In the consumer goods industry, the appearance inspection of home appliance panels is crucial. WeLinkirt brings new breakthroughs to the appearance inspection of home appliance panels with its advanced AI AOI software system.
User Scenario: A leading home appliance manufacturer, on its home appliance panel production line, mainly produces operation panels for various home appliances. The inspection objects are the appearance of home appliance panels, including surface scratches, stains, printing defects, and other flaws. These panels are widely used in common home appliances such as refrigerators, washing machines, and air conditioners, and their appearance quality directly affects the overall image and user experience of the products.
Pain Points: Under the traditional inspection method, the manufacturer faces many difficulties. The miss rate is relatively high, about 3%, which causes some defective products to enter the market and affects the brand reputation. The false alarm rate also reaches 20%, which makes a large number of qualified products need to be re-inspected, increasing labor and time costs. At the same time, when producing different models of home appliance panels during model change, a large amount of time is required for reprogramming and debugging, and the model change time is as long as 30 minutes. Combining with the current hot topic of improving inspection efficiency and accuracy in chip production using industrial vision inspection, the manufacturer urgently needs a more efficient and accurate inspection solution.
Technical Principle
WeLinkirt's AI AOI software system uses an advanced visual foundation model for feature recognition. This model is trained with a large amount of image data and can accurately identify various features of the appearance of home appliance panels. During the inspection process, the system uses the APDT positive-sample/few-sample learning algorithm. Only 1-20 good samples are needed, and it can quickly learn the features of normal panels. This is because the algorithm can extract key feature information from a small number of samples and use it as a benchmark for inspection. At the same time, the semantic false alarm filtering technology can effectively filter false alarms based on the semantic information of defects, such as the length and width of scratches, improving the accuracy of inspection.
- The visual foundation model can conduct in - depth analysis of complex appearance features, and even tiny defects can be accurately identified.
- The APDT positive-sample/few-sample learning algorithm reduces the dependence on a large number of samples and improves the adaptability and flexibility of the system.
- The semantic false alarm filtering technology avoids false alarms caused by environmental factors or minor interferences through understanding the semantics of defects.
WeLinkirt's Solution and Product
Centered on the AI AOI software system, WeLinkirt provides a complete solution. This software system has the ability of zero-code automatic programming. Only one good product is needed, and the programming can be completed within 5 minutes, greatly shortening the model change time. At the same time, the system supports SDK/API/Docker deployment and can achieve 100% local privatization, ensuring that data does not leave the factory and meeting the enterprise's data security requirements. In terms of supporting facilities, it can be combined with the DaoAI 2D / 3D AI AOI equipment, which uses its self-developed 3D camera and 3D topography reconstruction technology to detect hidden defects of the panels.
The AI AOI software system brings new changes to the appearance inspection of home appliance panels with its efficient and accurate inspection ability.
Quantitative Results: By using WeLinkirt's AI AOI software system, the manufacturer's inspection results have been significantly improved. The detection rate has increased to 99%, and the miss rate has been reduced to <1%, effectively preventing defective products from entering the market. The false alarm rate has been reduced by -60%, reducing a large amount of re-inspection work and improving inspection efficiency. The model change time has been shortened from 30 minutes to 5 minutes, greatly improving the flexibility and response speed of production.
FAQ
How many samples does the AI AOI software system need for learning?
The system uses the APDT positive-sample/few-sample learning algorithm. Only 1-20 good samples are needed to quickly learn the features of normal panels, reducing the dependence on a large number of samples.
How much can the model change time of the system be shortened?
The system has the ability of zero-code automatic programming. Only one good product is needed, and the programming can be completed within 5 minutes, greatly shortening the model change time from the traditional 30 minutes to 5 minutes.
How does the system ensure data security?
The system supports SDK/API/Docker deployment and can achieve 100% local privatization, ensuring that enterprise data does not leave the factory and meeting data security requirements.