最近帮别人做一个项目，主要是使用摄像头做人脸识别

github地址：https://github.com/qugang/AVCaptureVideoTemplate

要使用IOS的摄像头，需要使用AVFoundation 库，库里面的东西我就不介绍。

启动摄像头需要使用AVCaptureSession 类。

然后得到摄像头传输的每一帧数据，需要使用AVCaptureVideoDataOutputSampleBufferDelegate 委托。

首先在viewDidLoad 里添加找摄像头设备的代码，找到摄像头设备以后，开启摄像头

1
2
3
4
5
6
7
8
9
10
11
12
13
captureSession.sessionPreset = AVCaptureSessionPresetLow
let devices = AVCaptureDevice.devices()
for device in devices {
  if (device.hasMediaType(AVMediaTypeVideo)) {
    if (device.position == AVCaptureDevicePosition.Front) {
      captureDevice = device as?AVCaptureDevice
      if captureDevice != nil {
        println("Capture Device found")
        beginSession()
      }
    }
  }
}

beginSession,开启摄像头：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
func beginSession() {
  var err : NSError? = nil
  captureSession.addInput(AVCaptureDeviceInput(device: captureDevice, error: &err))
  let output = AVCaptureVideoDataOutput()
  let cameraQueue = dispatch_queue_create("cameraQueue", DISPATCH_QUEUE_SERIAL)
  output.setSampleBufferDelegate(self, queue: cameraQueue)
  output.videoSettings = [kCVPixelBufferPixelFormatTypeKey: kCVPixelFormatType_32BGRA]
  captureSession.addOutput(output)
  if err != nil {
    println("error: \(err?.localizedDescription)")
  }
  previewLayer = AVCaptureVideoPreviewLayer(session: captureSession)
  previewLayer?.videoGravity = "AVLayerVideoGravityResizeAspect"
  previewLayer?.frame = self.view.bounds
  self.view.layer.addSublayer(previewLayer)
  captureSession.startRunning()
}

开启以后，实现captureOutput 方法：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
func captureOutput(captureOutput: AVCaptureOutput!, didOutputSampleBuffer sampleBuffer: CMSampleBuffer!, fromConnection connection: AVCaptureConnection!) {
  if(self.isStart)
  {
    let resultImage = sampleBufferToImage(sampleBuffer)
    let context = CIContext(options:[kCIContextUseSoftwareRenderer:true])
    let detecotr = CIDetector(ofType:CIDetectorTypeFace,  context:context, options:[CIDetectorAccuracy: CIDetectorAccuracyHigh])
    let ciImage = CIImage(image: resultImage)
    let results:NSArray = detecotr.featuresInImage(ciImage,options: ["CIDetectorImageOrientation" : 6])
    for r in results {
      let face:CIFaceFeature = r as! CIFaceFeature;
      let faceImage = UIImage(CGImage: context.createCGImage(ciImage, fromRect: face.bounds),scale: 1.0, orientation: .Right)
      NSLog("Face found at (%f,%f) of dimensions %fx%f", face.bounds.origin.x, face.bounds.origin.y,pickUIImager.frame.origin.x, pickUIImager.frame.origin.y)
      dispatch_async(dispatch_get_main_queue()) {
        if (self.isStart)
        {
          self.dismissViewControllerAnimated(true, completion: nil)
          self.didReceiveMemoryWarning()
          self.callBack!(face: faceImage!)
        }
        self.isStart = false
      }
    }
  }
}

