PCB (Printed Circuit Board) defect detection method applied to finite computing power edge equipment
Patent Information
- Application Number
- CN202411440463.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2025-01-28
AI Technical Summary
The prior art has problems such as low detection efficiency, high labor costs, difficult data recovery, high equipment requirements and large energy consumption in PCB circuit board defect detection, especially in edge limited computing equipment.
Through training large parameter models and using knowledge distillation technology, it is trained to train small parameter models suitable for marginal limited computing power equipment to achieve circuit board defect detection. This method includes collecting and preprocessing data, training YOLOV5 models, and generating small parameter models through knowledge distillation technology, and finally transplanted it to the edge device for detection.
The efficient and automated testing on the edge limited computing power equipment has significantly reduced the problems of labor costs, missed inspections, misunderstandings, and has the characteristics of flexibility and decentralization. ability.
Smart Images

Figure CN119379633A_ABST