| Summary: | 碩士 === 國立交通大學 === 光電科技學程 === 106 === The solar industry developement is the common goal and tendency of every country through out the world. To improve the quality and durability of solar module, defects inspection plays a very important role in it , while the failure of module may attribute to the abnormal appearance that occurs during the cell production. Spontaneuously, the inspection is no doubt a critical issue for the solar cell manufacturers in the quality control and pre-shipment inspection . Therefore, for solar module quality and reliability, the appearance defect inspection is quite important
In recent years, many domestic and foreign testing machine manufacturers have invested in the development of many detection speed and both high detection rate of automatic automatic optical inspection (Automatic Optical Inspection AOI) [1 ~ 3], these machines have also been widely developed , The detection of these light source equipment, the main use of LED red light as a major inspection of the flaw detection of light source, red light source system can be effectively detected about 90% of the appearance of solar cell defects at the same time to achieve low false positive rate of stability testing, But in the solar components of the battery chip appearance quality requirements continue to improve, want to improve the effective detection rate to the existing red light source architecture is difficult to achieve at the same time to achieve high detection rate and low rate of judgment and to meet the client module appearance Conditional requirements
In view of this, the research direction of this paper is the application of solar cell flaw detection composite light source, the use of single-crystal silicon chip surface color characteristics, to be tested for the two types of pollution defects (fingerprint, scratch, Acid, blue spot), for (fingerprints, scratches, light sticky) type of defects to the algorithm to simulate the enhanced light enhancement mode, to improve (flaws and battery background) gray-scale contrast difference and effectively highlight the flattery grayscale eigenvalues (Residual acid, blue spot), the experimental results show that the algorithm to simulate the multiples of the enhanced pull gray scale contrast method can improve the detection rate of 30 to improve the detection rate of the new blue light source to improve the detection of red light can not detect defects % (Fingerprint, scratch, light sticky glue), the new Blu-ray flaw detection system can improve the detection rate of up to about 75% (residual acid, blue spot), the two sets of test results can actually use its production line produce
|
|---|
