Random Picking Solution —— 3D Vision-Guided Weld Workpiece Randomly Picking

The new-generation vision solution based on the Mixed AI physics engine can effectively cope with complex working conditions such as a large variety of workpieces, reflective surfaces, complex structures, disorderly stacking, and ambient light interference.

ꄴPrevious： Loading And Unloading Solutions - Component Loading And Unloading

ꄲNext： null

3D Vision-Guided Weld Workpiece Random Picking

Video For The Project

In recent years, the traditional automobile industry seeks to upgrade, new energy vehicles continue to develop, and the transformation of the automobile industry has created a large number of automation needs. Retrofitting with high standard requirements, complex process flow and a large variety of workpieces are long-term challenges faced by the automation of the automotive industry. Various typical transformations, such as workpiece loading and unloading, positioning assembly, etc., have high requirements for technical indicators such as cycle time, precision, and stability.

Project Description

Welding production in automobile manufacturing has large batches, many types of welded parts, fast production speed, and high precision requirements. It is necessary to support flexible production changes of various welded parts while considering accuracy and performance.

Project Specification

Beat: <2s, first <10s
Accuracy: <=3mm, <=3 degrees
Stability: >1500 times per day
Accuracy: 1500 times without recognition error

Solution Advantage

Using the Mixed AI platform for data generation and labeling, the trained model has fewer parameters; and the exclusive support provides CAD to quickly train a new algorithm model, compatible with a variety of different types of workpieces, and supports flexible production changes.

Greatly reduce the work intensity of front-line operators, can improve production efficiency and reduce costs.