本文針對在戶外拍攝時會產生的影像角度及光線不均的情形做研究，並在影像拍攝後以機械視覺的方式做圖像的優化，使光線的影響變小，且能做到影像偏移角度的偵測。並透過程式邏輯的判斷可以辨識出9種不同的情形及7種缺陷。透過本文發展的系統可以初步的為焊接表面做一個簡單的檢測。並且可以應用在建立焊接品質的基本規範上。;As the improvement on machine vision, we can use the technique to help us assure the quality of producing and the efficiency of manufacturing.
We developed a system called “AWI” (Aeronautic Welding Inspection system) to improve welding quality inspection. Through this system, some welding products without normal production line, such as aeronautic repairing or building structure could be inspected. For users convenience, we use portable device like cell phone、 camera to take pictures.
This system aims at fixing the angle leaning and uneven brightness by optimized images when using outdoor shooting. It can detect 9 cases and 7 external defects of welding component through this system.
