Design Of Automatica Gripping Device For Medical Latex Gloves

Medical latex glove,as a commonly used protective tool in current medical treatment,is increasingly in demand and has more stringent requirement in quality.The existing quality inspection process is to manually place the gloves in the turnover box on the inspection platform for testing.With the increase of production demand and the increase of labor cost,it will be of practical significance to design an automatic grabbing device to realize the automatic feeding in the inspection link.Since the latex glove is soft and easy to deform,it is difficult to complete the automatic feeding by using a fixed model,and there is no good solution at present.In this thesis,a fast target detection method based on minimum circumscribed rectangle was proposed.Meanwhile,an adaptive weighted filtering algorithm based on gray-scale difference ratio was also given.The machine vision technology is used to locate the specific position of the latex gloves,and the manipulator controlled by PLC grasps them to complete the automatic inspection feeding.The main contents are as follows:Firstly,the control system of the gripping device was designed according to the step of adjusting the posture of the glove.Since the position of each grab is uncertain,the thesis used machine vision technology to determine the specific position of the latex glove,and the coordinate data was given to the control system to achieve the grab.Secondly,in order to avoid the influence of the target object other than the latex glove on the visual positioning system of the grasping device,the latex glove needs to be taken out from the collected image.Due to the contrasting background and foreground of the image studied in the thesis,an adaptive weighted average filtering algorithm based on gray-scale ratio was proposed to solve the effect of outlier noise on the image.At the same time,a fast target detection algorithm based on minimum circumscribed rectangle was given.By calculating the minimum circumscribed rectangle of the extracted contour as the target candidate frame,and then ResNet deep convolution network was used to identify the objects in the selected candidate frame.And finally,the determined area where the latex glove was located.Thirdly,because of the instability of industrial environmental factors,it will have an impact on the quality of the picture.Aiming at this problem,an adaptive weighted average filtering algorithm based ongray ratio was proposed to filter the image.And then,the algorithm of arc detection + feature point detection was used to detect the finger area of the latex glove for determining the final grab point.Finally,the visual positioning algorithm of the design was realized.The operation interface of the automatic grabbing device was written,and the whole machine debugging was completed.The experimental test results show that the recognition rate and recognition speed of the finger algorithm designed by the thesis can basically reach the expected goal.