Research On Target Location Of Yarn Hanging System Based On Machine Vision

With the rapid development of machine vision technology,binocular vision,as one of the important fields,has been gradually applied to all fields of our production and life.Binocular vision uses two cameras to form a vision system to observe the world instead of human eyes,observes the target from different angles,extracts the two-dimensional image feature information by using computer image processing technology and speculates the three-dimensional information of the target,so as to obtain the real coordinates of the target world.At present,most textile workshops use manpower to hang yarn,and its automation level is low.Based on a self-designed hanging yarn device and binocular vision technology,this dissertation uses stereo matching to identify and locate spindle in textile shop,obtains the real coordinate of the center circle point of the spindle in the world coordinate system and completes the operation of yarn loading and unloading according to the hardware device.Firstly,comparing the binocular camera calibration methods,selecting the Zhang Zheng-you camera calibration model according to the experimental requirements,using the appropriate black-and-white checkerboard and images collected by binocular camera to demarcate binocular camera and obtaining the respective optimal internal parameters of the two cameras,then calculating their external parameters according to the position and pose parameter information of the spatial 3D projection of the two cameras.Finally,leading into stereo correction technology to construct parallel binocular X-ray.Then,preprocessing the images collected by binocular camera.Completing the image grayscale,enhancing the image by histogram normalization according to the image grayscale histogram,and denoising the image by median filter.After the image preprocessing is completed,analyzing and comparing the current stereo matching technology and selecting the SGBM algorithm to calculate the parallax map of the preprocessed image.In addition,the Canny operator is used to replace the Sobel operator in the original algorithm for edge detection,so as to enhance the robustness of the algorithm and maximize the weak edge information of the target and better complete the edge information fitting.The median filter is used to fill the holes in the parallax map,so that the parallax map tends to be smooth and complete,improves the density,and obtains more significant parallax results.Finally,marking the center after edge detection and circular fitting,inferring the real-world coordinates of each center according to the principle of triangulation,and then using the distance between the centers of circles to complete the error measurement and analysis identification.Finally,completing the identification and positioning of the barrel spindle rod.The automatic yarn hanging mechanical device is designed and the hardware module selection is completed according to the specific requirements,then,using the calibration results of binocular vision system to complete the yarn loading operation and measuring the error and positioning time in the direction of three-dimensional coordinate axis.