AI AOI Software System Solves Glass Surface Defect Detection with Few Samples

In the chemical and materials industry, the detection of glass surface defects is crucial. Traditional detection methods perform poorly with few samples. WeLinkirt's AI AOI software system provides an effective solution to this problem.

User scenario: A leading manufacturer in the chemical and materials industry has a glass production line. Its main products are various types of optical glass and architectural glass. The detection objects are micro-scratches, bubbles, impurities and other defects on the glass surface. These defects not only affect the appearance quality of the glass but also may have a serious impact on the optical performance of optical glass.

Pain points: In the traditional acoustic emission acquisition system, when combined with AI signal processing technology for industrial visual quality inspection, the manufacturer faces many difficulties. Due to the small number of glass defect samples, traditional algorithms are difficult to accurately learn defect features, resulting in a missed detection rate of up to 5% and a false alarm rate of 10%. Excessive missed detections may cause unqualified products to enter the market, increasing the company's after-sales cost. A large number of false alarms require manual re-inspection, which not only consumes a lot of manpower but also reduces the detection efficiency of the production line. The model change time is as long as 30 minutes, seriously affecting the continuity and efficiency of production.

Technical principle

WeLinkirt's AI AOI software system uses an advanced visual foundation model for feature recognition. Based on deep learning algorithms, this model can automatically extract feature information such as the texture and shape of the glass surface. It uses the APDT positive-sample/few-sample learning algorithm. With only 1-20 good samples, it can quickly learn the characteristic distribution of normal glass. When encountering a defect, it can accurately identify the abnormality by comparing with normal features. At the same time, the semantic false alarm filtering technology can effectively filter false alarms according to the semantic information of defects, such as the size, location and shape of defects. This filtering method can not only reduce the workload of manual re-inspection but also improve the accuracy of detection.

• The visual foundation model can perform high-precision recognition and analysis of the microscopic features on the glass surface, and even tiny defects can be accurately captured.
• The APDT positive-sample/few-sample learning algorithm can quickly establish a feature model of normal samples through learning from a small number of good samples, thus better identifying defects.
• The semantic false alarm filtering technology combines deep learning and semantic analysis, and can make judgments according to the actual situation of defects, avoiding false alarms caused by factors such as noise.
• The model uses a combination of unsupervised learning and supervised learning to further improve the detection ability of unknown defects.

WeLinkirt's solution and product

Centered around the AI AOI software system, this system has the ability of 5-minute zero-code automatic programming. Operators do not need to write complex codes. They can quickly complete programming through simple operations to achieve rapid model change of the system. At the same time, the system supports 100% local private deployment of SDK/API/Docker, and the data does not leave the factory, ensuring the security of enterprise data. In practical applications, this system can cooperate with the DaoAI 2D/3D AI AOI equipment to conduct all-round detection of the glass surface. The self-developed 3D camera of the DaoAI 2D/3D AI AOI equipment can achieve three-dimensional topography reconstruction, detect hidden solder joints, coplanarity, micron-level topography, etc., providing richer detection data for the AI AOI software system.

The AI AOI software system brings new breakthroughs to glass inspection in the chemical and materials industry with its efficient few-sample learning ability and accurate defect detection technology.

Quantitative results: After introducing WeLinkirt's AI AOI software system, the detection rate of glass surface defects of the manufacturer has increased to 98%, the missed detection rate has decreased to <2%, and the false alarm rate has decreased by -60%. At the same time, the rapid model change ability of the system has shortened the model change time to 5 minutes, greatly improving the production efficiency of the production line.