一种基于Kinect的步态分析方法可行性研究

发布时间：2018-01-20 19:39

本文关键词： Kinect 步态分析运动轨迹稳定性对称性　出处：《延边大学》2017年硕士论文　论文类型：学位论文

【摘要】：在全民健身计划纲要举国实施的背景下,针对目前的步态分析系统在实际应用中所暴露出来的成本高、操作繁琐、受外界环境影响大等问题,本文提出一种基于Kinect的步态分析方法,以期解决以上问题为步态分析在全民健身中的普及提供一种新的思路。首先,运用可视化分析法,借助知识图谱可视化分析工具,从一个全新的研究视角对国内外步态分析的发展以及kinect的相关研究成果进行计量与可视化分析,力求科学、清晰的呈现其研究动态前沿及发展趋势,然后根据分析结果运用文献资料法搜集和整理相关文献资料,在有了丰富的理论知识的前提下开始进行基于Kinect的步态分析方法的模型构建,结合步态分析的相关内容和Kinect骨骼跟踪技术原理,最终确定主要从身体姿势的稳定性和两侧肢体的对称性两个维度进行步态分析,从而确定了本方法需要获取的12个关节点的三维运动坐标,依据坐标值和时间,计算出速度、步幅、步宽、步行时相等步态时空参数。利用Unity3D 5.3.4f绘制出两侧肩关节点运动轨迹、两侧髋关节点运动轨迹、两侧膝关节点运动轨迹、两侧踝关节点运动轨迹、重心和头部运动轨迹,利用Kinect for Windows SDK 2.0所提供的功能获取两侧肘关节角度变化,并绘制成曲线,结合欧几里得距离公式计算出两侧肘关节点与髋关节点之间的距离变化,并绘制成曲线,结合方差公式计算出头部和重心运动轨迹的离散程度。最后得出结论,本文在全民健身的大背景下,针对目前步态分析系统在实际应用中所表现出来的各种不足之处,提出一种基于Kinect的步态分析方法,使步态分析过程表现出操作简单、价格低廉、受环境影响低等特点,同时通过研究者设计的三个实验分别证明该方法具有良好的信度、效度和效率,从而验证了该方法的可行性。同时,本文的不足之处是,由于研究者的个人能力和精力有限,导致本方法仅仅实现了从运动学的角度对人体步行过程中身体姿势的稳定性和两侧肢体的对称性两个维度的分析,而不能对受试者的步态进行全面的分析;受制于Kinect的精度不够高,导致本方法所获取的步态信息没有高精度步态分析仪器那么高。
[Abstract]:In the context of the national implementation of the National Fitness Program, the current gait analysis system in the practical application exposed the high cost, cumbersome operation, by the outside environment and other problems. This paper presents a gait analysis method based on Kinect in order to solve the above problems and provide a new way of thinking for the popularization of gait analysis in national fitness. Firstly, visual analysis method is used. With the aid of visual analysis tools of knowledge map, the development of gait analysis at home and abroad and the related research results of kinect are measured and visualized from a new research perspective, in order to be scientific. Clearly presents its research dynamic front and the development trend, then according to the analysis result uses the literature material method to collect and collate the related literature material. On the premise of abundant theoretical knowledge, we begin to build the model of gait analysis method based on Kinect, combining the relevant contents of gait analysis and the principle of Kinect bone tracking technology. Finally, the gait analysis is carried out mainly from the stability of body posture and symmetry of both limbs, and the 3D motion coordinates of 12 nodes need to be obtained in this method. According to the coordinate value and time, the velocity, stride width, stride width and the equal gait space-time parameters were calculated. Using Unity3D 5.3.4f, the motion track of the shoulder points on both sides was plotted. The motion track of bilateral hip joint, knee joint, ankle joint, center of gravity and head. Using the function provided by Kinect for Windows SDK 2.0, the angle change of both sides of elbow joint was obtained, and the curve was drawn. Combined with Euclidean distance formula to calculate the distance between the two elbow points and hip joint points, and draw a curve, combined with the variance formula to calculate the head and center of gravity trajectory discrete degree. Finally, the conclusion is reached. In this paper, under the background of national fitness, a gait analysis method based on Kinect is proposed in view of the shortcomings of gait analysis system in practical application. The gait analysis process shows the characteristics of simple operation, low price, low environmental impact and so on. At the same time, three experiments designed by researchers show that the method has good reliability, validity and efficiency. Therefore, the feasibility of this method is verified. At the same time, the deficiency of this paper is that the individual ability and energy of the researcher is limited. As a result, this method can only analyze the stability of body posture and the symmetry of both limbs in the course of walking from the kinematics point of view, but it can not analyze the gait of the subjects comprehensively. Because the precision of Kinect is not high enough, the gait information obtained by this method is not as high as that of high precision gait analysis instrument.
【学位授予单位】：延边大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】