基于卷积神经网络的室内虚拟控制系统设计

Design of indoor virtual control system based on convolutional neural network

为了能够实现在智能家居背景下的远程虚拟控制,解决传统家居安防系统中所存在的资源浪费,无法高效率控制的技术弊端,文中提出一种基于卷积神经网络的室内虚拟控制系统设计思路。将RaspberryPi3b+和CIS摄像头连用作为设计的硬件配置构成,通过摄像头获取室内的相关图像数据,并运用计算机系统对图像数据进行判断,根据判断结果实现对室内虚拟控制。文中首先概述了室内虚拟控制采用的机器人抓取卷积神经网络模型,并说明此次室内虚拟控制系统设计的软硬件系统配置。在系统硬件配置部分选用STM32F103C8T6型号作为系统的主控制器,设计语音控制功能模块和串口屏用于存储控制代码、图片、字库等内容。在系统软件部分设计行人检测,对人脸识别的图像进行预处理、特征提取、匹配识别。实验结果表明,所提系统能够提高室内虚拟控制移位抓取的精准度,提高控制操作简便性。

In order to realize the remote virtual control in the background of smart home,reduce wasting of resources existing in the traditional home security system,and improve the technical disadvantages that cannot be controlled with high work efficiency,a design thought of indoor virtual control system based on convolutional neural network is proposed.The combination of RaspberryPi3b+and CIS camera is used as the hardware configuration of this design,the related image data in indoor can be obtained by the camera,the image data is determined by using the computer system to,and the indoor virtual control can be realized according to the judgment results.The convolutional neural network model with robot grasping used in the indoor virtual control system is summarized,and the hardware and software system configuration designed in the indoor virtual control system is described.In the hardware configuration part of the system,STM32F103C8T6 model is selected as the main controller of the system,and the voice control function module and serial port screen are designed to store control code,pictures,font library and other contents.In the software part of the system,the pedestrian detection is designed to execute the preprocessing,feature extraction and matching recognition for the face recognition image.The design of indoor virtual control system based on convolutional neural network proposed in this paper can improve the accuracy of indoor virtual control shift capture and improve the control operation simplicity.The control system has a good prospect of promotion and application.