- 學位論文
利用影像處理改進焊道循邊控制精度
An Enhancement of Contour Welding Precision Using Image Processing
摘要
本研究利用影像視覺對於物件的接縫處進行攝影，而後循邊設計焊道，並將此軌跡傳送到伺服XY Table，對物件進行循邊燒焊的仿真驗證，確認燒焊軌跡的精度無誤。在實驗過程中瞭解到影像解析度不足將影響焊道的精度，因而使用可調式高畫素的變焦鏡頭及背光源進行影像拍攝，並配合自行自製的暗房設備防止外部光線的干擾，如此可以清楚辨識出物件接縫處的圖像。 本文量測的對象是直線型的切割物件，切割長度約數十公分。在影像處理過程中，先利用NI Vision使彩色的影像變成灰階，得到每張影像以8 Bit灰階值所構成的。之後透過影像處理對物件的接縫寬度進行量測是否有超出所預期的設定值，量測結果會顯示在PC上。此結果乃經由長期分析不良品的接縫寬度，歸納出每張圖像所測得的接縫寬度可適合燒焊的最大值與最小值，由這項數據可以辨別出物件的良品與不良品。在影像擷取過程中，還包括了Pattern Matching的步驟，以判斷物件擺置位置是否有偏移。之後依照圖面上接縫的長度，依序決定熔接的起始位置、熔接點數、熔點間距與熔接時間，進而在影像上產生循邊的熔接焊道軌跡。最後本研究利用雙軸同動模式XY Table做為熔接軌跡測試。首先，在Y軸架設簽字筆，並將所列印出來的物件影像固定在平台上，XY Table會依照PC上所顯示的焊道軌跡畫在物件圖面上的接縫處進行軌跡模擬驗證。這項研究可以減少固定物件機具和焊槍的調整時間，提升焊接品質及減少不良品的發生。
並列摘要
In this study, image processing technology is used to record images of the seam and the welding track is then designed to follow the contour of the seam. The track data is then transmitted to the server XY Table for emulation verifications to confirm the precision and absence of error in the welding contour. Throughout the course of the experiments, it was discovered that inadequate image resolution lowered the precision of the weld run, thus a variable-focus lens, which is adjustable and provides high resolution, and back-lighting was used to capture the image. The shooting was conducted in a dark room facility that was built in the lab to prevent interference from peripheral light sources. An experimental ensemble like this enables the team to take clearly identifiable images of the seam between objects. In the experiment, objects tested were specifically shapes with straight-lined and partial straight-lined cuts whose lengths reach up to multiples of ten centimeters. During image processing, NI Vision first desaturated colored images to gray-scale, producing 8-bitimages of black or white gradients. Then the seam widths between solids were measured via image processing to see if the measurements have exceeded the expected values. Results were then displayed on the PC screen. The results available (here) were collected over a long period of time by analyzing seam widths of defective products, inducing maximum and minimum welding values suitable for the measured seam width of each image. Whether a product is good or defective can be determined with this data. During every process of capturing an image the Pattern Matching step is included to diagnose any displacement. It is followed by generating the welding track of the contour on the image. Due to the fact that no welding machine has been designed for this specific topic, the team makes use of XY Table with synchronized dual-axis module for welding track tests. First of all, mount a felt tip pen Y Axis hand, place the tested object on the platform and then the XY Table will draw the welding track on the seam of the object image according to the welding track that has been displayed on the PC so that track simulation can be verified.This study successfully reached its target of designing a system that can decrease the time spent adjusting fixed object mechanism and the welding blowpipe. Also, the system can raise welding quality while lower tendencies for defects.