In the wave of intelligent manufacturing, industrial vision inspection is becoming increasingly important for the electronics/PCBA industry. WeLinkirt's AI AOI software system provides an effective solution to solve PCBA assembly problems.
User scenario: The PCBA assembly line of a leading electronics manufacturing manufacturer mainly produces printed circuit board assemblies (PCBA) for various electronic products. In the assembly process, it is necessary to detect missing components on the PCBA and incorrect installation of connectors. The detection objects include various surface-mount components, through-hole components, and connectors.
Pain points: In the current wave of intelligent manufacturing, traditional PCBA detection methods face many difficulties. Previously, the manufacturer used a combination of manual visual inspection and traditional AOI equipment. The missed detection rate was as high as 3%, and the false alarm rate reached 20%. This not only led to a large number of products needing re-inspection, increasing labor costs but also affecting production efficiency. In addition, when changing product models, the traditional programming method took more than 30 minutes for reprogramming, seriously affecting the flexibility of the production line. At the same time, as the industry's requirements for product quality continue to improve, traditional detection methods are difficult to meet strict compliance standards.
Technical principle
WeLinkirt's AI AOI software system is based on an advanced visual foundation model for feature recognition. Through deep-learning algorithms, the model learns and analyzes a large amount of PCBA image data and can accurately identify the features of various components. For missing-component detection, it can quickly detect components that should exist on the PCB but are actually missing; for incorrect connector installation detection, it can make accurate judgments based on the shape, position, and pin features of the connector.
- Five - minute zero-code automatic programming for a good product: The system uses advanced image analysis technology to automatically learn and program good products in a short time without writing complex codes, greatly shortening the programming time.
- APDT positive-sample/few-sample learning (1-20 good products): With only a small number of positive samples (only 1-20 good products), the system can quickly learn the normal features of the product and effectively identify defective products.
- Semantic false-alarm filtering: The system can filter false alarms based on semantic information, reducing unnecessary false alarms and improving the accuracy of detection.
- Support for SDK/API/Docker for 100% local private deployment: The system supports multiple deployment methods and can achieve 100% local private deployment, ensuring that data does not leave the factory and protecting the security and privacy of data.
WeLinkirt's solution and product
Centered on the AI AOI software system, a complete PCBA detection solution is provided for the manufacturer. The system can be integrated with the factory's existing detection equipment for rapid deployment. When changing product models, new programming can be completed in only 5 minutes, greatly improving the production line's model-changing efficiency. At the same time, the APDT positive-sample/few-sample learning ability of the system allows model training to be completed with only a small number of good-product samples, reducing the difficulty and cost of sample collection. In addition, the semantic false-alarm filtering function effectively reduces false alarms and improves detection efficiency. The supporting DaoAI 2D/3D AI AOI equipment can provide more comprehensive detection and accurately detect hidden solder joints.
The AI AOI software system has become a powerful tool for improving the quality and efficiency of PCBA detection with its efficient programming method and accurate detection ability.
Quantitative results: After applying WeLinkirt's AI AOI software system, the detection rate of the manufacturer's PCBA detection increased to 99%, the missed detection rate decreased to less than 1%, the false alarm rate decreased by -70%, and the product model-changing time was shortened from more than 30 minutes to 5 minutes, greatly improving production efficiency and product quality.
FAQ
How many good-product samples are needed for the training of the AI AOI software system?
The system uses APDT positive-sample/few-sample learning. Only 1-20 good products are required to complete model training, greatly reducing the difficulty and cost of sample collection.
How long does it take to program the AI AOI software system when changing product models?
The system supports five-minute zero-code automatic programming for a good product. When changing product models, new programming can be completed in only 5 minutes, improving the production line's model-changing efficiency.
How does the AI AOI software system reduce the false alarm rate?
The system has a semantic false-alarm filtering function, which can filter false alarms based on semantic information, effectively reducing unnecessary false alarms and improving the accuracy of detection.