| Nowadays, problems of energy and environment are increasingly serious, new source ofenergy, especially solar energy, are becoming increasingly popular for their good magneticproperties. The solar cells as the final product of photovoltaics, directly effect photovoltaic’sdevelopment. In the production process of solar cells, from ingots making to moduleencapsulation, the defects injected on each link will produce bad effects on the solar cells, forinstance, reducing transfer efficiency or shortenning service life, even becaming waste cells.Based on the technique of using infrared imaging for solar cells defect detection, this paperpresents a design scheme of image processing and data high-speed acquisition based onhigh-speed DSP micro-processor.First of all, the article gives an account of the manufacturing processes of solar cells,various defects which appears easily in solar cells manufacturing are summarized. Then, onthe basis of classical experimental instrument for electroluminescence, the DSP has been usedin image processing, and a design scheme of on-line inspection and recognition system ofSolar Cell Defects based on DSP is presented, including the hardware structure schema andthe software structure schema. Image pretreatment and recognition has been introduced,including image filter histogram equalization, image sharpening, image segmentation andfeature extraction, recognition, etc. Then, it completes the feature extraction with the methodof Hu moment invariant. The research uses a modified moment invariant algorithm tocomplete the feature extraction. Besides, the infrared image library also has been set up. Thearticle uses the method of support vector machine to complete the classifier training of thedefections. The results are very good. In order to reach the goal of human-computerinteraction, the article make a diagnostic system design of soar modules based on thegraphical user interface window of matrix laboratory. |
PDF Full Text Request