中文摘要：

针对传统基于物联网技术的智能家居系统用户参与感不强、无法进行自定义手势操控等问题，设计一套基于Kinect的智能家庭管理系统。通过深度传感器得到彩色数据流、深度数据流以及骨骼数据流，从而获取人体的动作轨迹，并自定义手势操控方式来实现家居的智能控制。由于深度数据对光照和背景干扰等因素有较强的鲁棒性，提高了手势及轨迹识别的准确性和可靠性。在明暗两种光照条件下，对该文定义的4种手臂姿势、3种动作任务和5种微小手势的识别实验表明，该文设计的系统能够满足智能家居的基本需求。

英文摘要：

Since the traditional smart home system based on Internet of Things technology has the poor user participation,and can’t perform the user?defined gesture control,a smart home management system based on Kinect was designed. The colordata flow,depth data flow and skeleton data flow are got by means of the depth transducer,so as to obtain the movement trackof human behavior. The gesture control pattern is self?defined to realize the home smart control. The depth data has strong robust?ness for light and background disturbance,which can improve the accuracy and reliability of gesture and track recognition. Thedefined four arm gestures,three movement missions and five tiny hand gestures were recognized in the experiments at the brightand dark lighting conditions. The results show that the designed system can satisfy the basic demand of smart home.