在每一帧图片上使用CIDetector 得到人脸，CIDetector 还可以得到眨眼，与微笑的人脸，如果要详细使用去官方查看API

上面就是关键代码，设置了有2秒的延迟，2秒之后开始人脸检测。

全部代码：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
//
//  ViewController.swift
//  AVSessionTest
//
//  Created by qugang on 15/7/8.
//  Copyright (c) 2015年 qugang. All rights reserved.
//
 
import UIKit
import AVFoundation
class AVCaptireVideoPicController: UIViewController,AVCaptureVideoDataOutputSampleBufferDelegate {
  var callBack :((face: UIImage) ->())?
  let captureSession = AVCaptureSession()
  var captureDevice : AVCaptureDevice?
  var previewLayer : AVCaptureVideoPreviewLayer?
  var pickUIImager : UIImageView = UIImageView(image: UIImage(named: "pick_bg"))
  var line : UIImageView = UIImageView(image: UIImage(named: "line"))
  var timer : NSTimer!
  var upOrdown = true
  var isStart = false
  override func viewDidLoad() {
    super.viewDidLoad()
    captureSession.sessionPreset = AVCaptureSessionPresetLow
    let devices = AVCaptureDevice.devices()
    for device in devices {
      if (device.hasMediaType(AVMediaTypeVideo)) {
        if (device.position == AVCaptureDevicePosition.Front) {
          captureDevice = device as?AVCaptureDevice
          if captureDevice != nil {
            println("Capture Device found")
            beginSession()
          }
        }
      }
    }
    pickUIImager.frame = CGRect(x: self.view.bounds.width / 2 - 100, y: self.view.bounds.height / 2 - 100,width: 200,height: 200)
    line.frame = CGRect(x: self.view.bounds.width / 2 - 100, y: self.view.bounds.height / 2 - 100, width: 200, height: 2)
    self.view.addSubview(pickUIImager)
    self.view.addSubview(line)
    timer =  NSTimer.scheduledTimerWithTimeInterval(0.01, target: self, selector: "animationSate", userInfo: nil, repeats: true)
     
    NSTimer.scheduledTimerWithTimeInterval(2, target: self, selector: "isStartTrue", userInfo: nil, repeats: false)
  }
  func isStartTrue(){
    self.isStart = true
  }
  override func didReceiveMemoryWarning(){
    super.didReceiveMemoryWarning()
    captureSession.stopRunning()
  }
   
