Vue+tracking.js 實現(xiàn)前端人臉檢測功能
項目中需要實現(xiàn)人臉登陸功能,實現(xiàn)思路為在前端檢測人臉,把人臉照片發(fā)送到后端識別,返回用戶token登陸成功
前端調(diào)用攝像頭使用tracking.js檢測視頻流中的人臉,檢測到人臉后拍照上傳后端。
后端使用face_recognition人臉識別庫,使用Flask提供restfulAP供前端調(diào)用
實現(xiàn)效果如下圖:
登陸界面:
攝像頭檢測人臉界面:
前端代碼如下:
<template> <div id="facelogin"> <h1 class="title is-1">{{FaceisDetected}}</h1> <!-- <p>{{FaceisDetected}}</p> --> <div class="content-cam"> <div class="camera-wrp sec"> <video width="320" height="320" ref="videoDom" id="video_cam" preload autoplay loop muted></video> <canvas width="320" height="320" ref="canvasDOM" id="face_detect"></canvas> <div class="control-btn"></div> </div> <div class="images-wrp sec"> <!-- <p class="title is-5">Image taken</p> --> <div :class="`img-item img-item-${index}`" v-for="(image, index) in images" :key="`img-wrp-${index}`" :style="`background-image: url('${image}')`" ></div> </div> </div> </div> </template> export default { name: 'facelogin', data() { return { count: 0, isdetected: '請您保持臉部在畫面中央', videoEl: {}, canvasEL: {}, images: [], trackCcv: false, trackTracking: false, autoCaptureTrackTraking: false, userMediaConstraints: { audio: false, video: { // ideal(應(yīng)用最理想的) width: { min: 320, ideal: 1280, max: 1920 }, height: { min: 240, ideal: 720, max: 1080 }, // frameRate受限帶寬傳輸時,低幀率可能更適宜 frameRate: { min: 15, ideal: 30, max: 60 }, // 攝像頭翻轉(zhuǎn) facingMode: 'user' } } } }, computed: { FaceisDetected() { return this.isdetected } }, created() { this.changeView() }, mounted() { // The getUserMedia interface is used for handling camera input. // Some browsers need a prefix so here we're covering all the options navigator.getMedia = navigator.getUserMedia || navigator.webkitGetUserMedia || navigator.mozGetUserMedia || navigator.msGetUserMedia this.init() }, methods: { async init() { this.videoEl = this.$refs.videoDom this.canvasEL = this.$refs.canvasDOM await navigator.mediaDevices .getUserMedia(this.userMediaConstraints) .then(this.getMediaStreamSuccess) .catch(this.getMediaStreamError) await this.onPlay() }, async onPlay() { debugHelper.log('onPlay') this.onTrackTracking() }, changeView() { this.setTitle('刷臉登陸') this.setBackDisabled(false) this.setBackIcon('arrow_back') msgbus.vm.setBottomNavVisible(false) msgbus.vm.setBottomBtnVisible(false) msgbus.vm.setMsgInputVisible({ value: false }) }, onTrackTracking() { const context = this const video = this.videoEl const canvas = this.canvasEL const canvasContext = canvas.getContext('2d') let tracker = new tracking.ObjectTracker('face') video.pause() video.src = '' tracker.setInitialScale(4) tracker.setStepSize(2) tracker.setEdgesDensity(0.1) tracking.track('#video_cam', tracker, { camera: true }) tracker.on('track', function(event) { const { autoCaptureTrackTraking } = context canvasContext.clearRect(0, 0, canvas.width, canvas.height) event.data.forEach(function({ x, y, width, height }) { canvasContext.strokeStyle = '#a64ceb' canvasContext.strokeRect(x, y, width, height) canvasContext.font = '11px Helvetica' canvasContext.fillStyle = '#fff' }) if (!isEmpty(event.data) && context.count <= 10) { if (context.count < 0) context.count = 0 context.count += 1 //debugHelper.log(context.count) if (context.count > 10) { context.isdetected = '已檢測到人臉,正在登錄' //context.