1. Project description

This work is one of the projects I was responsible for during my work in 2019 at Beijing, China. The goal of this project is to develop an appearance defect detection system with the deep learning tools of ViDi Suite from Cognex. The system can be deployed in the production line to automatically detect the different types of defects of soap.

Samples of the soap

Samples of the defects

2. Responsibilities

• Developed an intelligent appearance defect detection system based on ViDi Suite from Cognex.
• Deployed the system in the production line.

3. Contributions

• Designed the scheme of appearance inspection according to the requirements of the customer.
• Collected and labeled the real samples of the soap from the production line.
• Designed the intelligent appearance defect detection system with the tools of ViDi Suite.
• Trained and deployed the system.

Tools: C#, ViDi Suite

4. Some Results

• Setup:
  • Industrial camera: Basler 1.3MP
  • Lighting: polarized strip light
  • GPU: NVIDIA RTX2070 8G
  • CPU: Intel i7-8700K

• Scheme layout of the appearance defect detection system
  
  

• Results of the appearance defect detection system