﻿ 基于拉班空间的跌倒方位检测方法
 计算机系统应用  2018, Vol. 27 Issue (11): 224-230 PDF

1. 西安理工大学 计算机科学与工程学院, 西安 710048;
2. 陕西省网络计算与安全重点实验室, 西安 710048

Falling Position Detection Method Based on Laban Space
TU Peng-Jia1, LI Jun-Huai1,2, WANG Huai-Jun1,2, JI Wen-Chao1,2, WANG Kan1,2
1. School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710048, China;
2. Shaanxi Key Laboratory for Network Computing and Security Technology, Xi’an 710048, China
Foundation item: National Natural Science Foundation of China (61771387); National Key Research and Development Plan of China (2017YFB1402103); Searle Nnetwork Innovation Project (NGII20150707, NGII20160704)
Abstract: In moving object detection process, it needs to automatically judge whether it has detected the moving object, although there is no moving object in the current scene, detection result wrongly judges that it has detected the moving object. In order to find the source of the error, optical flow disturbance effect is found through experiment. The optical flow disturbance effect detection algorithm is designed, and the effect of optical flow perturbation is clearly detected. Next, through the binarization method of image it eliminates optical flow disturbance effect. The ideal results of the moving object detection are obtained. This research proves that the optical flow perturbation effect exists in the space, which can cause interference to the detection of moving object. It also can eliminate the effect of optical flow disturbance and improve the accuracy and reliability of moving object detection and judgment.
Key words: falling detection     moving average filtering     labanotation     position judgment

1 跌倒检测方法

1.1 拉班空间

1.2 跌倒方位检测算法

 图 1 拉班空间

1.2.1 姿态角计算

 图 2 坐标转换图

 $R = {R_{\rm Z}}\left( r \right) = \left[ {\begin{array}{*{20}{c}} {\cos r}&{\sin r}&0 \\ { - \sin r}&{\cos r}&0 \\ 0&0&1 \end{array}} \right]$ (1)
 $P = {R_{\rm X}}\left( p \right) = \left[ {\begin{array}{*{20}{c}} 1&0&0 \\ 0&{\cos p}&{\sin p} \\ 0&{ - \sin p}&{\cos p} \end{array}} \right]$ (2)
 $Y = {R_{\rm Y}}\left( y \right) = \left[ {\begin{array}{*{20}{c}} {\cos y}&0&{ - \sin y} \\ 0&1&0 \\ {\sin y}&0&{\cos y} \end{array}} \right]$ (3)

 $\begin{array}{l}N = {(RPY)^{ - 1}} = \left[ {\begin{array}{*{20}{c}}{\cos r\cos y - \sin r\sin p\sin y}\\{\sin r\cos y + \cos r\sin p\sin y}\\{ - \cos p\sin y}\end{array}} \right.\\\begin{array}{*{20}{c}}{ - \sin r\cos p}\\{\cos r\cos p}\\{\sin p}\end{array}\;\;\;\left. {\begin{array}{*{20}{c}}{\cos r\sin y + \sin r\sin p\cos y}\\{\sin r\sin y - \cos r\sin p\cos y}\\{\cos p\cos y}\end{array}} \right]\end{array}$ (4)

 $N = \left[ {\begin{array}{*{20}{c}} {{T_{11}}}&{{T_{12}}}&{{T_{13}}} \\ {{T_{21}}}&{{T_{22}}}&{{T_{23}}} \\ {{T_{31}}}&{{T_{32}}}&{{T_{33}}} \end{array}} \right]$ (5)

 $pitch = p = \arcsin ({T_{32}})$ (6)

${T_{12}} = - \sin r\cos p$ , ${T_{22}} = \cos r\cos p$ , 可得人体姿态角中的偏航角:

 $roll = r = \arctan \left( { - \frac{{{T_{12}}}}{{{T_{22}}}}} \right)$ (7)

${T_{13}} = - \cos p\sin y$ , ${T_{33}} = \cos p\cos y$ , 可得人体姿态角中的俯仰角:

 $yaw = y = \arctan \left( - \frac{{{T_{31}}}}{{{T_{33}}}}\right)$ (8)
1.2.2 动作差异分析

 图 3 不同行为动作角度分布

 图 4 跌倒曲线图

 图 5 不同跌倒方位roll角变化图

1.2.3 跌倒检测及方位判定算法

(1) 计算获取三个姿态角, 判断pitchroll是否超过阈值.

