| In the process of long-term operation management and line maintenance,it is necessary to check the construction gauge and equipment gauge of the rail transit tunnel for the whole line.For the traditional contact measurement method,not only detection efficiency is very low and data is not easy to save,but also it needs more manpower and material resources and there are scheduling difficulties in the large vehicle-carrying measurement system and other disadvantages.Therefore,on the basis of sufficient site investigation and previous theoretical research,this paper proposed to adopt machine vision detection method and carry portable motion detection platform so as to realize the non-contact motion detection of the rail transit tunnel gauge for the whole line.Here follows the main work:(1)Aiming at the large-scale target photogrammetry in near view,firstly,the related technologies based on machine vision measurement were studied,including the establishment of camera imaging model and the research of camera calibration method.Then summarizes its measuring features,studies the triangulation principle suitable for large-scale target contour measurement in near view and establishes the theoretical foundation for the design of portable tunnel gauge detector system and key technologies.(2)This paper designed and analyzed the overall scheme of the portable tunnel-gauge detector and the implemented function and detection principle.It adopted the triangulation measurement principle,utilized the Pro/E 3D simulation design software to design a portable motion platform and realized the software design and system integration of the portable tunnel-gauge detector under the DALSA camera development environment and the VC++6.0MFC development platform through the model selection of each component and system hardware design.Finally,the overall design of the portable tunnel-gauge detector was completed,which could meet the multi-stage multi-periodtunnel gauge detection requirements in the whole life cycle of the rail transit.(3)Aiming at the key technical problems in the process of system development,firstly,this paper used the image fuzzy processing technology and proposed an efficient processing method for tunnel section image.It adopted Sobel algorithm and boundary midpoint method to seek and obtain the fine solid curve of the tunnel laser section contour,and filtered the light noise according to the regular changes of the lamplight in the tunnel so as to solve the image recognition interference problem.Secondly,this paper studied and designed a large multi-array round calibration plate to do the system calibration work,proposed the region growing gravity method to realized extraction of the datum point on the calibration image,and utilized the datum point matching method to realize the calibration of the detection system.Finally,for the standard gauge matching problem,through the angle and the decision method to realize the decision of the beyond limit point,this paper detected the beyond limit point of manual setup and realize the real-time beyond limit detection of the tunnel gauge through designing the algorithm model in VC++.(4)This paper conducted the detection error analysis on the tunnel-gauge detector from the two perspective of indoor and field,through the disassembly and assembly repeatability test,compared and analyzed the manual measurement data,and the error values are within a reasonable range.For the gauge measurement,the error values are within ±1mm and for the tunnel section distance measurement,the error values are assured within ±3mm,which can completely meet the detection requirements of the rail transit tunnel gauge. |
PDF Full Text Request