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. 格而知之16：我所理解的Block（3）

    23.在前文中的例子中,Block结构体里的isa指针还没有详细讲解,这个指针都被置向了_NSConcreteStackBlock,它标识了Block的类型. 其实除了_NSConcreteStack ...

  2. VirtualBox 扩展包卸载或安装失败(VERR_ALREADY_EXISTS)

    最近在卸载VirtualBox出现了无法卸载的错误.提示为Failed to install the extension. The installer failed with exit code 1: ...

  3. Unity之Bmob云存储一

    无论我们做软件还是做游戏,少不了的就是和数据打交道,对于要保存到本地的数据,我们可以采用的载体太多了.例如:txt,Xml,Sqlite,SqlServer,Mysql等等,具体使用什么那就视情况而定 ...

  4. js获取某个标签中的信息

    <!DOCTYPE html> <html> <head> <meta charset="utf-8"> <meta http ...

  5. JavaScript ----------------- 寄生式继承

    寄生式继承 寄生式继承是于原型式继承紧密相关的一种思路.寄生式基础的思路与寄生构造函数和工厂模式类似,既创建一个仅用于封装继承过程的函数,该函数内部以某种方式来增强对象,最后再像真地是它做了所有工作一 ...

  6. App签名--- Android

    步骤: 下面就Next即可

  7. flex 调用WebService1(基于.net)

    以.net平台下C#语言开发的WebService为web服务,使用flex  actionscript语句访问webservice接口 Flex:  Temp.mxml部分代码 //调用WebSer ...

  8. Lucene学习总结之三：Lucene的索引文件格式(1)

    Lucene的索引里面存了些什么,如何存放的,也即Lucene的索引文件格式,是读懂Lucene源代码的一把钥匙. 当我们真正进入到Lucene源代码之中的时候,我们会发现: Lucene的索引过程, ...

  9. [原]此程序专用来说明C++模板的用法

    #include using namespace std; //此程序专用来说明模版的使用 template void swap1(T& a,T& b){     T temp=a; ...

  10. Cleaning Shifts（POJ 2376 贪心）

    Cleaning Shifts Time Limit: 1000MS   Memory Limit: 65536K Total Submissions: 15143   Accepted: 3875 ...