安卓车辆VIN码扫描

首页

文档扫描与图像处理

安卓车辆VIN码扫描

2024年12月18日

分类：数据捕获

VIN是车辆识别号码的英文的缩写。这是一个独特的17个字符的代码，用于识别车辆。车辆识别码包含汽车品牌、型号、年份、制造国家/地区等信息。

VIN code39

可以在驾驶员侧挡风玻璃附近或汽车底盘上找到VIN标签。

大多数VIN标签仅包含文本。有些标签还具有以下格式的条形码：Code 39、QR码、Data Matrix码。

在本文中，我们将构建一个安卓应用，使用Dynamsoft Capture Vision SDK扫描条形码或文本以读取VIN码。

演示视频：

新建项目

使用Android Studio创建一个基于Java的空的activity的应用项目。

添加依赖项

添加Dynamsoft的maven仓库。

repositories {
    maven {
        url "https://download2.dynamsoft.com/maven/aar"
    }
}

添加Dynamsoft Capture Vision SDK。

implementation 'com.dynamsoft:dynamsoftcapturevisionbundle:2.6.1001'

添加VIN的OCR模型。

implementation 'com.dynamsoft:dynamsoftvin:3.4.20'

初始化许可证

初始化许可证以使用Dynamsoft的SDK。可以在此处申请许可证。

//use one-day trial
LicenseManager.initLicense("DLS2eyJoYW5kc2hha2VDb2RlIjoiMjAwMDAxLTE2NDk4Mjk3OTI2MzUiLCJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSIsInNlc3Npb25QYXNzd29yZCI6IndTcGR6Vm05WDJrcEQ5YUoifQ==", this, (isSuccess, error) -> {
    error.printStackTrace();
});

访问摄像头

在activity_main.xml中添加以下内容，添加CameraView以显示摄像机预览：

<com.dynamsoft.dce.CameraView
    android:id="@+id/camera_view"
    android:layout_width="match_parent"
    android:layout_height="match_parent">
</com.dynamsoft.dce.CameraView>

根据ID获取CameraView。

CameraView cameraView = findViewById(R.id.camera_view);

创建一个Camera Enhancer实例来控制摄像头，并将CameraView与之绑定。

private CameraEnhancer mCamera;
mCamera = new CameraEnhancer(cameraView, this);

请求摄像头权限。

PermissionUtil.requestCameraPermission(this);

打开或关闭摄像头。
```
mCamera.open();
//mCamera.close();
```

设置扫描区域，只处理帧的一部分以提高扫描效率。

DSRect region = new DSRect(0,0.4f,1,0.6f,true);
mCamera.setScanRegion(region);

读取条形码

接下来，读取VIN标签上的条形码。

创建Capture Vision Router以调用各种图像处理任务。

private CaptureVisionRouter mRouter;
mRouter = new CaptureVisionRouter(this);

创建一个新的JSON模板文件，用来配置router以读取Code 39、QR code和Data Matrix格式的条形码。将其命名为vin_barcode_template.json，存放在 res/raw。

{
  "CaptureVisionTemplates": [
    {
      "Name": "ReadVINBarcode",
      "ImageROIProcessingNameArray": [
        "roi-read-vin-barcodes"
      ]
    }
  ],
  "TargetROIDefOptions": [
    {
      "Name": "roi-read-vin-barcodes",
      "TaskSettingNameArray": ["task-read-vin-barcodes"]
    }
  ],
  "BarcodeReaderTaskSettingOptions": [
    {
      "Name": "task-read-vin-barcodes",
      "BarcodeFormatIds": ["BF_CODE_39", "BF_QR_CODE", "BF_DATAMATRIX"],
          "LocalizationModes": [
        {
          "Mode": "LM_CONNECTED_BLOCKS"
        },
        {
          "Mode": "LM_SCAN_DIRECTLY",
          "ScanStride": 0,
          "ScanDirection": 0,
          "IsOneDStacked": 0
        },
        {
          "Mode": "LM_STATISTICS"
        },
        {
          "Mode": "LM_LINES"
        }
      ],
      "ExpectedBarcodesCount": 1
    }
  ]
}

