Flaw detection for disorderly feeding of hub

Publisher: Shiyan  Time:2021-12-11 17:49:03

Case background

The hub workpiece is placed in the material frame disorderly, fed to the belt pulley manually, and whether it needs to be turned over is confirmed according to the front and back of the workpiece, and then sent to the flaw detector for flaw detection. The self weight of the workpiece is 10-20kg, which is completely loaded and unloaded manually according to the beat of 8s. The labor intensity of the workers is high, and they need to bend down constantly, which causes great damage to their bodies. Therefore, automatic transformation is required.

Technical difficulties

01. The terminal incoming material is 1000 * 800 * 800mm full frame incoming material mode, and the workpiece is completely disordered in the frame.
02. Three types and 16 types of workpieces need to be flexible. There will be dirt and other interference on the workpiece surface.
03. The beat of the whole gripping workpiece shall be controlled within 7 seconds, and the gripping accuracy shall be m ±1。

Solution

  • 01. The visual part adopts 3D structured light camera + depth learning algorithm.
  • 02. Use 3D vision to locate the products in the material frame to avoid collision points and generate grab points.
  • 03. 2D vision is used for accurate positioning twice to ensure feeding accuracy.
  • 04. The 20kg manipulator is equipped with variable angle gripper to meet the workpiece grasping at different angles.


Copyright © Shiyan Intelligent Technology (Guangzhou) Co., Ltd.
top