An Approach for Human Posture Recognition Based on the Fusion PSE-CNN-BiGRU Model

This study proposes a pose estimation-convolutional neural network-bidirectional gated recurrent unit(PSECNN-BiGRU)fusion model for human posture recognition to address low accuracy issues in abnormal posture recognition due to the loss of some feature information and the deterioration of comprehensive performance in model detection in complex home environments.Firstly,the deep convolutional network is integrated with the Mediapipe framework to extract high-precision,multi-dimensional information from the key points of the human skeleton,thereby obtaining a human posture feature set.Thereafter,a double-layer BiGRU algorithm is utilized to extract multi-layer,bidirectional temporal features from the human posture feature set,and a CNN network with an exponential linear unit(ELU)activation function is adopted to perform deep convolution of the feature map to extract the spatial feature of the human posture.Furthermore,a squeeze and excitation networks(SENet)module is introduced to adaptively learn the importance weights of each channel,enhancing the network’s focus on important features.Finally,comparative experiments are performed on available datasets,including the public human activity recognition using smartphone dataset(UCIHAR),the public human activity recognition 70 plus dataset(HAR70PLUS),and the independently developed home abnormal behavior recognition dataset(HABRD)created by the authors’team.The results show that the average accuracy of the proposed PSE-CNN-BiGRU fusion model for human posture recognition is 99.56%,89.42%,and 98.90%,respectively,which are 5.24%,5.83%,and 3.19%higher than the average accuracy of the five models proposed in the comparative literature,including CNN,GRU,and others.The F1-score for abnormal posture recognition reaches 98.84%(heartache),97.18%(fall),99.6%(bellyache),and 98.27%(climbing)on the self-builtHABRDdataset,thus verifying the effectiveness,generalization,and robustness of the proposed model in enhancing human posture recognition.

A Survey on Artificial Intelligence in Posture Recognition

Over the years,the continuous development of new technology has promoted research in the field of posture recognition and also made the application field of posture recognition have been greatly expanded.The purpose of this paper is to introduce the latest methods of posture recognition and review the various techniques and algorithms of posture recognition in recent years,such as scale-invariant feature transform,histogram of oriented gradients,support vectormachine(SVM),Gaussian mixturemodel,dynamic time warping,hiddenMarkovmodel(HMM),lightweight network,convolutional neural network(CNN).We also investigate improved methods of CNN,such as stacked hourglass networks,multi-stage pose estimation networks,convolutional posemachines,and high-resolution nets.The general process and datasets of posture recognition are analyzed and summarized,and several improved CNNmethods and threemain recognition techniques are compared.In addition,the applications of advanced neural networks in posture recognition,such as transfer learning,ensemble learning,graph neural networks,and explainable deep neural networks,are introduced.It was found that CNN has achieved great success in posture recognition and is favored by researchers.Still,a more in-depth research is needed in feature extraction,information fusion,and other aspects.Among classification methods,HMM and SVM are the most widely used,and lightweight network gradually attracts the attention of researchers.In addition,due to the lack of 3Dbenchmark data sets,data generation is a critical research direction.

A Human Body Posture Recognition Algorithm Based on BP Neural Network for Wireless Body Area Networks

Human body posture recognition has attracted considerable attention in recent years in wireless body area networks(WBAN). In order to precisely recognize human body posture,many recognition algorithms have been proposed.However, the recognition rate is relatively low. In this paper, we apply back propagation(BP) neural network as a classifier to recognizing human body posture, where signals are collected from VG350 acceleration sensor and a posture signal collection system based on WBAN is designed. Human body signal vector magnitude(SVM) and tri-axial acceleration sensor data are used to describe the human body postures. We are able to recognize 4postures: Walk, Run, Squat and Sit. Our posture recognition rate is up to 91.67%. Furthermore, we find an implied relationship between hidden layer neurons and the posture recognition rate. The proposed human body posture recognition algorithm lays the foundation for the subsequent applications.

Moving Human Posture Recognition Based on Joint Quaternion

Posture recognition plays an important role in many applications,such as security system and monitoring system.Joint quaternion combined with support vector machine(SVM) can solve the problem of moving human posture recognition.It is a simple and effective algorithm that only three joints are used as the feature points in the whole human skeleton.Using the quaternion of the three joints,a feature vector with five parameters in gait cycle is extracted.The efficiency of the proposed method is demonstrated through an experimental study,and walking and running postures can be distinguished accurately.

