Item description:

This project has a large number of detection object, and most of the defective defects are so small that they can hardly be judged correctly by the naked eye. Based on the deep learning algorithm of Handdle AI and the visual solution provided by artificial intelligence technology, Dongsheng finally achieves the detection accuracy of more than 99.99%. The CT requirement: 1.5 seconds/box.

     Defect types:

Bubbles, thin wires, loose ends, poor joints, open wires, sundries, etc.

                                                                OK sample picture                                                                                                                                                                                         Inspection picture

                  Overall intelligent detection, with an increase of 35% in production efficiency                 Continuous optimization of model