| Tractor is the typical agricultural machinery and the power machinery of agricultural production,and its quality promote the development of agriculture.The drive train is a key component of the tractor.The small end face of the gearbox and the end face of the gearbox combine to form the drive train.The quality of the joint surface affects the overall performance of the tractor.The pin-hole is the positioning foundation of the joint surface,which effectively measures the hole diameter and position of the pin-hole,which is the key to ensure the quality of the drive train.At present,in the gearbox manufacturing enterprises,the measurement of the aperture is made by plug gauges and calipers.The measurement of the position of the hole group is mainly performed by the three-coordinate measuring instrument,which cannot meet the online real-time measurement requirements.The existing machine vision measurement technology limits the field of view of a single-camera high-precision measurement system because the image resolution and the field of view cannot be simultaneously satisfied.Aiming at the above problems,this paper developed a dual camera-based vision measurement system for the pin hole machining quality of tractor gearbox,realizing the online real-time measurement of aperture and position degree;analyzing the influencing factors affecting the processing quality,proposing early warning rules and establishing the processing quality The early warning model provides decision support for quality management during the manufacture of tractor gearboxes.The main research contents and conclusions of this paper are as follows:1.Three measuring schemes are designed and the measurement device is constructed to select the best scheme and developed software system.The requirement of pinhole machining accuracy for the end face of transmission box is analyzed.Three measurement schemes based on checkerboard standard board,cross laser and high precision fine-tuning device are designed.Hardware selection of camera and lens is completed,and three machine vision measurement devices are constructed.Because of the depth of field problem,the measuring device based on chessboard can not meet the requirements of processing accuracy.Because of the requirement of accumulating errors and high precision measurement of pin-holes,the measurement results based on cross laser can not meet all the accuracy requirements.The measurement device based on high precision fine-tuning device is selected,and the software system of machine vision measurement is developed.The field environment verifies that the measurement system meets the requirement of on-line measurement of pin aperture and position of transmission box end face.2.The image processing algorithm is studied and an improved Zernike moment sub-pixel edge detection algorithm is proposed.In order to achieve high precision measurement of pin-hole,the filtering and denoising algorithms are compared and analyzed.Bilateral filtering algorithm is selected to highlight image edges.The image enhancement algorithms are compared and analyzed,and Gamma transform is selected to enhance the image contrast.According to the gray histogram of the image,threshold is set to convert the image into binary image.The edge detection algorithm and effect are compared and analyzed.Canny edge detection is selected to extract the pixel-level edge of pin-hole.In order to extract edges more accurately,an improved Zernike moment sub-pixel edge detection algorithm is proposed to detect pin-hole sub-pixel edges.The experimental verifies that the average deviation of aperture measurement in the measurement system is 0.0026 mm and 0.0032 mm.The mean deviation of the measurement of position degree is 0.0090 mm and 0.0133 mm,which meets the requirements of pin-hole machining accuracy.3.The influencing factors of pin-hole processing quality are analyzed and the early warning model of pin-hole positioning quality is established.The factors affecting the processing quality of pin-hole on the end face of tractor gearbox are analyzed.The main factors affecting the processing quality are determined – the environmental temperature and the number of drill tools.By analyzing the relationship between the influencing factors and the location degree,the quality early warning rules are put forward.LSSVM and PSO are studied and analyzed,and a pin-hole position quality early warning model based on LSSVM is established.PSO is used to optimize the unknown parameters of LSSVM.The experimental results show that the RMSE of the model is 0.0058 mm and the RE is 0.075%,which verifies the validity of the established processing quality early warning model. |
PDF Full Text Request