(2) 当连续1 s时间内pitchroll超过阈值, 认为人体姿态出现疑似跌倒状态, 但还不能判定测试者已跌倒, 先预报警.

(3) 延时采集2 s时间数据, 对姿态角进行二次检测. 该步骤的必要性在于判断测试者是否处于异常姿态的终态, 即倒地. 当测试者一直倒地时, 人体相对地面静止, 姿态角将一直处于异常状态, 这时就会判定测试者已跌倒, 进行报警.

(4) 记录跌倒姿态角值, 判定具体跌倒方位, 最后依据跌倒方位预诊可能受伤部位, 达到及时精准救治目的.

 图 6 算法流程图

2 实验验证与分析 2.1 实验设计

2.2 实验结果

 图 7 不同滤波处理前后对比图

 图 8 不同跌倒方位曲线图

3 结束语

 [1] Wang J, Zhang ZQ, Li B, et al. An enhanced fall detection system for elderly person monitoring using consumer home networks. IEEE Transactions on Consumer Electronics, 2014, 60(1): 23-29. DOI:10.1109/TCE.2014.6780921 [2] 沈秉乾, 武志勇, 贺前华, 等. 人体姿势状态判决的跌倒检测方法. 计算机应用, 2014, 34(S1): 223-227, 264. [3] Qu WH, Lin F, Xu WY. A real-time low-complexity fall detection system on the smartphone. Proceedings of IEEE 1st International Conference on Connected Health: Applications, Systems and Engineering Technologies. Washington, DC, USA. 2016. 354–356. [4] Mazurek P, Morawski RZ. Application of naïve Bayes classifier in fall detection systems based on infrared depth sensors. Proceedings of the 8th International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications. Warsaw, Poland. 2015. [5] 白勇, 孙晓雯, 秦昉, 等. 基于SVD特征降维和支持向量机的跌倒检测算法. 计算机应用与软件, 2017, 34(1): 247-251. DOI:10.3969/j.issn.1000-386x.2017.01.045 [6] 王荣, 章韵, 陈建新. 基于三轴加速度传感器的人体跌倒检测系统设计与实现. 计算机应用, 2012, 32(5): 1450-1452, 1456. [7] He J, Bai S, Wang XY. An unobtrusive fall detection and alerting system based on Kalman filter and Bayes network classifier. Sensors, 2017, 17(6): 1393. DOI:10.3390/s17061393 [8] De La Concepción MÁÁ, Morillo LMS, García JAÁ, et al. Mobile activity recognition and fall detection system for elderly people using Ameva algorithm. Pervasive and Mobile Computing, 2017, 34: 3-13. DOI:10.1016/j.pmcj.2016.05.002 [9] Xu T, Zhou Y. Fall detection based on skeleton data. Proceedings of the 3rd International Conference on Human Aspects of IT for the Aged Population. Applications, Services and Contexts. Vancouver, Canada. 2017. 475–483. [10] Young J, Wood LL. Laban: A guide figure between dance/movement therapy and drama therapy. The Arts in Psychotherapy, 2018, 57: 11-19. DOI:10.1016/j.aip.2017.09.002 [11] Kangas R, Pöllänen E. Physical activity responsive miRNAs-potential mediators of training responses in human skeletal muscle?. Journal of Sport and Health Science, 2013, 2(2): 101-103. DOI:10.1016/j.jshs.2013.04.002 [12] Özdemir AT. An analysis on sensor locations of the human body for wearable fall detection devices: Principles and practice. Sensors, 2016, 16(8): 1161. DOI:10.3390/s16081161 [13] 程兰, 王志远, 陈杰, 等. 基于粒子滤波和滑动平均扩展Kalman滤波的多径估计算法. 电子与信息学报, 2017, 39(3): 709-716.