Design Of Roundness Detection System For Spline Shaft Based On Machine Vision

Spline Shaft is an important transmission component of Automobile Gearbox.Its machining accuracy directly affects the service life of gearbox.Large dimensional error will cause automobile transmission system failure.Therefore,it is necessary to recheck the finished spline shaft to ensure that its end face roundness and concentricity meet the use requirements.The traditional manual detection method is not only slow in detection speed,but also inevitable in reading estimation,false detection and missing detection.The contact detection method requires the inspection tool to directly contact the workpiece,which is easy to cause Scratch on the workpiece.The non-contact detection method of workpiece based on photoelectric principle requires the workpiece to reflect the detection light sent by photoelectric sensor,which is not suitable for workpieces with low light return rate,and has high requirements for the detection environment.In order to meet the current needs of highprecision and high-efficiency detection of roundness and concentricity of gearbox spline shaft end face,a detection method of roundness and concentricity of gearbox spline shaft end face based on machine vision is proposed.Firstly,analyze the system requirements and find out the problems to be solved,that is,detect the roundness and concentricity of the gear side end face of the gearbox spline shaft.The construction principle of detection platform is introduced from the aspects of algorithm,software and hardware.Secondly,the shortcomings of the traditional subpixel edge detection algorithm based on moment method are analyzed.On the basis of single pixel edge detection,subpixel edge location with detection accuracy less than one pixel is carried out.Aiming at the circular object to be detected,puts forward an improved sub-pixel edge detection algorithm which is faster and more suitable for the system,so as to further improve the measurement accuracy of the detection system.Then,the traditional Hough circle detection algorithm can only find the circle,but can not detect the roundness,and can not detect whether multiple circles are concentric is optimized,and an improved Hough roundness detection algorithm more suitable for this system is proposed.While improving the traditional Hoff roundness detection,through the steps of rough positioning of the circle center,screening the points on the suspected circumference,accurate positioning of the circle center,roundness and concentricity detection,the data to be processed can be effectively reduced,which can not only meet the actual needs of roundness and concentricity detection of the system,but also improve the detection speed.Finally,combined with the built transmission spline shaft end face roundness and concentricity detection system,the anti manufacturing and detection accuracy of the improved algorithm are analyzed.The error sources in the experiment are analyzed and the improvement measures are put forward.It further shows that the system can not only meet the experimental needs,but also meet the actual industrial detection requirements.