基于卷积神经网络的电力查询设备手势操控算法研究  

Research on gesture manipulation algorithm of power inquiry equipment based on convolution neural network

作　　者：张凡[1,2] 刘文达 蔡涛 范亚雷 Zhang Fan;Liu Wenda;Cai Tao;Fan Yalei(Hubei Collaborative Innovation Center for High-efficiency Utilization of Solar Energy,Hubei University of Technology,Wuhan 430068,Chin a;Hubei Key Laboratory for High-efficiency Utilization of Solar Energy and Operation Control of Energy Storage System.Hubei University of Technology,Wuhan 430068,China;Hubei Power Grid Intelligent Control and Equipm ent Engineering Technology Research Center,Hubei University of Technology,Wuhan 430068,China)

机构地区：[1]湖北工业大学太阳能高效利用湖北省协同创新中心,武汉430068 [2]湖北工业大学太阳能高效利用及储能运行控制湖北省重点实验室,武汉430068 [3]湖北工业大学电网智能控制与装备工程技术研究中心,武汉430068

出　　处：《电子测量技术》2019年第24期98-102,共5页Electronic Measurement Technology

基　　金：国家自然科学基金(11605051);湖北省高等学校优秀中青年科技创新团队计划项目(T201805)资助。

摘　　要：当电力设备噪声大且不在可接触范围内时,工作人员难以通过触屏、触键或语音方式操控查询设备。探索了一种基于卷积神经网络的手势识别算法。首先采用基于YCrCb的颜色空间变换和基于OSTU的自适应图像二值化对图像进行压缩处理,然后采用两层卷积层对特征进行提取,最后用Softmax进行分类。为了提高模型的计算速度,激活函数优化为ReLu,并添加了Dropout层。结果表明,算法能够识别自定义的9种手势,平均识别率高于99.6%。因满足精度要求,故可用于电力查询设备的操控。In some specific working environments, it is difficult for a worker to mainpulate the inquiry device by touch screen, touch buttons or voice. To solve the above problems, a gesture recognition algorithm based on convolution neural network is studied. Firstly,the image is compressed by the YCrCb-based color space transform and OSTU-based image binarization,and then the features are extracted by two convolution layers.Lastly,the gestures are classified by Softmax. In order to improve the speed of calculation,the activation function is optimized to ReLu and the Dropout layer is added.Results show that the model can recognize 9 gestures and the AP is higher than 99.6%.It can be used for the control of power inquiry equipment because it meets the accuracy requirements.

关 键 词：电力查询设备 卷积神经网络 平均识别率 手势识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]