如何用MediaCapture解决二维码扫描问题

Posted 2020-10-11

原文:如何用MediaCapture解决二维码扫描问题

二维码扫描的实现，简单的来说可以分三步走：“成像”、“截图”与“识别”。

UWP开发中，最常用的媒体工具非MediaCapture莫属了，下面就来简单介绍一下如何利用MediaCapture来实现扫描和截图并且利用Zxing识别二维码，以及会遇到的问题和需要注意的地方。

1. 初始化与成像

 1 private async void InitMediaCaptureAsync()
 2         {
 3             //寻找后置摄像头
 4             var allVideoDevices = await DeviceInformation.FindAllAsync(DeviceClass.VideoCapture);
 5             var cameraDevice = allVideoDevices.FirstOrDefault(x => x.EnclosureLocation != null && x.EnclosureLocation.Panel == Windows.Devices.Enumeration.Panel.Back);
 6 
 7             if (cameraDevice == null)
 8             {
 9                 Debug.WriteLine("No camera device found!");
10 
11                 return;
12             }
13 
14             var settings = new MediaCaptureInitializationSettings
15             {
16                 StreamingCaptureMode = StreamingCaptureMode.Video,
17                 //必须，否则截图的时候会很卡很慢
18                 PhotoCaptureSource = PhotoCaptureSource.VideoPreview,
19                 VideoDeviceId = cameraDevice.Id
20             };
21 
22             _mediaCapture = new MediaCapture();
23 
24             try
25             {
26                 await _mediaCapture.InitializeAsync(settings);
27                 _initialized = true;//初始化成功
28             }
29             catch (UnauthorizedAccessException)
30             {
31                 Debug.WriteLine("The app was denied access to the camera");
32             }
33             catch (Exception ex)
34             {
35                 Debug.WriteLine("Exception when initializing MediaCapture with {0}: {1}", cameraDevice.Id, ex.ToString());
36             }
37 
38             if (_initialized)
39             {
40                 var focusControl = _mediaCapture.VideoDeviceController.FocusControl;
41 
42                 if (focusControl.Supported)
43                 {
44                     var focusSettings = new FocusSettings()
45                     {
46                         Mode = focusControl.SupportedFocusModes.FirstOrDefault(f => f == FocusMode.Continuous),
47                         DisableDriverFallback = true,
48                         AutoFocusRange = focusControl.SupportedFocusRanges.FirstOrDefault(f => f == AutoFocusRange.FullRange),
49                         Distance = focusControl.SupportedFocusDistances.FirstOrDefault(f => f == ManualFocusDistance.Nearest)
50                     };
51 
52                     //设置聚焦，最好使用FocusMode.Continuous，否则影响截图会很模糊，不利于识别
53                     focusControl.Configure(focusSettings);
54                 }
55 
56                 captureElement.Source = _mediaCapture;
57                 captureElement.FlowDirection = FlowDirection.LeftToRight;
58 
59                 try
60                 {
61                     await _mediaCapture.StartPreviewAsync();
62                     _previewing = true;
63                 }
64                 catch (Exception ex)
65                 {
66                     Debug.WriteLine("Exception when starting the preview: {0}", ex.ToString());
67                 }
68 
69                 if (_previewing)
70                 {
71                     try
72                     {
73                         if (_mediaCapture.VideoDeviceController.FlashControl.Supported)
74                         {
75                             //关闭闪光灯
76                             _mediaCapture.VideoDeviceController.FlashControl.Enabled = false;
77                         }
78                     }
79                     catch
80                     {
81                     }
82 
83                     if (focusControl.Supported)
84                     {
85                         //开始聚焦
86                         await focusControl.FocusAsync();
87                     }
88                 }
89             }
90         }

View Code

2. 截图与识别

1 private void InitTimer()
2         {
3             _timer = new DispatcherTimer();
4             //每50毫秒截一次图
5             _timer.Interval = TimeSpan.FromMilliseconds(50);
6             _timer.Tick += _timer_Tick;
7             _timer.Start();
8         }

