Quantitative Research On Detection Of Limb Tremor Based On Binocular Stereo Vision

Tremor is the most widespread movement disorder.The quantification and recognition of tremor is helpful to the early diagnosis and differential treatment of nervous system diseases and has important social significance.Although the quantification and identification of tremor has been widely studied,it is still a difficult problem in clinical studies due to its many limitations and deficiencies.Therefore,based on the previous research,this paper proposes a human gesture tremor quantization recognition research based on binocular stereo vision.Early classification and diagnosis of disease can be realized by efficient tremor detection method and effective quantitative index,which can reduce the rate of misdiagnosis and missed diagnosis,and improve the diagnosis and treatment level of pathological tremor.This article obtains from the current limb tremors detection method,analyzed the research status at home and abroad,find out their advantages and disadvantages,around the existing way of measuring current practice will cause patient discomfort,and the gyroscope has inherent errors such as zero offset,the non-contact detection method based on image is designed,which eliminated the sensor binding on tester of discomfort and fear,and there is no inherent error of the sensor compared with the traditional detection methods.The main research contents and results are as follows:(1)This paper analyzes in detail the objective existence of limb tremor and the diseases that suffer form tremor at present,narrates the common detection methods at home and abroad from the classification of tremor,introduces the design scheme of this paper through the characteristics of tremor itself and the shortcomings of current common practices,to selects the use of technology.(2)After selecting the binocular stereo vision technology,the basic principle of binocular stereo vision was studied in depth to explore its feasibility.Founding the fitting point of binocular stereo vision and tremor detection.Moving light spot was used to replace the tremor signal of the limbs to achieve indirect collection,which was both convenient and efficient.(3)Researched the objective errors caused by the traditional camera model and lens distortion.Used the camera calibration method to obtain the camera's internal and external parameters,corrected the errors caused by the above problems via the parameters in software,and calibration tests are carried out to realize the above process.(4)According to the clinical needs,the detection scheme was designed,and then the corresponding quantitative index to evaluate tremor was put forward according to the detection scheme.On the basis of the selection of clinical need to design the related hardware and software,and collected vibration signal,according to the existing stereo matching algorithm is proposed in this paper the use of the stereo matching algorithm based on region,restore the limb tremor trajectory by the 3 d reconstruction algorithm,and then according to the design of the path to solve the quantitative index of the algorithm,finally calculate the quantitative indicators,to reach the goal of tremor detection.