In response to the needs of physical health data management in the context of the Internet of Everything, this article first uses cloud computing, big data, mobile Internet and other technologies to build a physical health smart management system. When the system is deployed, edge nodes are introduced in each data collection area, and the system is composed of data collection, transmission, and query and analysis modules. Secondly, it uses convolutional neural network to learn features from body measurement data unsupervised. Then, based on the Gaussian mixture distribution, a three-level physical fitness assessment model was established. Finally, input the learned features into the evaluation model to get the result of physical fitness evaluation. The results show that the system not only has a better response to the family, but also can reduce operating costs and improve work efficiency. Moreover, the algorithm in this paper is not affected by individual physical fitness assessment methods and results, and provides new ideas and methods for physical fitness assessment.

针对万物互联背景下身体健康数据管理的需求，本文首先利用云计算、大数据、移动互联网等技术构建身体健康智能管理系统。系统部署时，在每个数据采集区域引入边缘节点，系统由数据采集、传输和查询分析模块组成。其次，它使用卷积神经网络从无监督的身体测量数据中学习特征。然后，基于高斯混合分布，建立三级体能评估模型。最后，将学习到的特征输入到评估模型中，得到体能评估结果。结果表明，该系统不仅对家庭有较好的反应，还能降低运营成本，提高工作效率。而且，本文算法不受个体体能评估方法和结果的影响，为体能评估提供了新的思路和方法。