Research And Development Of Vision-based Motion Measurement System For Soft Executor

In this paper,a machine vision measurement system is proposed to extract and measure the motion parameters of a soft-bodied robot unit.The recognition of moving objects,the calibration of binocular cameras and the reconstruction of three-dimensional coordinates of the system are introduced.In view of the different motion modes of the soft-bodied unit,this study focuses on extracting the motion parameters by vector calculation.Firstly,A general scheme for measuring soft-bodied unit motion with machine vision measurement system is proposed..The system is introduced in four aspects: equipment selection,platform construction,image acquisition and software processing.Secondly,the platform of the whole system is built to process the collected source image and extract soft-bodied unit object.The theory and common algorithms of image segmentation are analyzed and summarized.This paper introduces a hybrid algorithm based on genetic algorithm and OTSU threshold method,which is used to segment the foreground and background of the image.And,the moving object of the soft-bodied unit is extracted by region growing method.Then,the machine vision measurement system is calibrated and the camera's internal and external parameter matrix are obtained,including the introduction of camera model and polar constraint,single camera calibration,binocular calibration and stereo matching.This study also uses different calibration boards to test the calibration accuracy of the system.Finally,the soft executor motion mode is analyzed.The methods of characterizing the angle of bend motion and the elongation of stretch motion are specified.The method of threedimensional coordinate vector and two-dimensional image pixel coordinate are used to calculate the bending angle,and the accuracy of angle extraction is verified by comparing with the bending angle extracted by the vertical camera.The method of calculating the elongation with three-dimensional coordinates is used,and the accuracy of the elongation extraction is verified by comparing with the marker of known length.And,the whole system is debugged and the experiment of parameter extraction precision is carried out to verify the feasibility and validity of the system.