Low-Cost Posture Recognition of Moving Hands by Profile-Mold Construction in Cluttered Background and Occlusion

In this paper, we propose a low-cost posture recognition scheme using a single webcam for the signaling hand with nature sways and possible oc-clusions. It goes for developing the untouchable low-complexity utility based on friendly hand-posture signaling. The scheme integrates the dominant temporal-difference detection, skin color detection and morphological filtering for efficient cooperation in constructing the hand profile molds. Those molds provide representative hand profiles for more stable posture recognition than accurate hand shapes with in effect trivial details. The resultant bounding box of tracking the signaling molds can be treated as a regular-type object-matched ROI to facilitate the stable extraction of robust HOG features. With such commonly applied features on hand, the prototype SVM is adequately capable of obtaining fast and stable hand postures recognition under natural hand movement and non-hand object occlusion. Experimental results demonstrate that our scheme can achieve hand-posture recognition with enough accuracy under background clutters that the targeted hand can be allowed with medium movement and palm-grasped object. Hence, the proposed method can be easily embedded in the mobile phone as application software.

IoMT-Enabled Fusion-Based Model to Predict Posture for Smart Healthcare Systems

Smart healthcare applications depend on data from wearable sensors(WSs)mounted on a patient’s body for frequent monitoring information.Healthcare systems depend on multi-level data for detecting illnesses and consequently delivering correct diagnostic measures.The collection of WS data and integration of that data for diagnostic purposes is a difficult task.This paper proposes an Errorless Data Fusion(EDF)approach to increase posture recognition accuracy.The research is based on a case study in a health organization.With the rise in smart healthcare systems,WS data fusion necessitates careful attention to provide sensitive analysis of the recognized illness.As a result,it is dependent on WS inputs and performs group analysis at a similar rate to improve diagnostic efficiency.Sensor breakdowns,the constant time factor,aggregation,and analysis results all cause errors,resulting in rejected or incorrect suggestions.This paper resolves this problem by using EDF,which is related to patient situational discovery through healthcare surveillance systems.Features of WS data are examined extensively using active and iterative learning to identify errors in specific postures.This technology improves position detection accuracy,analysis duration,and error rate,regardless of user movements.Wearable devices play a critical role in the management and treatment of patients.They can ensure that patients are provided with a unique treatment for their medical needs.This paper discusses the EDF technique for optimizing posture identification accuracy through multi-feature analysis.At first,the patients’walking patterns are tracked at various time intervals.The characteristics are then evaluated in relation to the stored data using a random forest classifier.

复杂人机共融场景中人体姿态识别及避碰策略综述

智能机器人与人类智慧的融合,即人机协作共融,已经实现了将机器人的机械优势和人类的高级认知能力集中于同一个工作架构之中,能够在复杂环境中协同作业,从而提高效率。针对复杂的人机共融场景,特别是机器人在诸如光线条件变化、背景干扰以及运动过程,对比总结了基于机器视觉的人体姿态识别方法和基于机器学习的避碰策略,详细比较各类方法的研究现状及应用,并探讨了基于深度学习的目标识别和避碰方法的发展及应用。

煤矿防冲钻孔机器人全自主钻进系统关键技术

针对高地应力矿井钻孔卸压作业智能化程度低的技术难题,总结分析了国内外钻孔卸压技术和装备的研究现状,指出研发高性能、高可靠、高效率的防冲钻孔机器人全自主钻进系统是破解冲击地压防治难题的重要发展方向。为此,凝练了影响钻进系统性能的“孔位精准识别、钻具姿态精确感知、无线电磁随钻智能检测、钻具运行状态智能识别和钻进系统精确控制”五大关键技术,并给出了解决思路和方法。针对在复杂恶劣环境下卸压孔的精确识别问题,设计了融合图像尺寸调节和多阶段训练模式的卸压孔图像样本扩充SinGAN模型,引入多层特征融合优化的FasterRCNN,构建了基于改进SqueezeNet轻量级网络架构的孔位识别模型,以实现卸压孔位的准确快速识别;针对钻具姿态精确感知问题,提出了基于改进梯度下降法算法优化无迹卡尔曼滤波的惯性测量单元(Inertial Measurement Unit,IMU)初始对准方法,设计了多个IMU的空间阵列布局方式,研究了基于BP神经网络的钻具姿态误差补偿方法,旨在提高钻具姿态的解算精度,实现精准钻孔卸压;针对复杂地质环境下钻进工况的精确检测问题,搭建了煤矿井下随钻测量无线电磁传输系统架构,探讨了微弱电磁波信号自适应调制和随钻高速双向电磁传输技术原理,研究了孔底地质参数、几何参数和工程参数的测量原理和实现过程;针对钻进系统运行状态识别问题,构建了钻进信号时域、频域、时频域的多域特征和深度网络高级特征提取架构,提出了钻进系统关键零部件健康状态评估和故障诊断技术,构建了基于改进蝙蝠优化长短期记忆网络的卡钻风险因子预测模型,实现对卸压钻具卡钻状态的准确预测;针对钻进系统的精确控制问题,分析了钻进系统的液压系统工作原理,构建了考虑煤岩性状的钻进系统精确控制方案,探讨了基于转矩和位置的钻进系统最优控制参数求解原理,旨在实现钻进回转系统和给进系统的智能协同控制和并行作业。

