Company News

"In the production process, improving efficiency and yield rate is key. Using AI to replace the human eye aims to reduce errors from manual judgment. On the other hand, human experience varies, and we hope to standardize accumulated experience through algorithm models to improve accuracy." ---- He Jun, CEO of Gechuangdongzhi, quoted from Shanghai Securities News China Securities Network

On April 28, Gechuangdongzhi, Tencent Cloud, and TCL CSOT accepted a joint interview with a media group from Southern Daily, Shanghai Securities News China Securities Network, Securities Times, First Financial, and Jiemian on the ADC project.

TCL CSOT Factory

| Gechuangdongzhi/Tencent Cloud/CSOT Jointly Develop ADC Project

In 2018, Gechuangdongzhi and Tencent Cloud collaborated to help CSOT implement AI in AI-based defect judgment, fully undertaking the ADC system implementation projects for t1, t2, and t6 factories in certain processes. The ADC project is the first AI-based automatic defect classification system applied in the LCD panel industry.

Panels are precision instruments with very high requirements for product yield, but some subtle defects are tiny and hard to detect. In the past, defect detection at CSOT largely relied on manual judgment, with a dedicated team of employees examining images on computers. Using human eyes to judge results from AOI (Automatic Optical Inspection) equipment to classify defects based on morphology and determine quality compliance was tedious. This process faced several challenges: recruitment difficulties, time-consuming training, and room for improvement in the stability of manual judgment accuracy.

Now, over a hundred algorithm models have been deployed at CSOT. With AOI equipment and algorithm models, it’s like equipping the production line with 'eyes' and a 'brain.' Using AI technology for defect identification not only greatly improves detection efficiency but also eliminates issues like missed detections and misjudgments. Compared to traditional manual judgment, AI recognition speed has increased by 5-10 times, reducing manpower by 50%. This is a significant change in the production process after CSOT adopted the visual inspection system jointly developed by Tencent Cloud and Gechuangdongzhi.

Using ADC, anomaly interception time is reduced by 1-2 hours, and production cycle at inspection stations is shortened by 60%

| Introduction to Gechuangdongzhi ADC Visual Inspection Tool

The ADC product is based on Gechuangdongzhi's self-developed Tianshu AI, an intelligent application tool that uses AI technology to address the shortcomings of traditional AOI in complex scenarios. Through ADC, benefits include reduced manual input, improved judgment efficiency, enhanced accuracy, lower missed detection rates, and overall increased comprehensive benefits.

Current application fields: semiconductor, chip, textile, automotive, solar energy, and other industries.

Original media report links:

1. Southern Daily Report: Digital New Momentum⑧| Equipping Industry with Eyes and a Brain, Tencent Uses AI to Read Defects

https://static.nfapp.southcn.com/content/202004/30/c3477856.html

2. Shanghai Securities News China Securities Network: TCL CSOT Leverages Tencent Cloud to Create a New Infrastructure Implementation Model

http://company.cnstock.com/company/scp_gsxw/202004/4527327.htm

3. First Financial: AI Replaces Quality Inspectors to Improve Efficiency, but Industry Expects More from Industrial Internet

https://m.yicai.com/news/100614053.html

Click to read the original article and apply for a free product experience~

https://jinshuju.net/f/YRKWJ1