What are the blind spots of 3DAOI during testing?

The development of AOI technology in the direction of intelligence is an inevitable requirement brought by the development of SMT. Under the development characteristics of SMT's miniaturization, high density, rapid assembly, and variety diversification, the amount of detection information is large and complex. Whether it is in terms of real-time detection feedback, or in terms of the accuracy of analysis and diagnosis, it depends on manual verification. It is almost impossible to analyze and diagnose the quality information obtained by AOI. Intelligent AOI technology that replaces manual automatic analysis and diagnosis has become an inevitable development.

Research on the feature extraction and defect recognition and classification of various defects; in view of the small defects to be detected in the high-density PCB visual inspection system, the variety of defects, the feature is not easy to determine, etc., the feature extraction technology and Research on various classification methods, use machine learning methods, design different classifiers, compare and analyze the classification effects and errors of different classifiers, and use optimized classifiers to achieve rapid detection and accurate classification of defects , And improve the intelligence level of the classifier as much as possible.

Nowadays, more and more factories are equipped with AOI testing equipment to ensure the quality of products. The basic principle is to use light reflection to check whether components such as BGA are mounted correctly, whether the soldering is good, whether there is missing, reverse or short circuit. Defects such as phenomenon. When a defect occurs through the AOI inspection, the on-site staff needs to make a visual judgment. After the judged good product, the test result is manually modified to be qualified, and subsequent production is continued, while the defective product is repaired.

So, what are the blind spots of 3D AOI in the testing process?

1. Welding: the reason is that the crotch light of the components causes insufficient light reflection, which is difficult to grasp with any of the five cameras;

2. And corner distortion: the reason is that the gloss difference between the pin and the solder is not very large, and it is difficult to grasp the corner distortion and warping;

3. Empty welding: The reason is that although empty welding can also be caught, the false judgment rate of empty welding is relatively high. Only when the parameters are set properly can be caught, and the misjudgment is reduced;

4. Lifting feet: The reason is that the board will expand correspondingly after the PCB board is out of the reflow furnace, which causes a misunderstanding. Although the image comparison method can be used to catch it, it is relatively misjudgmental and more troublesome to grasp;

5. Multiple items: The reason is that AOI can only set the detection frame on the coordinates we give, and only grab the detection frame where there is no detection frame and not grab it;

6: Too much tin: The reason AOI detection has no area range, no matter what kind of detection frame is used, it is impossible to catch tin at the same time, but only a little;