基于CenterNet的跑步姿态鉴别系统的设计

为了改善目前大众跑步姿势普遍不规范的现状,提出了一种基于CenterNet的跑步姿态鉴别系统。首先,通过截图、拍照的方式自制数据集,并对数据集进行清洗、标注和分析,消除数据无关信息与简化数据。其次,引入多尺度通道注意力机制与添加十字星变形卷积2种方式改进CenterNet算法模型,将动作图像转化为数字信息和特征向量,并以此为基础,利用KNN(K-nearest neighbors)算法对跑步姿态类型进行分类。最后,与经典模型方案进行对比,验证改进CenterNet算法鉴别系统的有效性。结果表明:改进的CenterNet模型的精确率与召回率都有所提升,其参数量与计算量降低。所提算法模型能够对大多数不良姿势作出及时、准确反馈,有效帮助跑步爱好者发现问题,从而改善跑步姿态、提高运动效率、预防伤病。

采用静态数据增强的AGV定位与姿态修正研究

搭载激光雷达的移动机器人(AGV)被大量应用于智能工厂的工件运输,但受障碍物和移动物体的影响,AGV在车间内位姿辨别能力较弱。为了解决定位和姿态识别问题,提出一种全新的点云数据融合位姿辨别策略,利用高精度的静态全站仪整体数据融合激光雷达局部、低精度数据,并提出向量权重匹配法来完成AGV的室内定位。设计一种抽样网格卷积法实现异构数据的快速初步定位;建立自适应搜索全站仪数据的基准区域,将它映射到激光雷达数据的对应区域;最后通过向量权重匹配获取AVG的位姿参数。上述方法在6 m×8 m室内空间进行实验。结果表明:所提方法可达到±7 mm的定位精度与±1.4°的姿态控制识别精度,且能准确补偿激光雷达的扫描误差,提高AGV的位姿识别能力。

基于深度学习的人体姿态识别算法研究

人体姿态识别技术具有巨大的研究价值及广泛的应用前景。提出一种基于深度学习的人体姿态识别算法及能量序列改进Openpose算法,提取人体骨骼关键点,结合卷积神经网络(CNN)和长短期记忆网络(LSTM)捕获人体骨架数据空间与时间特征,得到最终的分类和识别结果,在数据集上验证。实验结果表明,提出的算法在人体姿态识别任务上取得了较好的性能。

改进的双流多模态信息融合坐姿识别方法

不正确的坐姿通常会导致青少年近视、脊柱侧弯和退行性疾病。研究能够快速、准确识别不规律坐姿的智能监测技术,有助于保持正确的姿势并预防健康问题。为了解决RGB图像易受光照强度以及遮挡因素的干扰并造成的识别率不高等问题,通过采用双流RGB-D图像作为双输入,利用ResNet网络中的残差结构改进EfficientNet基线网络结构,提出了一种基于改进R-EfficientNet的双流RGB-D多模态信息融合的坐姿识别方法。试验结果表明,提出的R-EfficientNet融合方法模型对8种坐姿的识别均值平均精度(mean average precision,mAP)达到了98.5%。与CNN、Vgg16、ResNet18、EfficientNet、RGB-D不同的输入方法相比,所提方法获得了最高的识别率。该方法不仅可以用于坐姿客观监测,具有医学和社会效益,此外还为人体工学研究者们提供改进办公家具的方案。

压电复合薄膜触觉传感器的研制及其坐姿识别应用

对溶液流延法和静电纺丝法制备出的聚偏二氟乙烯(PVDF)/钛酸钡(BaTiO_(3))复合薄膜进行压电性能分析,并且对不同质量分数BaTiO_(3)的压电复合薄膜进行性能测试,确定最佳的掺杂比例,使用丝网印刷方法制备出传感器阵列。设计信号转换电路、复位电路、供电电路、多路选通电路,使用PyQt5编写上位机图形可视化界面。在坐姿的识别过程中对五种不同的坐姿(左倾、右倾、正常、前倾、后倾)进行数据采集,对采集到的数据进行线性插值滤波,使用支持向量机(SVM)对数据进行分类,通过添加不同的核函数进行对比分析,实验结果表明识别准确率达到88.33%。

