Research On Hand Tremor Evaluation Technology Based On Binocular Vision System

Parkinson’s disease is a common dysfunctional disease caused by basal ganglia disease.About 70%of Parkinson’s disease patients have tremor as the first symptom.The current routine diagnosis of tremor in Parkinson’s disease patients is mainly clinical diagnosis.Using the binocular vision system to visualize and quantitatively evaluate tremor can help doctors observe and classify tremor and perform a certain quantitative evaluation within a short visit time.This paper is based on the binocular vision system to evaluate the limb tremor,focusing on the camera calibration and stereo matching of the binocular vision system,analyzes the recognition and tracking of moving targets,and studies the detection of limb tremor by the binocular vision system And analysis technology.This article specifically includes the following work content:First,according to the working principle and calibration method of binocular vision,complete the calibration of the equipment,obtain the camera parameters,analyze the basic steps of binocular matching,analyze the matching process of the SGBM algorithm and the BM algorithm,and conduct experiments on the two algorithms Compared.The image disparity maps processed by the two algorithms are obtained.Through the analysis of the disparity map,the SGBM stereo matching algorithm has a better matching effect,the disparity map is clearer than the BM algorithm disparity map,and the image depth is more accurate.Secondly,analyze the principle of the inter-frame difference method to detect the target,and improve the traditional inter-frame difference method.The moving area image obtained from the three-frame difference is processed by median filtering,and the Canny edge detection algoritlun is used to obtain the edge mask of the current frame.,Use adaptive mathematical morphology to connect the edges of the image to achieve the segmentation of the moving target.Thirdly,the commonly used human hand capture methods are described,and the marking method suitable for the tracking and tremor evaluation research in this article is selected as the capture method.Introduce the principle and characteristics of the Mean-shift algorithm,explain the tracking steps of the Mean-shift algorithm,and design experiments with the hand as the tracking object.In the process of holding the ball,apply the Mean-shift algorithm to the entire hand and the hand to mark the points separately For tracking,through the analysis of the results of the tracking experiment,the Mean-shift algorithm can effectively and accurately track the selected target during the entire process of hand movement.Finally,trough the experimental analysis of the tremor group and the static control group,the tracking of the binocular vision system and the determination of the three-dimensional space coordinates can achieve the acquisition of the space trajectory coordinates of the tremormovement of the hand markers,and use the Fourier transform to quickly analyze the time series data The frequency amplitude ratio,through the tremor spectrogram analysis of the experimental group and the control group,obtained the feasibility and accuracy of the binocular visual system for the analysis and evaluation of tremor motion.