Design Of On-line Detection Device For Bobbin Yarn Density Based On Machine Vision

The bobbin yarn is easy to store and transport,and can be produced in small batch.Compared with the hank yarn,it is more widely used in the textile field.The winding quality of the bobbin yarn affects the subsequent warping and dyeing process.One of the basic indicators to evaluate the winding quality of the bobbin yarn is the average winding density of the bobbin yarn.At present,the contact measurement methods such as manual measurement are basically used in many domestic textile companies for the detection of the average winding density of the bobbin yarn.In order to improve the detection efficiency,the sampling testing method need be used for the same batch of bobbin yarn.The automation level of manual detection is low,and the measurement results depend on the subjectivity of the operator,which is easy to be detected by mistake.Moreover,the contact measurement may cause the bobbin yarn deformation and affect the subsequent production process.Machine vision measurement is a convenient and efficient non-contact measurement method,which is widely used in the field of industrial detection in recent years.Its advantages such as high automation and high reliability provide a new idea for the detection of bobbin density.Aiming at the problems of low efficiency and low automation of the traditional density measurement method of bobbin winding,this thesis designed an on-line detection device of bobbin yarn density based on machine vision for the endless bobbin winding which is widely used in the textile industry.The main research contents are as follows:(1)The overall scheme of the on-line density detection device of bobbin yarn based on machine vision was designed.The limitations of the current detection methods of the average winding density of the bobbin yarn at home and abroad were analyzed.The operation scheme and hardware structure of the whole device were designed according to the actual needs of the domestic textile enterprises and the characteristics of the bobbin yarn,and carries out the hardware selection.The selection of hardware devices was completed and the sensor signal processing circuits and communication system were designed.Finally the controllable hardware system was built,which implemented the automatic control of bobbin density online detection device based on machine vision.(2)The calibration method of the device and the preprocessing algorithm for the image of bobbin yarn were designed.The calibration contents of this device include the parameters calibration of cylinder core and bobbin template,the parameters calibration of image acquisition subsystem,the parameters calibration of weight measurement subsystem as well as the pixel equivalent,which provided the reliable basis for the subsequent image acquisition,weight measurement,image processing and bobbin yarn density calculation.According to the characteristics of the actual bobbin winding shape and the image acquired by the image acquisition subsystem,the image preprocessing algorithms were designed,including the ROI extraction,the edge preserving filter and the sub-pixel precision edge extraction.Through the calibrated device and the image preprocessing algorithm,the bobbin yarn winding profile image with complete profile information and less interference information was obtained.(3)The algorithm of bobbin winding profile correction and the calculation method of bobbin yarn winding density were studied.The imaging principle,as well as the cause and influencing factors of image distortion in the image acquisition subsystem,was analyzed.The advantages and disadvantages of image splicing method and model correction method to calibrate the bobbin yarn winding profile were compared.The volume calibration model of bobbin yarn based on perspective projection theory was designed.The model parameters were obtained based on the the camera space geometry information.The preprocessed bobbin image was then calibrated inversely based on this model to obtain the front projection image without distortion.The adaptability of cylindrical shell method and the cylindrical shell method to the volume calculation of bobbin was compared and demonstrated respectively.The cylindrical shell method was then chosen to calculate the volume of bobbin yarn based on the bobbin yarn front projection image and finally the average winding density of the bobbin yarn was calculated according to its weight.(4)The PC computer software of the on-line detection device of bobbin yarn density based on machine vision was designed,and the experimental analysis of the device was completed.The requirements of the on-line detection device for bobbin yarn density were analyzed.Then the image processing program and the model calibration program based on OpenCV were written by C++and was packaged into a DLL file.The PC computer software interface was developed by using C#which called the DLL files for image processing and density calculation so as to improve the operation efficiency of the program.The stability coefficient of the device is 0.003?0.039,which was obtained in laboratory.The accuracy of the device was tested in the production workshop of a textile company for 150 endless bobbin yarn.where the relative error of the device was within ± 3%.Experimental results show that the reliability,accuracy and stability of the device can meet the actual production requirements,and the improvement direction is clear.