基于图卷积网络的儿童坐姿检测学习桌椅设计方法研究

儿童长期处于伏案学习的状态,不良坐姿对儿童生长发育容易造成严重影响,对儿童坐姿的矫正已刻不容缓。通过对儿童在学习桌椅上的坐姿行为调研,基于OpenPose姿态估计算法进行坐姿检测分析,引入ST-GCN模型进行坐姿识别与评判,结果表明:使用ST-GCN模型能够快速准确识别儿童的八种坐姿,并根据识别结果对儿童进行有效的错误坐姿提示,其Macro-F1和Micro-F1评价指标分别提高了6.8%和7.4%。同时表明儿童坐姿矫正在自适应儿童学习桌椅上应用的可行性及有效性,可为智能儿童课桌椅的设计提供技术支撑。

基于关键点检测的坐姿识别方法

针对直接应用深度学习分类算法所得模型泛用性较差的问题,提出了一种分两个步骤完成人体坐姿识别的方法。所提方法首先提取图片中人体上身关键点的坐标信息,在关键点检测环节采用高分辨率主干网络,并进一步改进了模型结构,在下采样环节中引入SE(Squeeze-and-Excitation)注意力机制,加强了空间位置特征的表达,取得了更高的检测平均精准度;然后采用随机森林算法对关键点进行坐姿分类。实验结果表明:所提方法识别准确率可以达到94%以上,并且在陌生场景下有更好的泛用性,能适应实际应用中复杂的人物环境。

基于Mediapipe的人体姿态跟随机器人系统的设计与实现

本研究的机器人系统采用计算机视觉对人体姿态进行动态标识,从现有的单目相机获得图像或视频,利用Mediapipe方法对32个人体关键部位进行了定位和可信度的测量。控制系统通过反馈关节位置标记进行人机关节映射,控制驱动元件复现人体实时自然交互,实现人体跟随的功能。该系统的随机实验结果证明了系统设计的正确性和可行性,这对人体姿态跟随机器人系统的仿真研究有一定的指导作用。

基于不同虚拟技术的掘进机培训方案对比研究

传统掘进机培训方式效果较差,专业高效的培训方案可以提高对掘进机的掌握程度。设计了四种基于不同虚拟技术的掘进机培训方案,并通过对比分析加以实验验证选择出了最佳方案。硬件部分主要设计了矩阵键盘和控制手柄以及多种视觉显示设备,软件部分使用建模技术构建掘进机模型,使用虚拟技术构建培训场景,使用Open-Pose技术规范员工姿态。结果表明:在四种掘进机培训方案中,物理实体与混合现实技术的掘进机培训方案的培训效率最高。

面向煤矿井下作业场景的安全帽佩戴识别算法

为提高煤矿井下作业场景的工人安全帽佩戴识别准确度,提出一种人体姿态分析的安全帽佩戴识别算法和机器视觉系统优化方法。首先,选择单步多框目标检测(SSD)作为多目标识别的基础模型,利用网络压缩模型(SqueezeNet)削减改进模型参数,以形成高效的识别模型,在提高对工人安全帽的识别准确度的同时,维持识别准确度与计算速度之间的平衡;然后,引入多人姿态估计算法,定位每个工人的人体关节点,判断工人的复杂行为状态;最后,基于多目标识别与多人姿态估计的融合模型,提取目标人员的上肢关节点,结合其与安全帽包围框的空间拓扑结构关系,判定安全帽的佩戴情况,选取3000条图像数据验证所提方法的有效性。结果表明:机器视觉系统优化方法能优化硬、软件配置,提高系统性能。该方法识别准确度达到91.1%,明显优于最新的安全帽佩戴识别算法,能够满足煤矿井下作业的工人安全帽佩戴识别准确度要求。

基于关键点亲和场与支持向量机的人体姿态识别

针对传统动作姿态识别仍需物理数据采集设备或深度体感设备进行手工提取特征的问题,提出一种基于关键点亲和场与支持向量机的人体姿态识别方法。以关键点亲和场为核心进行关节点检测,获取各种姿态下的18个关节点坐标信息,使用标准化后的坐标数据对支持向量机模型加以训练,选择不同的高斯核函数进行对比。实现在没有人体深度信息和无穿戴设备的情况下,只使用普通RGB图片便可对人体姿态进行分类识别的效果。实验表明它在KTH数据集、Weizmann数据集中的识别效果良好;在自采集数据集中与带有传感器的方法相比,缩减操作步骤的同时准确率提高了7百分点。另外,还在保持关节点检测不变的情况下,使用随机森林、KNN算法进行姿态分类对比,实验结果证明该方法优于后两者。

基于人体姿态识别的无人机控制系统设计与实现

研究了一种基于人体姿态识别的无人机控制系统,该系统能够通过识别人体姿态来实现对无人机的控制。该控制系统具有较高的精度和健壮性,能够实现无人机的自主控制和高效运行。首先介绍无人机控制系统的基本概念和原理,接着详细描述了基于人体姿态识别的无人机控制系统的构建过程和实现方法,并分析了该系统的优缺点和应用前景。最后,对未来的研究方向进行了展望。