應用電腦視覺於道路標線品質評鑑之研究

Applications of Computer Vision on Road Marking Quality Assessment

陳佶宏(Chi-Hung Chen)

指導教授 ： 林國良

義守大學/理工學院/土木與生態工程學系/碩士(2014年)

https://doi.org/10.6343/ISU.2014.00287

摘要

道路標線依據公共工程委員會所頒之施工綱要規範第02898章與交通部公路總局公路養護手冊之規定，標線完工後外觀驗收以及後期是否進行養護的評估方式主要依賴檢查人員之目視判斷，並無系統且科學化的評估方法與準則，因此容易產生標準不一之疑慮。本研究利用電腦視覺技術建立一套客觀的自動化道路標線品質評鑑系統。在此研究中以數位相機擷取道路標線影像，再運用影像處理技術與系統化的二維量測方法來分析標線的幾何特性，最後給予適當評分，藉以判斷道路標線繪製品質之優劣。
　　本研究建立之道路標線品質評鑑系統針對兩個幾何面向進行分析：(1)表面劣化程度：標線之影像灰階值被用來作為表面劣化程度判斷之依據；本研究首先擷取優良的道路標線影像，分析其灰階值訂定「基準灰階值」，接著再將欲量測之標線實測灰階值與「基準灰階值」進行比較，實測灰階值與「基準灰階值」之偏差越大則代表標線表面均勻度越差。(2)邊緣平整度：以標線邊緣點與迴歸直線間之平均離差作為邊緣平整度判斷之依據；本研究運用J.F.Canny所提出的邊緣偵測方法尋找出標線之邊緣點座標，接著計算邊緣點與迴歸直線之平均離差，當平均離差越大則表示邊緣平整度越差。本研究利用Matlab分別將標線表面均勻度與邊緣平整度之分析程序建立了自動化平台，使用者只要匯入標線影像、選擇分析區域後，系統即可自動產生0-100之量化評估結果。
　　本研究針對各種不同品質之道路標線進行分析比較，驗證了系統之可行性，且此自動化系統讓道路標線進行品質評鑑時有合理之依據，並排除了可能發生的人為誤差。

關鍵字

道路標線 ； 電腦視覺 ； 表面劣化程度 ； 基準灰階值 ； 邊緣平整度 ； 邊緣偵測

並列摘要

According to Section 02898 road marking of the Standard Construction Specifications and Codes by Public Construction Commission and Maintenance Manual by Directorate General of Highways, the quality inspection of road marking has been relied on subjective visual examination. It is likely to cause different.
This project is to develop standards a systematic road marking quality assessment mechanism, utilizing machine vision technologies in Matlab. The system receives digital images of finished road marking surface and has the images processed and analyzed to capture the geometric characteristics of the marking. The geometric characteristics are then further evaluated to determine the quality level of the finished work.
The proposed road marking mechanism considers twe geometric categories to assess an appropriate score for a finished road marking work: 1) marking surface deterioration degree, and 2) marking edge smoothness.
Marking surface deterioration degree evaluation system as follows:
(1) Formulate a basis of gray values.
(2)Calculate the average deviation grayscale values of marking. Large deviation represents poor setting quality.
(3)Show quantitative results (0-100).
Marking edge smoothness evaluation system as follows:
(1)Use Corner detector to plot the coordinates of marking edge.
(2)Reduce noise with filtering utilities.
(3)Calculate average deviation of the points which are supposed to be aligned. Large deviation represents poor setting quality.
(4)Show quantitative results (0-100).
This study demonstrates the feasibility of the system.Application of the proposed mechanism can effectively reduce improper manipulation of road marking acceptance procedures.

並列關鍵字

road marking ； computer vision ； marking surface deterioration degree ； formulate a basis of gray values ； marking edge smoothness ； edge detection

參考文獻

13. 黃敬群、富博超，「以電腦視覺技術建構之停車場管理服務」，Icl Technical

1. Haas, C.T.,"Evolution of an automated crack sealer: a study in construction technology development", Automation in Construction, Vo4, Issue 4, January 1996, pp.293-305.

2. Kim, Y. S., Yoo, H. S., Lee, J. H., and Han S. W.,"Chronological development history of X–Y table based pavement crack sealers and research findings for practical use in the field". Automation in Construction, V18, Issue 5, August 2009, pp.513-524.

4. Moon, S., Yang, B., Kim, J., and Seo, J."Effectiveness of remote control for a concrete surface grinding machine". Automation in Construction, V19, Issue 6, October 2010, pp.734-741.

5. Victores, J.G. , Martínez, S., Jardón, A. , Balaguer, C."Robot-aided tunnel inspection and maintenance system byvision and proximity sensor integration ". Automation in Construction, In Press, Corrected Proof, Available online 6 January 2011.