使用模板更新Capture Vision Router的运行时设置。

@Override
protected void onCreate(Bundle savedInstanceState) {
    String template = readTemplate(R.raw.vin_barcode_template);
    mRouter.initSettings(template);
}

private String readTemplate(int ID) throws IOException {
    InputStream inp = this.getResources().openRawResource(ID);
    BufferedReader reader = new BufferedReader(new InputStreamReader(inp));
    StringBuilder out = new StringBuilder();
    String line;
    while ((line = reader.readLine()) != null) {
        out.append(line);
    }
    String content = out.toString();
    reader.close();
    return content;
}

将Camera Enhancer设置为Capture Vision Router的图像源。
```
mRouter.setInput(mCamera);
```

添加结果接收器以接收扫描到的条码结果。

mRouter.addResultReceiver(new CapturedResultReceiver() {
    @Override
    public void onCapturedResultReceived(@NonNull CapturedResult result) {
        String barcode = "";
        for (CapturedResultItem item:result.getItems()) {
            if (item.getType() == EnumCapturedResultItemType.CRIT_BARCODE) {
                barcode = ((BarcodeResultItem) item).getText();
            }
        }
    }
});

从摄像头获取图像并检测条形码。
```
mRouter.startCapturing("ReadVINBarcode");
```

识别文本

接下来，识别VIN标签上的文本。

创建一个新的JSON模板文件，用来配置router以读取17个字符的文本。将其命名为vin_text_template.json，存放在 res/raw。我们需要指定VIN模型和图像处理参数。例如，VIN的文本通常是反色的（黑底白字），我们需要配置GrayscaleTransformationModes来反转颜色。

