· Detection object: proton membrane surface coating products

Using a large line-scanning camera and a high-brightness light source, after the coating is dried and before entering the compound process, the encoder is used to trigger the camera to take pictures, and the acquisition card is used to obtain and link the pole piece images, and the surface of the coated surface is detected by software algorithms Quality, automatically identify defects and determine whether it is NG or not, label and mark NG incoming materials by itself, and self-check whether the label is affixed. At the same time, it has the function of statistics and storage of bad data.

· Detection accuracy: defect size>0.3mm×0.3mm.

· Applicable products: Detection of the proton membrane after coating and drying.

· Detection accuracy rate: missed judgment rate ≤ 1%, false positive rate ≤ 1%.

· Model switching: the entered model switching ≤15min.