When it comes to the industrial digital transformation, the first thing people think of is the Industrial Internet, as if the Industrial Internet is the whole of industrial transformation.
In fact, with the acceleration of the digitalization process, intellectualization is the future of industrial development. The deep fusion of AI and industry has become an important topic in the academia industry. AI will become a measure for intelligent manufacturing for a long time in the future.
When the concept of intelligent manufacturing was born, people were exploring the application of AI in various scenarios of industrial manufacturing, such as intelligent sorting, intelligent inspection, energy efficiency optimization, preventive maintenance, intelligent defect detection, etc. Among the applications, the quality inspection is a challenge in the process of achieving intelligent manufacturing.
According to the report, the market of AI powered industrial visual quality inspection has entered a growth period in China. The market size of China’s industrial quality inspection software and service has reached 142 million U.S. dollars, a growth of nearly 32% compared with 2019. The market size will maintain a growth rate of more than 30% in the next five years.
This is mainly due to the fact that China has many industrial subdivisions and the large differences in R&D, production, and management and other links in each field. For industrial scenarios, the biggest challenge of integrating AI is that the AI applications requires equipment, network and computing power as the basis condition. However, the factory cannot spend huge costs to transform the production line and deeply integrate with AI.
Another reason is that quality inspection has always been a rigid part of product internal control. As we all know, most factories now rely on workers to complete quality inspection, with a large amount of repetitive labor and cost. In terms of labor quality inspection, the difficulties in recruiting and employing labor is a problem for many factories.
These quality inspectors spend a lot of time every day to inspect the quality of industrial parts, which is not only harmful to the eyesight, but also has problems such as poor speed and stability. Meanwhile, traditional industrial quality inspection makes judgments based on human eyes and subjective experience, which cannot form refined inspection data to assist process optimization. And manual inspection experience is difficult to replicate and inherit.
Compared with traditional manual quality inspection, AI quality inspection has the advantages of high efficiency, high detection accuracy, and stability. In the process of industrial upgrading, using AI to complete quality inspection is undoubtedly the best choice.
Take contact lenses as an example. Most manufacturers are using random sampling methods to test whether the products are defective. But this method is not applicable in the production line of contact lenses, as every lens needs to be inspected. The quality control personnel can only inspect up to 4000 lenses each shift, which creates a production bottleneck. In addition, false detections and missed detections of manual inspection are inevitable.
Since contact lenses are transparent, the detection method using machine data has always been a major challenge for this industry. Traditional AOI relies on fixed geometric algorithms to find defects, but it is difficult to obtain high-quality images from transparent objects. As a result, the detection performance cannot be accepted by customers.
Using AI-based smart cameras to collect data to train AI algorithms, and continue to iterate the performance of detection has been proven better solutions. The quality inspection system based on AI can identify common defects, such as burrs, bubbles, rough edges, particles, scratches, etc.
As shown in the photo above, the AOI based on AI can even detect tiny defects in transparent contact lenses. Compared with the manual quality inspection process, the inspection efficiency has been significantly improved. Compared with manual visual inspection, each smart camera can detect more than 50 times the number of contact lenses, and the detection accuracy has been increased from 30% to 95%.
As the amount of data increases, the accuracy of AI detection will continue to be optimized and provide data support for inspection process optimization, which will fully empower the industry.
The accurate recognition of AI based quality inspection requires large amount of high-quality training data. As a world’s leading AI data services provider, Datatang has deployed 3 data labeling centers in China, and has more than 5000 experienced data labeling experts. Datatang provides customers with high-quality training data and helps customers quickly improve the accuracy of AI quality inspection.
As industrial manufacturers are using computer vision technology to detect industrial products and continuously improve the efficiency and accuracy of quality inspection system, AI will be a powerful tool in the process of intelligent manufacturing.
If you need data services, please feel free to contact us: email@example.com