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. 对于Android Service 生命周期进行全解析

    应用程序组件有一个生命周期——一开始Android实例化他们响应意图,直到结束实例被销毁.在这期间,他们有时候处于激活状态,有时候处于非激 活状态:对于活动,对用户有时候可见,有时候不可见.组件生命周 ...

  2. Python进阶之路---1.1python简介

                            Python简介 Python简介 Python (发音:[ 'paiθ(ə)n; (US) 'paiθɔn ]n.蟒蛇,巨蛇 ),是一种面向对象的解释 ...

  3. [core Java学习笔记][第一二三章基本语法]

    基本语法 1 Java 简单的类型 1.1 一些常量 正无穷大 Double.POSITVE_INFINITY 负无穷大 Double.NEGATIVE_INFINITY 不存在 Double.NaN ...

  4. sqlserver2012一直显示正在还原（Restoring）和从单用户转换成多用户模式（单用户连接中）

    如果不需要还原,则使用: restore database test with recovery如果只需要还原,则使用: restore database test with norecovery U ...

  5. OD调试2---TraceMe

    OD调试2---TraceMe 拆解一个Windows程序要比拆解一个DOS程序容易得多,因为在Windows中,只要API函数被使用,想对寻找蛛丝马迹的人隐藏一些东西是比较困难的.因此分析一个程序, ...

  6. VS2015预览版中的C#6.0 新功能（三）

    VS2015预览版中的C#6.0 新功能(一) VS2015预览版中的C#6.0 新功能(二) Using static 使用using StaticClass,你可以访问StaticClass类里的 ...

  7. hdu 2504

    Problem Description 有三个正整数a,b,c(0<a,b,c<10^6),其中c不等于b.若a和c的最大公约数为b,现已知a和b,求满足条件的最小的c. Input 第一 ...

  8. for（）循环

    今天发现自己一直以来都搞错了for()循环的执行顺序.这么简单的问题一直都错了,我也是醉了. ;i>&&a[i]>a[i-];--i) { } //即 for(init_s ...

  9. (转) 学习C++ -> 指针初步

    学习C++ -> 指针初步 一.指针    1. 什么是指针?        我们知道, 计算机的内存是由一个个独立的存储单元组成, 并且系统会对每一个存储单元分配一个唯一的号码, 称为这个存储 ...

  10. codeforces432D Prefixes and Suffixes(kmp+dp)

    转载请注明出处: http://www.cnblogs.com/fraud/          ——by fraud D. Prefixes and Suffixes You have a strin ...