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. web项目的两个创建形式website和webapplication

    前言 在利用VS2010创建web项目的时候,会有两个选择.可以选择直接创建website网站,还可以选择使用 webapplication应用程序.刚刚接触web开发,看到这两个就疑惑了,既然是都可 ...

  2. kaggle之人脸特征识别

    Facial_Keypoints_Detection github code facial-keypoints-detection, 这是一个人脸识别任务,任务是识别人脸图片中的眼睛.鼻子.嘴的位置. ...

  3. 如何搞定前端资源服务跨域问题之nginx篇

    问题描述 1.首先让我们先看一张图 2.从图中,我们可以很清楚的看到当http请求的站点访问https的资源的时候会报出“Cross-Origin”跨域的问题.为什么会出现这样的错误,这是因为涉及到“ ...

  4. 如何在WP8模拟器中连接本地的web服务

    这个问题困扰了很久,查找答案一度找偏方向. 其实连接web服务对于wp7不是问题,因为wp7使用的网络就是本机的网络,但是到了wp8模拟器,纯粹的虚拟机,独立的设备,也就有了自己的网络连接,要当做虚拟 ...

  5. Flex 扩展combobox 实现复选功能

    前段时间,老大召唤,给个任务,研究一下flex的combobox控件 实现复选功能(flex自带的combobox控件不可以复选),小兵能力有限,哪里可以扩展呀,网上看了别人写的,发现总是有点瑕疵(关 ...

  6. css基础-背景文本

    css背景 1. background-color:#6495ed; 2. background-image:url('bgdesert.jpg'); 3. background-repeat:rep ...

  7. 优化：代码移动code motion

    代码移动code motion-一种常见的优化-这种优化是把(一种需要执行多次但计算结果不会改变)的计算移到前面-这种优化一般需要程序员自行移动代码,不能依靠编译器(编译器担心会有副作用) 看看代码就 ...

  8. HDU-简单计算器-1237

    这道题我做了一天,把中缀表达式转化为后缀表达式,但遇到了储存的问题,考虑了好久,写出后又调试,弄了一天,下面说一下中缀表达式转换后缀表达式: 算法: 中缀表达式转后缀表达式的方法: 1.遇到操作数:直 ...

  9. 添加jar

    file->project structure->'+'添加jar 在.gradle中配置

  10. PHP基础示例:用PHP+Mysql编写简易新闻管理系统

    实现目标:使用php和mysql操作函数实现一个新闻信息的发布.浏览.修改和删除操作 实现步骤: 一.创建数据库和表 1.创建数据库和表:newsdb 2.创建表格:news 字段:新闻id,标题,关 ...