  func animationSate(){
    if upOrdown {
      if (line.frame.origin.y >= pickUIImager.frame.origin.y + 200)
      {
        upOrdown = false
      }
      else
      {
        line.frame.origin.y += 2
      }
    } else {
      if (line.frame.origin.y <= pickUIImager.frame.origin.y)
      {
        upOrdown = true
      }
      else
      {
        line.frame.origin.y -= 2
      }
    }
  }
  func beginSession() {
    var err : NSError? = nil
    captureSession.addInput(AVCaptureDeviceInput(device: captureDevice, error: &err))
    let output = AVCaptureVideoDataOutput()
    let cameraQueue = dispatch_queue_create("cameraQueue", DISPATCH_QUEUE_SERIAL)
    output.setSampleBufferDelegate(self, queue: cameraQueue)
    output.videoSettings = [kCVPixelBufferPixelFormatTypeKey: kCVPixelFormatType_32BGRA]
    captureSession.addOutput(output)
    if err != nil {
      println("error: \(err?.localizedDescription)")
    }
    previewLayer = AVCaptureVideoPreviewLayer(session: captureSession)
    previewLayer?.videoGravity = "AVLayerVideoGravityResizeAspect"
    previewLayer?.frame = self.view.bounds
    self.view.layer.addSublayer(previewLayer)
    captureSession.startRunning()
  }
  func captureOutput(captureOutput: AVCaptureOutput!, didOutputSampleBuffer sampleBuffer: CMSampleBuffer!, fromConnection connection: AVCaptureConnection!) {
    if(self.isStart)
    {
      let resultImage = sampleBufferToImage(sampleBuffer)
      let context = CIContext(options:[kCIContextUseSoftwareRenderer:true])
      let detecotr = CIDetector(ofType:CIDetectorTypeFace,  context:context, options:[CIDetectorAccuracy: CIDetectorAccuracyHigh])
      let ciImage = CIImage(image: resultImage)
      let results:NSArray = detecotr.featuresInImage(ciImage,options: ["CIDetectorImageOrientation" : 6])
      for r in results {
        let face:CIFaceFeature = r as! CIFaceFeature;
        let faceImage = UIImage(CGImage: context.createCGImage(ciImage, fromRect: face.bounds),scale: 1.0, orientation: .Right)
        NSLog("Face found at (%f,%f) of dimensions %fx%f", face.bounds.origin.x, face.bounds.origin.y,pickUIImager.frame.origin.x, pickUIImager.frame.origin.y)
        dispatch_async(dispatch_get_main_queue()) {
          if (self.isStart)
          {
            self.dismissViewControllerAnimated(true, completion: nil)
            self.didReceiveMemoryWarning()
            self.callBack!(face: faceImage!)
          }
          self.isStart = false
        }
      }
    }
  }
  private func sampleBufferToImage(sampleBuffer: CMSampleBuffer!) -> UIImage {
    let imageBuffer: CVImageBufferRef = CMSampleBufferGetImageBuffer(sampleBuffer)
    CVPixelBufferLockBaseAddress(imageBuffer, 0)
    let baseAddress = CVPixelBufferGetBaseAddressOfPlane(imageBuffer, 0)
    let bytesPerRow = CVPixelBufferGetBytesPerRow(imageBuffer)
    let width = CVPixelBufferGetWidth(imageBuffer)
    let height = CVPixelBufferGetHeight(imageBuffer)
    let colorSpace: CGColorSpaceRef = CGColorSpaceCreateDeviceRGB()
    let bitsPerCompornent = 8
    var bitmapInfo = CGBitmapInfo((CGBitmapInfo.ByteOrder32Little.rawValue | CGImageAlphaInfo.PremultipliedFirst.rawValue) as UInt32)
    let newContext = CGBitmapContextCreate(baseAddress, width, height, bitsPerCompornent, bytesPerRow, colorSpace, bitmapInfo) as CGContextRef
    let imageRef: CGImageRef = CGBitmapContextCreateImage(newContext)
    let resultImage = UIImage(CGImage: imageRef, scale: 1.0, orientation: UIImageOrientation.Right)!
    return resultImage
  }
  func imageResize (imageObj:UIImage, sizeChange:CGSize)-> UIImage{
    let hasAlpha = false
    let scale: CGFloat = 0.0
     
    UIGraphicsBeginImageContextWithOptions(sizeChange, !hasAlpha, scale)
    imageObj.drawInRect(CGRect(origin: CGPointZero, size: sizeChange))
    let scaledImage = UIGraphicsGetImageFromCurrentImageContext()
    return scaledImage
  }
}

swift通过摄像头读取每一帧的图片，并且做识别做人脸识别的更多相关文章

  1. OpenCV摄像头人脸识别

    注: 从外设摄像装置中获取图像帧,把每帧的图片与人脸特征进行匹配,用方框框住识别出来的人脸 需要用到的函数: CvHaarClassifierCascade* cvLoadHaarClassifier ...

  2. 基于OpenCV读取摄像头进行人脸检测和人脸识别

    前段时间使用OpenCV的库函数实现了人脸检测和人脸识别,笔者的实验环境为VS2010+OpenCV2.4.4,opencv的环境配置网上有很多,不再赘述.检测的代码网上很多,记不清楚从哪儿copy的 ...

  3. QQ摄像头读取条码

    跟我学机器视觉-HALCON学习例程中文详解-QQ摄像头读取条码 第一步:插入QQ摄像头,安装好驱动(有的可能免驱动) 第二步:打开HDevelop,点击助手—打开新的Image Acquisitio ...

  4. 跟我学机器视觉-HALCON学习例程中文详解-QQ摄像头读取条码

    跟我学机器视觉-HALCON学习例程中文详解-QQ摄像头读取条码 第一步:插入QQ摄像头,安装好驱动(有的可能免驱动) 第二步:打开HDevelop,点击助手-打开新的Image Acquisitio ...

