On March 18th, at the China (Shanghai) Machine Vision Exhibition, Mr. Huang Kai, sales director of Cogtech East China, accepted an exclusive interview with the reporter of the sponsor. He talked about the company's efforts to develop deep learning technology and how deep learning products contribute to the intelligent manufacturing production line. Cognex's global competitiveness and other topics.

Interview with Cognex

Q: Is there a different theme for Cognex this year? What new products were on display?

A: As an American company, Cognex's theme is Smart Vision for Small Factory. Around the theme of intelligent manufacturing, how to reflect intelligence? Why do we want to show intelligence? Well, there's a reason.

We know that the development of human society is progressing with the progress of production tools, from farming to the steam age, to the machine age, to the current information age, and finally, I think, to the intelligent age. The opportunity is that the production process needs more and more intelligent products to meet the needs of that production. So in this year's show, we're highlighting intelligence in particular. We went from wanting machine vision to be faster and more stable, to being more intelligent and able to find more irregular patterns and patterns. In fact, in more and more production processes, we found that this demand is increasing, from the very beginning of mobile phone 3C to our cosmetics, medicine, including new energy batteries. So we hope that the battery surface of the new energy battery is completely free of any defects. The battery has 6 sides. How to check? Under this premise, we are increasingly launching deep learning products. For example, the D900 series and the L4000 line scanning laser series launched today are some of the products that our company has launched and displayed around this theme.

Q: Cognex has been deeply involved in the machine vision market for many years. How does deep learning help the production line of "intelligent manufacturing"? Please introduce the In-Sigh D900 product features, advantages, industrial application and market prospect based on deep learning technology In detail.

A: Why do we need to develop deep learning? The reason is that in the production process, we do have a part of very critical testing technology. Now we use manual inspection, such as the mobile phone, cosmetics and 3C mentioned just now. Manual testing, there is a very big problem is easy to fatigue. So how to combine the high speed and stability of machine vision with human intelligence, then how to apply this deep learning technology in intelligent manufacturing will be particularly meaningful.

What kind of product is Cognex D900? To solve that problem, it's an integrated deep learning hardware and software solution, the size of a fist. It doesn't need a controller. The big thing about the D900 is that it's integrated. I don't have a small box in the back, just a head. This header integrates the acquisition and processing of images. At the same time, the intelligent camera equipped with deep learning algorithm is one of the few integrated ones in the market at present. It is also a product that we mainly promote in VisionChina this year.