Research On Coaxiality Detection Method Of Arm Hole Group Of Loading Maneuver Based On Machine Vision

The boom is the main connecting component of the loader’s working device,which lifts the bucket and connects it to the frame.Due to the complex structure of the loader’s working device and its high requirements for motion characteristics,the coaxiality of the boom hole group directly affects the assembly and transmission efficiency of the entire machine.It can cause uneven force on the assembly and may lead to control failure,which seriously threatens the service life and personnel safety of the vehicle.Therefore,accurately obtaining the coaxiality error of the loading manipulator arm and timely feeding back the data to the manufacturing process can ensure the quality of the whole machine and improve the manufacturing level of construction machinery parts.In this paper,the current detection scheme of the enterprise is investigated,aiming at the problems such as poor efficiency,low automation level and poor information timeliness,the online rapid detection method of the coaxiality error of the arm hole group of loading motor based on machine vision is carried out.The main work of this paper is as follows:1.A coaxiality error model was established based on the structure and forming characteristics of the boom hole group,and a coaxiality detection scheme based on machine vision was proposed.The industrial camera resolution,industrial lens,light source,etc.were determined according to the hole group size,detection accuracy,and light irradiation conditions.An image acquisition device composed of a linear module,a laser rangefinder,an image acquisition module,and other components was built.2.Basic image processing algorithms such as image enhancement,image filtering,image denoising,and threshold segmentation were described.An improved Retinex image enhancement algorithm was proposed for enhancing the contrast and quality of images in poor image acquisition environments.To address the shortcomings of traditional filtering algorithms,a guided filtering algorithm based on joint wavelet transform was used to filter images,and comparison experiments were conducted with traditional filtering algorithms in MATLAB.The sharpening algorithm was improved by using the Laplace algorithm to sharpen the filtered image and enhance its edge contours and details,which is beneficial for the accuracy of subsequent edge detection.3.The camera’s internal and external parameters.A coaxiality vision detection system software was built using the HALCON operator library and Lab VIEW development platform,including image acquisition and display,hole group data processing,motion control,and output and storage of detection results.The obtained spatial coordinates of the circle center were processed using MATLAB,and the minimum least squares algorithm based on singular value decomposition was used for baseline axis fitting.The coaxiality error can be calculated according to the error definition.The results of the algorithm’s measurement of coaxiality error were compared with those of three-coordinate measurement,and the stability of the measurement system was analyzed.The results showed that the precision error of the visual detection results was within the acceptable range,and the maximum fluctuation of the measurement values for the same workpiece was 0.04 mm.Therefore,it can be considered that the visual detection method can meet the measurement requirements in terms of accuracy and stability.At the same time,the detection time for the entire machine was reduced from 50 minutes to 15 minutes,greatly improving detection efficiency.