Image acquisition technology of mobile phone FPC!

As the core component of mobile phone products , mobile phone FPC plays a key role in the quality of mobile phone products. As a new and fast defect detection method for mobile phone soft board, automatic optical inspection can avoid various defects of manual visual inspection, but it is not enough for the collected data. The image processing requirements are relatively high, and a slight improper image processing method will make the false detection rate of the FPCB detection very high. Therefore, the selection of image preprocessing and segmentation methods has become the key technology of automatic optical detection of mobile phone FPCB defects.

In the process of image acquisition, due to the influence of foreground light, background light and ambient light source, as well as the relative humidity and temperature in the environment, there will inevitably be differences between the acquired image and the real image; at the same time, in the process of image transmission and conversion, may cause image quality degradation. Therefore, the acquired degraded images must be improved, including image smoothing and image sharpening.

The system uses the ring light as the light source, and adopts the combination of front light and back light. The upper light source is used as the forward light source to obtain the surface features such as the wires, pads, and vias of the flexible board ; the lower light source is used as the back light source to strengthen the wires. Contour dimensional features. At the same time, the Microvision MVC3O00 camera is used (its basic performance parameters are: resolution 2048x1536, 10 frames per second, CMOS sensor, USB interface), supplemented by a stepper motor to drive the workbench to realize the block acquisition of the soft board. The function of image smoothing is mainly to suppress the noise of the image. The noise in the mobile phone soft board image is mostly noise points, that is, salt and pepper noise. According to the principle of median filtering, it can be known that through the method of data sorting, the probability of replacing the value of the noise point with the point that is not polluted by noise in the image is relatively large, and the noise The suppression effect is very good, the picture clarity is basically maintained, and the algorithm is concise and the detection speed is fast, which is suitable for real-time detection.

Any FPC request email to [email protected]