視覺輔助H型鋼橫向尺寸量測系統之研發

On the Measurement of the Cross-Sectional size of H-Shaped Steel Using a Vision System

林建昌(Chane-Chang Lin)

指導教授 ： 林榖欽

崑山科技大學/工程學院/機械工程研究所/碩士(2004年)

https://doi.org/10.6828/KSU.2004.00011

摘要

型鋼是現代化建設及大眾傳輸系統中重要的建材，我國對型鋼之年需求量很大，尤其近幾年因高鐵建設以及「九二一」大地震後，耐震H型鋼之需求量將顯著地增加。我國有生產各類型型鋼的能力，但生產線上型鋼的橫向輪廓尺寸須仰賴人力做離線檢測，當生產中的H型鋼尺寸不合時，往往因未能即時地修改設定或中斷生產，以致產出大量尺寸不合的H型鋼，導致嚴重的經濟損失。為能避免或減少此類的經濟損失，發展H型鋼橫向尺寸之線上檢測系統實有其必要性，以期能即時地檢測生產中H型鋼的橫向尺寸，並能適時地警告操作員或自動修改設定、或暫時中斷生產流程。
本論文探討一視覺輔助H型鋼橫向尺寸量測系統，系統中包含一組電腦、兩台攝影機、一塊影像擷取卡、一組線性雷射光模組及一組旋轉機構量測平台；以量測H型鋼局部的輪廓外形尺寸，而所有的輪廓尺寸資料經由3D重建來決定H型鋼之橫向輪廓尺寸。本論文引用Zhang[Zhang, 1998]校正法以估算系統之攝影機參數，並建立一適用本系統攝影機校正用之L型校正器，校正器正面黏貼一張由高解析雷射印表機列印出陣列配置的點圖案，各點圖案之中心點即為座標已知可供校正用之控制點。本論文利用Epipolar Geometry以尋找特徵點對應之投影座標，並進而計算特徵點之世界座標。
本論文主要的目的，是發展一線上即時視覺輔助H型鋼之橫向尺寸量測系統，藉由型鋼橫向輪廓之三維空間座標及疊合，重建H型鋼之橫向輪廓尺寸。由本研究的成果顯示，其典型誤差值為-0.317~0.294 mm，誤差百分比約為1 %，此一量測精度較一般型鋼製造之要求精度3 %為低，因此發展立體視覺輔助型鋼橫向尺寸之線上量測實屬可行。

關鍵字

型鋼橫向尺寸量測 ； 立體視覺 ； 攝影機模型 ； 攝影機校正

並列摘要

Shaped steel is the fundamental material in a variety of modern constructions including a massive transportation system. The amount of shaped steel demand yearly in Taiwan is large, especially in the future few years, since the high-speed railway was built and the strong earthquake 921 occured. Taiwan produces shaped steel, but the task on checking the steel cross-sectional size is still conducted off line and by manpower. This can cause significant economical loss if the size of the steel produced is incorrect but cannot be adjusted or quitted in good time. It is desired to develop a system which can sense the cross-sectional size of shaped steel on line and promptly alert human or automatically shut down the process in case incorrect product size arises.
This proposal is to confer a stereo-vision system of measuring the cross-sectional size of H-Shaped steel. The vision system contains a personal computer, two cameras, a frame grabber, a laser-line projector, and a personal device rotator of the measure disk. It is designed to measure a portion of the H-Shaped steel cross-sectional contour. All the contour data is measured by 3-D reconstruction to determine the cross-sectional overall size of the H-Shaped steel. In this article, the camera models and calibration algorithm proposed by Zhang [Zhang, 1998] are utilized .An L-shaped device is designed for calibrating all the cameras in the proposed vision system. The obverse side of the L-shaped device is pasted a paper which is printed a 2-D matrix array of black point pattern with a high resolution laser printer. The control points for camera calibration are the center of the points and their coordinates are known. The paper is using epipolar geometry to find correspondence projection coordinate of feature point, and it can computer world coordinate of feature point.
The paper is a purpose of expanding a stereo-vision system of measuring the cross-sectional size of H-Shaped steel on line. Which is the cross-sectional contour 3-D world coordinate of Shaped steel to overlap and reconstruction is cross-sectional overall size of the H-Shaped steel.The result of the study shows that the experimental result has shown that The experimental results of the study show that the proposed vision approach typically can achieve the measuring error of -0.317~0.294 mm and 1% accuracy. The accuracy fulfills the industrial demand (ie 3% accuracy). It is practicable to develop a stereo-vision system for on-line measuring the cross-sectional size of shaped steel

並列關鍵字

camera model ； stereo vision ； measurement of cross-sectional size of shaped st ； camera calibration.

參考文獻

[35] 陳彥良，“即時立體視覺物體追蹤系統”，中原大學機械工程研究所碩士論文，2003。

[1] D.Nitzan, “Three-dimensional vision structure for robot applications,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 10, No.3,1988, pp. 291-309.

[6] Z. Zhang, “Flexible Camera Calibration By Viewing a Plane From Unknown Orientations”, Proceedings of the IEEE International Conference on Computer Vision, Vol. 1,1999, pp. 666-673.

[7] J. Salvi, X. Armangue, J. Batlle, “A comparative review of camera calibrating methods with accuracy evaluation”, Pattern Recognition 35,2002,pp.1617-1635.

[8] R. Y. Tsai, “A Versatile Camera Calibration Technique for High -Accuracy 3D Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses” , IEEE Trans. on Robotics and Automation, Vol. 3, No. 4 ,1987, pp. 323-344.