Research On Detection System For Trucks' Side Safety Guard Based On Machine Vision

With the particular mixed traffic road circumstances of the vehicles and pedestrians,installation of the trucks' side safety guard is an effective measure to prevent non-motorized vehicles and pedestrians involve in the goods vehicle when the side collision happening.At present,the side safety guard detection is carried out in a specified place and specified time by a specific person,which leads to a large human factors and inefficient.This method could neither prevent the removal of the side safety guard privately after the truck annual inspection nor avoid the behavior that rental the side safety guard to conduct the annual inspection.It has become particularly important to measure the side safety guard size fast,effectively and comprehensively.Therefore,a truck side safety guard installation size detection method was raised based on sheet of light measurement principle,which avoids the measurement error due to the side safety guard and front and rear tires are not coplanar in the original two-dimensional visual detection scheme.On the basis of theoretical research,an experimental platform was designed and built,the experimental result shows that the measurement error is less than 1 mm,which meets the system design requirements.The main contents are as follows:Firstly,based on the research of truck side guards and machine vision status,visual inspection will be applied to truck side safety guard installation position and size detection.According to the national standard for truck side safety guard requirements,the size detection scheme was raised for t side safety guard based on the sheet of light measurement principle,and corresponding experimental platform was built.Secondly,based on the fundamental principle of sheet of light and system design,a simplified sheet of light measurement model was proposed,which could greatly reduce the amount of computation.After,researched on the camera perspective imaging model and the common camera calibration principle,and Zhang Zhengyou camera calibration method is applied to camera calibrate.Then,studied and compared the common extraction method of piece light strip center and the gravity gradient method was mainly studied and applied to extract the center of ling strip.Then,a gray mapping dimensionality reduction method was proposed to map the points in the three-dimensional coordinate into the two-dimensional image.Finally,the two-dimensional image processing methods,including image preprocessing,Blob analysis,sub-pixel edge detection with Canny operator,oval and line fitting algorithm and so on was applied to calculate the distance to the front and rear tires of the side safety guard respectively.Again,the system detection software has been written based on the MFC class library.The corresponding class for camera and PCI-1761 was designed respectively to realize their functions,the joint programming for OpenCV libraries and Halcon with VC was achieved,as well as the image processing algorithms.Finally,the calibration experiments were conducted,including camera calibration experiments,the light plane pose calibration experiment and the motion direction calibration experiment.And the experimental results were compared and analyzed.