$router.push({ name: 'pwdlogin' }) } } else { context.count -= 1 if (context.count < 0) context.isdetected = '請您保持臉部在畫面中央' //this.isdetected = '已檢測到人臉,正在登錄' } }) }, onDownloadFile(item) { const link = document.createElement('a') link.href = item link.download = `cahyo-${new Date().toISOString()}.png` link.click() link.remove() }, onTakeCam() { const canvas = document.createElement('canvas') const video = this.$el.querySelector('#video_cam') const canvasContext = canvas.getContext('2d') if (video.videoWidth && video.videoHeight) { const isBiggerW = video.videoWidth > video.videoHeight const fixVidSize = isBiggerW ? video.videoHeight : video.videoWidth let offsetLeft = 0 let offsetTop = 0 if (isBiggerW) offsetLeft = (video.videoWidth - fixVidSize) / 2 else offsetTop = (video.videoHeight - fixVidSize) / 2 // make canvas size 300px canvas.width = canvas.height = 300 const { width, height } = canvas canvasContext.drawImage( video, offsetLeft, offsetTop, fixVidSize, fixVidSize, 0, 0, width, height ) const image = canvas.toDataURL('image/png') this.images.push(image) } }, onDetectFace(param, index) { const imgItem = document.querySelector(`.img-item-${index}`) const image = new Image() image.src = param const tracker = new tracking.ObjectTracker('face') tracker.setStepSize(1.7) tracking.track(image, tracker) tracker.on('track', function(event) { event.data.forEach(function(rect) { window.plot(rect.x, rect.y, rect.width, rect.height) }) }) window.plot = function(x, y, w, h) { const rect = document.createElement('div') document.querySelector(`.img-item-${index}`).appendChild(rect) rect.classList.add('rect') rect.style.width = w + 'px' rect.style.height = h + 'px' rect.style.left = x + 'px' rect.style.top = y + 'px' rect.style.border = '2px solid yellow' rect.style.position = 'absolute' } }, getMediaStreamSuccess(stream) { window.stream = stream // make stream available to browser console this.videoEl.srcObject = stream debugHelper.log('getMediaStreamSuccess1') //this.$store.commit('setVideoCanvasObject', this.videoEl) debugHelper.log('getMediaStreamSuccess2') }, // 視頻媒體流失敗 getMediaStreamError(error) { alert('視頻媒體流獲取錯誤' + error) }, // 結(jié)束媒體流 stopMediaStreamTrack() { clearInterval(this.timeInterval) if (typeof window.stream === 'object') { this.videoEl.srcObject = null //this.$store.commit('setVideoCanvasObject', '') window.stream.getTracks().forEach(track => track.stop()) } },
總結(jié)
到此這篇關(guān)于Vue+tracking.js 實現(xiàn)前端人臉檢測功能的文章就介紹到這了,更多相關(guān)vue tracking.js 人臉檢測內(nèi)容請搜索腳本之家以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持腳本之家!
相關(guān)文章
詳解用vue2.x版本+adminLTE開源框架搭建后臺應(yīng)用模版
這篇文章主要介紹了用vue2.x版本+adminLTE開源框架 搭建后臺應(yīng)用模版,小編覺得挺不錯的,現(xiàn)在分享給大家,也給大家做個參考。一起跟隨小編過來看看吧2019-03-03vue antd的from表單中驗證rules中type的坑記錄
這篇文章主要介紹了vue antd的from表單中驗證rules中type的坑記錄,具有很好的參考價值,希望對大家有所幫助。如有錯誤或未考慮完全的地方,望不吝賜教2023-04-04Vue 實現(xiàn)CLI 3.0 + momentjs + lodash打包時優(yōu)化
今天小編就為大家分享一篇Vue 實現(xiàn)CLI 3.0 + momentjs + lodash打包時優(yōu)化,具有很好的參考價值,希望對大家有所幫助。一起跟隨小編過來看看吧2019-11-11vue實現(xiàn)一個獲取按鍵展示快捷鍵效果的Input組件
這篇文章主要介紹了vue如何實現(xiàn)一個獲取按鍵展示快捷鍵效果的Input組件,幫助大家更好的理解和使用vue框架,感興趣的朋友可以了解下2021-01-01Vue集成three.js并加載glb、gltf、FBX、json模型的場景分析
這篇文章主要介紹了Vue集成three.js,并加載glb、gltf、FBX、json模型,本文通過實例代碼給大家介紹的非常詳細,對大家的學(xué)習(xí)或工作具有一定的參考借鑒價值,需要的朋友可以參考下2022-09-09深入了解vue-router原理并實現(xiàn)一個小demo
這篇文章主要為大家詳細介紹了vue-router原理并實現(xiàn)一個小demo,文中示例代碼介紹的非常詳細,具有一定的參考價值,感興趣的小伙伴們可以參考一下,希望能夠給你帶來幫助2022-03-03