  5. opencv摄像头读取图片

    # 摄像头捕获图像或视频import numpy as np import cv2 # 创建相机的对象 cap = cv2.VideoCapture(0) while(True): # 读取相机所拍到 ...

  6. Opencv摄像头实时人脸识别

    Introduction 网上存在很多人脸识别的文章,这篇文章是我的一个作业,重在通过摄像头实时采集人脸信息,进行人脸检测和人脸识别,并将识别结果显示在左上角. 利用 OpenCV 实现一个实时的人脸 ...

  7. Python3利用Dlib19.7实现摄像头人脸识别的方法

    0.引言 利用python开发,借助Dlib库捕获摄像头中的人脸,提取人脸特征,通过计算欧氏距离来和预存的人脸特征进行对比,达到人脸识别的目的: 可以自动从摄像头中抠取人脸图片存储到本地,然后提取构建 ...

  8. matlab使用摄像头人脸识别

    #关于matlab如何读取图片.视频.摄像头设备数据# 参见:http://blog.csdn.net/u010177286/article/details/45646173 但是,关于摄像头读取,上 ...

  9. OpenCV视频读取播放，视频转换为图片

    转载请注明出处!!! http://blog.csdn.net/zhonghuan1992 OpenCV视频读取播放,视频转换为图片 介绍几个有关视频读取的函数: VideoCapture::Vide ...

随机推荐

  1. 关于java读取和写入properties配置文件的内容

    一般通过使用流的方式进行读取 代码示例如下: package com.zznode.transmit.util; import java.io.FileInputStream; import java ...

  2. android-sdk-windows版本号下载

    Android SDK 4.0.3 开发环境配置及执行 近期又装了一次最新版本号的ADK环境 眼下最新版是Android SDK 4.0.3 本文的插图和文本尽管是Android2.2的 步骤都是一样 ...

  3. Android系统进程间通信（IPC）机制Binder中的Server启动过程源代码分析

    文章转载至CSDN社区罗升阳的安卓之旅,原文地址:http://blog.csdn.net/luoshengyang/article/details/6629298 在前面一篇文章浅谈Android系 ...

  4. 阻止IOS自动识别页面上的电话号码、email地址

    之前写页面的时候碰到一个很恶心的情况,在6P上数字自动变色,后来找了一些资料: 在iOS的浏览器上,他们有时候会有一些“自作聪明”,自动把页面上的一串数字识别成电话号码,这样用户不小心点击这串数字,就 ...

  5. html进阶css（3）

    css的某些样式是具有继承性的,那么什么是继承呢?继承是一种规则,它允许格式不仅应用于某个特定html标签元素,而且应用于其后代. <!doctype html> <html> ...

  6. jwplayer去Logo、自定义公司信息、限制拖动

    function initplayer(){        jwplayer("mediaplayer").setup({            primary: "fl ...

  7. CTL_CODE 宏 详解

    CTL_CODE宏 CTL_CODE:用于创建一个唯一的32位系统I/O控制代码,这个控制代码包括4部分组成: DeviceType(设备类型,高16位(16-31位)), Function(功能2- ...

  8. 求a，b在区间上的公倍数个数

    给你两个数 a,b.问你区间 [1,N]中都是有多少个数是a,b的公倍数.当数据很大的时候,遍历肯定会超时.其实,我们可以首先求出 lcm(a,b).因为我们知道(a,b)公倍数都是它最小公倍数的倍数 ...

  9. 使用LAMP创建基于wordpress的个从博客网站

    参考: http://blog.csdn.net/ck_boss/article/details/27866117 一.mysql配置 1.安装mysql yum install mysql-serv ...

  10. Ecstore的微信账号绑定会员免登录

    在微信公众号中通过消息交互/自定义菜单,可直接登录Ecstore手机商城.如果想用原商城账号登录,可以通过登录绑定,在后台可以获取登录绑定的链接这个链接可以在微信公众号消息交互场景或自定义菜单中使用, ...