View Code

 1 private async void _timer_Tick(object sender, object e)
 2         {
 3             using (var stream = new InMemoryRandomAccessStream())
 4             {
 5                 var previewProperties = _mediaCapture.VideoDeviceController.GetMediaStreamProperties(MediaStreamType.VideoPreview) as VideoEncodingProperties;
 6                 //将截图写入内存流中
 7                 await _mediaCapture.CapturePhotoToStreamAsync(ImageEncodingProperties.CreateJpeg(), stream);
 8 
 9                 //利用Zxing识别，成功：停止timer；失败：继续
10                 var reader = new BarcodeReader();
11                 var bitmapWriteable = new WriteableBitmap((int)previewProperties.Width, (int)previewProperties.Height);
12                 bitmapWriteable.SetSource(stream);
13                 var result = reader.Decode(bitmapWriteable);
14 
15                 if (!string.IsNullOrEmpty(result.Text))
16                 {
17                     _timer.Stop();
18                 }
19             }
20         }

View Code

这里顺便说一下如何安装Zxing，打开nuget管理器 命令窗口输入 Install-Package ZXing.Net ，回车; 关于Zxing如何使用，到网上搜索一下有很多教程，这里不再赘述

3. 问题与优化

A） 截图有响声

使用CapturePhotoToStreamAsync来截取图片有的时候会有“咔擦咔擦”声，很影响用户体验，最理想的做法是找到一种能从视频流中直接截取图片的方法，在这里不得不说一句MediaCapture真的真的很强大，MediaCapture给我们提供了直接从视频流中取出其中一帧的方法GetPreviewFrameAsync，于是我把代码进行了如下修改，即流畅又没有烦人的“咔擦咔擦”声

 1 private async void _timer_Tick(object sender, object e)
 2         {
 3             var previewProperties = _mediaCapture.VideoDeviceController.GetMediaStreamProperties(MediaStreamType.VideoPreview) as VideoEncodingProperties;
 4 
 5             using (var videoFrame = new VideoFrame(BitmapPixelFormat.Rgba8, (int)previewProperties.Width, (int)previewProperties.Height))
 6             {
 7                 using (var currentFrame = await _mediaCapture.GetPreviewFrameAsync(videoFrame))
 8                 {
 9                     using (var previewFrame = currentFrame.SoftwareBitmap)
10                     {
11                         var buffer = new Windows.Storage.Streams.Buffer((uint)(4 * previewFrame.PixelWidth * previewFrame.PixelHeight));
12                         previewFrame.CopyToBuffer(buffer);
13 
14                         using (var stream = buffer.AsStream().AsRandomAccessStream())
15                         {
16                             //利用Zxing识别，成功：停止timer；失败：继续
17                             var reader = new BarcodeReader();
18                             var bitmapWriteable = new WriteableBitmap((int)previewProperties.Width, (int)previewProperties.Height);
19                             bitmapWriteable.SetSource(stream);
20                             var result = reader.Decode(bitmapWriteable);
21 
22                             if (!string.IsNullOrEmpty(result.Text))
23                             {
24                                 _timer.Stop();
25                             }
26                         }
27                     }
28                 }
29             }            
30         }

View Code

顺便提一下记得要使用如下两个命名空间

using System.IO;
using System.Runtime.InteropServices.WindowsRuntime;

否则无法实现buffer.AsStream().AsRandomAccessStream()

B） 连续聚焦

并不是所有机型都支持连续聚焦的（FocusMode.Continuous），这个时候只能自己实现间断性持续聚焦了

C） 截图之后图片处理

有的时候为了实现一些功能（比如说扫描框）或者提高识别率，我们需要对截取出来的图片进行一些二次处理，或剪裁或缩放或旋转，我们可以使用BitmapDecoder和BitmapEncoder来实现

using (var stream = buffer.AsStream().AsRandomAccessStream())
{
      var decoder = await BitmapDecoder.CreateAsync(BitmapDecoder.JpegDecoderId, stream);
      var destStream = new InMemoryRandomAccessStream();

      var encoder = await BitmapEncoder.CreateForTranscodingAsync(destStream, decoder);
                            
      //剪裁
      encoder.BitmapTransform.Bounds = new BitmapBounds() { X = 0, Y = 0, Width = 100, Height = 100 };
      //缩放
      encoder.BitmapTransform.ScaledWidth = 100;
      encoder.BitmapTransform.ScaledHeight = 100;
      //旋转
      encoder.BitmapTransform.Rotation = BitmapRotation.Clockwise90Degrees;

      await encoder.FlushAsync();
      await destStream.FlushAsync();
}

D） 挂起和唤醒

另外值得注意的是，程序在Suspending和Resuming还有Activated时出现的一系列状态转换，这时候很容易引起bug，需要处理好避免crash。

4. 最后

识别出来的字符串处理一般也就超链接和普通文本两种，当然也可以增加条码扫描功能，识别出的是编码，不管怎样，大家可以根据项目具体需求做相应的处理。