摘要： 物聯(lián)網(wǎng)時(shí)代下,，智能家居作為人們?nèi)粘Ｉ钪兄匾囊画h(huán),，正在飛速發(fā)展。設(shè)計(jì)了一款基于EAIDK310和OneNET云平臺(tái)的非接觸式智能門(mén)禁系統(tǒng),，采用Arm公司與Open AI Lab聯(lián)合開(kāi)發(fā)的EAIDK310人工智能開(kāi)發(fā)套件作為系統(tǒng)控制核心,，在主控板內(nèi)通過(guò)人臉識(shí)別、手勢(shì)識(shí)別,、語(yǔ)音交互等功能實(shí)現(xiàn)無(wú)接觸門(mén)禁進(jìn)出,。系統(tǒng)借助人工智能中發(fā)展較為成熟的卷積神經(jīng)網(wǎng)絡(luò)并通過(guò)百度AI開(kāi)放平臺(tái)的領(lǐng)先優(yōu)勢(shì)進(jìn)行人臉識(shí)別、手勢(shì)數(shù)字識(shí)別和語(yǔ)音實(shí)時(shí)交互,，在保證識(shí)別速度夠快的同時(shí)也能夠維持非常高的識(shí)別精準(zhǔn)度,，引入活體檢測(cè)避免不法分子使用照片欺騙系統(tǒng)，使得整個(gè)門(mén)禁系統(tǒng)更加安全,。系統(tǒng)通過(guò)Wi-Fi將數(shù)據(jù)上傳移動(dòng)OneNET云平臺(tái),，與微信小程序進(jìn)行數(shù)據(jù)圖像交互，主人可在任何地點(diǎn)通過(guò)手機(jī)對(duì)門(mén)禁系統(tǒng)進(jìn)行解鎖等操作,，并能監(jiān)控門(mén)禁外實(shí)時(shí)圖像,，能夠在住戶安全方面提供保障。
中圖分類(lèi)號(hào)： TN98,；TP274 文獻(xiàn)標(biāo)識(shí)碼： A DOI：10.16157/j.issn.0258-7998.211821 中文引用格式： 樓陳哲,，夏鯤，張文君,，等. 基于EAIDK310和OneNET云平臺(tái)的非接觸式智能門(mén)禁系統(tǒng)[J].電子技術(shù)應(yīng)用,，2022，48(5)：89-93. 英文引用格式： Lou Chenzhe,，Xia Kun,，Zhang Wenjun，et al. Non-contacted intelligent access control system based on EAIDK310 and OneNET cloud platform[J]. Application of Electronic Technique,，2022,，48(5)：89-93.
Non-contacted intelligent access control system based on EAIDK310 and OneNET cloud platform
Lou Chenzhe，Xia Kun,，Zhang Wenjun,，Mo Yingdong
Department of Electrical Engineering，University of Shanghai for Science and Technology,，Shanghai 200093,，China
Abstract： In the era of Internet of Things, smart home, as an important part of people′s daily life, is developing rapidly. This paper designs a contactless intelligent access control system based on EAIDK310 and OneNET cloud platform, using the EAIDK310 AI development kit jointly developed by Arm and Open AI Lab as the system control core, and realizing contactless access control through face recognition, gesture recognition, voice interaction and other functions within the main control board. This work uses the more mature convolutional neural network in artificial intelligence and the leading edge of Baidu AI open platform for face recognition, gesture recognition and voice interaction in real time, which can maintain a very high recognition accuracy while ensuring a fast enough recognition speed, and the introduction of live detection to avoid unscrupulous elements using photos to cheat the system, making the whole access control system more secure. This work uploads data to the mobile OneNET cloud platform through Wi-Fi, and interacts with WeChat applet for data images, so that the owner can unlock the access control system and other operations through his cell phone at any location, and can monitor real-time images outside the access control, which can provide protection in terms of occupant security.
Key words : home automation,；Internet of Things；intelligent access control system,；image recognition,；speech recognition
