π-bot: 自主式鋪面破損即時檢測機器人

鋪面檢測是對於長期的鋪面維護以及修繕事項的重要工作項目。這種長期的鋪面維護以及修繕事項通常是耗時且耗人力的。於此本研究提出了一種自主式的鋪面檢測方法，也就是 PI-bot。在本論文研究中提到了兩大貢獻中，第一個為開發一自主式之智能機器人，主要是檢測儀器之搭載及運動之用途。此機器人搭載了全球定位系統、慣性測量單元、雷射測距儀以及一個鋪面檢測的模組。而其系統軟體是在微軟之MSRDS底下做開發，於此則可以提供PI-bot開發上相當高之彈性以及擴充性，並且於其系統內部採用IPC機制處理資料的同步。除此之外，PI-bot還實作了RTK定位、避障系統和重複檢測的行為。為了評估此PI-bot之自主式機器人之可行性以及效果，本研究已經在硬體溝通、軟體架構以及行為的實作方面全方面進行測試。而其測試之結果也顯示此PI-bot系統能夠在短距、小範圍的鋪面檢測上正常運作。而PI-bot的第二個部分則是我們所研發適用於機器人檢測的彩色雙光源檢測方法(CDLI)，此破損辨識方法適合於自動式的檢測。此CDLI方法主要包含四個步驟，也就是影像擷取、影像相減、影像強化以及影像分類。為了能夠驗證這個方法，我們擷取504張的影像，並且利用這些影像來測試CDLI方法的成效。而實驗的結果顯示，此彩色雙光源檢測方法(CDLI)在一般正常鋪面、鱷魚狀裂縫和人手孔方面有相當好的成果，而在污漬鋪面以及縱向橫向裂縫的鋪面上，也可達到令人滿意的效果。而在將彩色雙光源檢測方法與PI-bot進行整合後，本研究也證明此即時自主式鋪面檢測系統可有效地執行鋪面檢測工作。

關鍵字

鋪面檢測系統；機器人學；自主式檢測；鋪面破損；影像處理；彩色雙光源檢測

並列摘要

Pavement inspection is an important part of long-term maintenance and rehabilitation (M&R) work on roadways. M&R activities are often time-consuming and labor-intensive. PI-bot, an autonomous inspection method for pavements is proposed. Two major contributions of this research are described. The first is the development of an autonomous, intelligent robot which carries inspection equipment in the field. The robot is equipped with a GPS receiver, an inertial measurement unit (IMU), a laser rangefinder, and an inspection module. PI-bot’s system software was developed using Microsoft Robotics Developer Studio (MSRDS), which provides a high degree of development flexibility and extensibility, along with an Inter-Process Communication (IPC) mechanism for data synchronization. PI-bot is also equipped with a Real Time Kinematic (RTK) positioning system, an obstacle avoidance system, and re-inspection behaviors. In order to verify feasibility and performance of the proposed system, the hardware communication, software architecture, and behavior implementations were comprehensively tested in the field. The test results show that PI-bot is effectively at performing a short range and small area pavement inspection task. The second contribution is the Chromatic Dual-Light Inspection (CDLI) method, a distress identification method suitable for automatic inspection. The CDLI method is comprised of four steps, namely image acquisition, image subtraction, image enhancement, and image classification. To validate the CDLI method, we recorded 504 images and evaluated its performance with them. The results indicate that the proposed CDLI method performs very well on normal pavement, alligator cracks, and manholes, and performs satisfactorily on spillage pavement, longitudinal cracks, and transverse cracks. With the CDLI method integrated into PI-bot, the proposed real-time autonomous pavement inspection system is demonstrated to work effectively to execute pavement inspection tasks.

並列關鍵字

Pavement inspection system ； robotics ； autonomous inspection ； pavement distresses ； image processing ； chromatic dual-light inspection

參考文獻

[1] Madanat, S. (1993) "Incorporating inspection decisions in pavement management," Transportation Research Part B: Methodological 27 (6), 425-438, Elsevier.

[2] Guralinick, S.A., Suen, E.S., Smith, C. (1993) "Automating inspection of highway pavement surfaces," Journal of Transportation Engineering 119 (1), 1-12, ASCE.

[4] Bursanescu, L., Blais, F. (1997) "Automated Pavement Distress Data Collection and Analysis: a 3-D approach." Proceedings of International Conference on Recent Advances in 3-D Digital Imaging and Modeling, Ottawa, Ontario, Canada.

[8] Tseng, Y.H., Kang, S. C., Chang, J. R., Lee, C. H. (2011) "Strategies for Autonomous Robots to Inspect Pavement Distresses," Automation in Construction 20 (8), 1156-1172.

[9] Chang, J.R., Tseng, Y. H., Kang, S. C., Tseng, C. H., Wu, P. H. (2007). "The Study in Using an Autonomous Robot For Pavement Inspection," Proceedings of International Symposium on Automation and Robotics in Construction, Kochi, Kerala, India. September 19-21

國際替代計量

π-bot: 自主式鋪面破損即時檢測機器人

全文下載

主題瀏覽