Research On Detection Method Of Tractor-trailer Train Braking Time Sequence Based On Monocular Vision

With the development of our country’s transportation industry,more and more automobiles are used in logistics engineering and intercity transportation.Tractor-trailer trains occupy a major position in today’s logistics engineering because of their higher fuel economy and ease of modular transportation.National standards stipulate that semi-trailer trains and central axle trailer train are allowed to run on highways,which further promotes the development of my country’s tractor-trailer trains.Compared with ordinary passenger cars and commercial vehicles,the dynamic model of tractor-trailer trains is more complicated,the braking instability,such as tail flick,head push,drag,is prone to occur in the braking process due to asynchronous braking between the axles of the tractor and the trailer.Braking time sequence detection is to detect the inter-axle braking time sequence of the tractor-trailer trains,and analyze the braking state of the tractor-trailer trains to determine whether the braking system needs to be adjusted or even repaired.For the detection of braking time sequence,the national standards also put forward clear requirements,so it is necessary to propose a truck-trailer braking time sequence detection method based on monocular vision.First of all,analyzing the relationship between slip rate and braking force with the help of a tire model,and uses the time when the slip rate reaches 20% to characterize the braking time sequence between the axles of tractor-trailer train.By analyzing the slip rate formula,it is known that solving the tire slip rate requires obtaining the translational speed and rotation speed of the tractor-trailer trains.The translational speed and rotation speed of the tire are obtained by visual inspection the pasting markers on the edge of the tire,then the tire slip rate is solved by analyzing the cycloid trajectory formed by the tire edge points during the tire rolling process.At the same time,two 180-degree markers are used to solve the tire rotation angle to improve the detection accuracy.According to the proposed detection method,the spatial arrangement scheme,system hardware scheme and detection process scheme of the monocular vision tractor-trailer train braking sequence detection system are designed.In terms of image processing,considering that there are too many markers in the same field of view,template matching is used to identify tire markers to improve detection accuracy,concentric circles in black and white are used as markers in order to improve the accuracy of template matching.In the aspect of image preprocessing,median filtering and shock filtering are used to denoise and sharpen the image to improve the contrast of the markers in the image.In order to obtain the area where the template is created,Blob analysis is used to obtain the marker area,the template is created and the contour-based template matching algorithm is used to identify the marker area in the image to be matched.In order to verify the matching effect,the sub-pixel contour of the template image is converted to the matching area with the help of the positioning information output by the template matching operator,the sub-pixel contour of the template basically coincides with the boundary of the marker in the matching area,indicating that the matching accuracy is reliable.At the same time,in order to verify the matching effect,the average gray level difference of the image in the different sub-pixel contours of the matching area is extracted and compared with the set gray level difference threshold to determine the accuracy of the match.In order to improve the accuracy of image acquisition,the Zhang’s calibration method is used to calibrate the camera,and the camera calibration test is carried out.The camera calibration effect is explained from three aspects: the reprojection error,the stability of the parameter matrix in multiple experiments,and the linearity of the same corner point of the checkerboard after distortion correction,and the camera space positioning scheme is proposed.Finally,a verification test was carried out to verify the proposed slip rate solution formula.The test results showed that the angle recognition error was below 3.3%,and the slip rate solution error was below 4.2%.In order to verify the effectiveness of tractor-trailer trains braking time sequence detection method,a real vehicle test was carried out,three tractor-trailer trains were selected for five sets of tests to verify the detection effect,and the test results were analyzed.