近年來,由於看好太陽電池產業的發展,國內有許多廠商都投入此產業,並且已形成完整的產業鏈。在太陽電池的整個製程當中,可以進行的自動檢測有許多,本論文針對太陽能電池表面瑕疵與內部瑕疵進行自動化檢測之研究。在進行表面瑕疵檢測時,本研究使用CCD相機擷取太陽能電池影像,接著辨識出導線的位置,進而找出表面瑕疵,根據實驗結果顯示,本系統可成功檢測出污漬、斑點、刮痕、白點、佈線不良及水平斷線等瑕疵。在檢測太陽能電池內部瑕疵時,本研究所提出的方法是先對電池施加偏壓,接著以CCD擷取太陽電池所散發的近紅外光譜影像,藉由處理此影像進而辨識出電池內部的瑕疵,根據實驗結果顯示,本研究可成功檢測出以人眼視覺無法找出的內部瑕疵檢測。為了提升檢測速度,本研究進一步提出以平行處理方式來加快檢測速度,並且利用繪圖卡來實現平行演算法,實驗結果顯示,以繪圖卡實現平行演算法可以有效提高檢測速度。
Solar energy technology is one of the most important areas in these years. Many countries put plentiful resources and budgets in this industry. Lots of companies in Taiwan devoted to this area as well, and thus shaped a solar energy industry chain. Since most Taiwan companies in this industry focus on mass production, quality control is thus an important task for these companies. Lots of inspection tasks are included in the quality control process, we propose a system which automatically detects surface and internal defects in solar cell. The surface defect inspection system uses a CCD camera to take images of solar cells, recognizes fingers on the surfaces, and identifies defect positions. Experimental results show that defects such as Stains, spots, scratches, white spots, bad wiring, and breaks in fingers, are successfully detected by the proposed system. On inspecting internal defects, we propose to apply bias flow to the solar cell, captures emissions of solar cell, and processes captured image to recognize the internal defects in the solar cell. The experimental results show that the proposed system can successfully detect internal defects which can not be found by visual inspection.To further increase the inspection speed, we proposed parallel processing algorithms and implement them in graphic card. Experimental results show that the inspection speed is successfully increased by using parallel processing algorithms.