figure 1: Model
Experimental models include: Camera 3D, camera 2D with manipulatorRV-2AJ 5DOF of Mitsubishi. Samples used are solid rectangular blocks.