{
  "CaptureVisionTemplates": [
    {
      "Name": "ReadVINText",
      "ImageROIProcessingNameArray": [
        "roi-read-vin-text"
      ],
      "ImageSource": "",
      "MaxParallelTasks": 4,
      "MinImageCaptureInterval": 0,
      "OutputOriginalImage": 0,
      "Timeout": 10000
    }
  ],
  "TargetROIDefOptions": [
    {
      "Name": "roi-read-vin-text",
      "TaskSettingNameArray": ["task-read-vin-text"]
    }
  ],
  "CharacterModelOptions": [
    {
      "CharSet": {
        "ExcludeChars": ["O", "Q", "I"]
      },
      "DirectoryPath": "",
      "Name": "VIN"
    }
  ],
  "ImageParameterOptions": [
    {
      "BaseImageParameterName": "",
      "BinarizationModes": [
        {
          "BinarizationThreshold": -1,
          "BlockSizeX": 0,
          "BlockSizeY": 0,
          "EnableFillBinaryVacancy": 1,
          "GrayscaleEnhancementModesIndex": -1,
          "Mode": "BM_LOCAL_BLOCK",
          "MorphOperation": "Close",
          "MorphOperationKernelSizeX": -1,
          "MorphOperationKernelSizeY": -1,
          "MorphShape": "Rectangle",
          "ThresholdCompensation": 10
        }
      ],
      "ColourConversionModes": [
        {
          "BlueChannelWeight": -1,
          "GreenChannelWeight": -1,
          "Mode": "CICM_GENERAL",
          "RedChannelWeight": -1,
          "ReferChannel": "H_CHANNEL"
        }
      ],
      "GrayscaleEnhancementModes": [
        {
          "Mode": "GEM_GENERAL",
          "Sensitivity": -1,
          "SharpenBlockSizeX": -1,
          "SharpenBlockSizeY": -1,
          "SmoothBlockSizeX": -1,
          "SmoothBlockSizeY": -1
        }
      ],
      "GrayscaleTransformationModes": [
        {
          "Mode": "GTM_ORIGINAL"
        },
        {
          "Mode": "GTM_INVERTED"
        }
      ],
      "IfEraseTextZone": 0,
      "Name": "ip_recognize_text",
      "RegionPredetectionModes": [
        {
          "AspectRatioRange": "[]",
          "FindAccurateBoundary": 0,
          "ForeAndBackgroundColours": "[]",
          "HeightRange": "[]",
          "ImageParameterName": "",
          "MeasuredByPercentage": 1,
          "MinImageDimension": 262144,
          "Mode": "RPM_GENERAL",
          "RelativeRegions": "[]",
          "Sensitivity": 1,
          "SpatialIndexBlockSize": 5,
          "WidthRange": "[]"
        }
      ],
      "ScaleDownThreshold": 2300,
      "ScaleUpModes": [
        {
          "AcuteAngleWithXThreshold": -1,
          "LetterHeightThreshold": 0,
          "Mode": "SUM_AUTO",
          "ModuleSizeThreshold": 0,
          "TargetLetterHeight": 0,
          "TargetModuleSize": 0
        }
      ],
      "TextDetectionMode": {
        "CharHeightRange": [5, 1000, 1],
        "Direction": "HORIZONTAL",
        "MaxSpacingInALine": -1,
        "Mode": "TTDM_LINE",
        "Sensitivity": 7,
        "StringLengthRange": null
      },
      "TextureDetectionModes": [
        {
          "Mode": "TDM_GENERAL_WIDTH_CONCENTRATION",
          "Sensitivity": 5
        }
      ]
    }
  ],
  "LabelRecognizerTaskSettingOptions": [
    {
      "Name": "task-read-vin-text",
      "TextLineSpecificationNameArray": ["tls_vin_text"],
      "SectionImageParameterArray": [
        {
          "ContinueWhenPartialResultsGenerated": 1,
          "ImageParameterName": "ip_recognize_text",
          "Section": "ST_REGION_PREDETECTION"
        },
        {
          "ContinueWhenPartialResultsGenerated": 1,
          "ImageParameterName": "ip_recognize_text",
          "Section": "ST_TEXT_LINE_LOCALIZATION"
        },
        {
          "ContinueWhenPartialResultsGenerated": 1,
          "ImageParameterName": "ip_recognize_text",
          "Section": "ST_TEXT_LINE_RECOGNITION"
        }
      ]
    }
  ],
  "TextLineSpecificationOptions": [
    {
      "BinarizationModes": [
        {
          "BinarizationThreshold": -1,
          "BlockSizeX": 11,
          "BlockSizeY": 11,
          "EnableFillBinaryVacancy": 1,
          "GrayscaleEnhancementModesIndex": -1,
          "Mode": "BM_LOCAL_BLOCK",
          "MorphOperation": "Erode",
          "MorphOperationKernelSizeX": -1,
          "MorphOperationKernelSizeY": -1,
          "MorphShape": "Rectangle",
          "ThresholdCompensation": 10
        }
      ],
      "CharHeightRange": [5, 1000, 1],
      "CharacterModelName": "VIN",
      "CharacterNormalizationModes": [
        {
          "Mode": "CNM_AUTO",
          "MorphArgument": "3",
          "MorphOperation": "Close"
        }
      ],
      "ConcatResults": 0,
      "ConcatSeparator": "\n",
      "ConcatStringLengthRange": [3, 200],
      "ExpectedGroupsCount": 1,
      "GrayscaleEnhancementModes": [
        {
          "Mode": "GEM_GENERAL",
          "Sensitivity": -1,
          "SharpenBlockSizeX": -1,
          "SharpenBlockSizeY": -1,
          "SmoothBlockSizeX": -1,
          "SmoothBlockSizeY": -1
        }
      ],
      "Name": "tls_vin_text",
      "OutputResults": 1,
      "StringLengthRange": [17, 17],
      "StringRegExPattern": "[0-9A-HJ-NPR-Z]{9}[1-9A-HJ-NPR-TV-Y][0-9A-HJ-NPR-Z]{2}[0-9]{5}",
      "SubGroups": null,
      "TextLinesCount": 1
    }
  ]
}

在结果接收器中获取文本结果。

mRouter.addResultReceiver(new CapturedResultReceiver() {
    @Override
    public void onCapturedResultReceived(@NonNull CapturedResult result) {
        String barcode = "";
        String textLine = "";
        for (CapturedResultItem item:result.getItems()) {
            if (item.getType() == EnumCapturedResultItemType.CRIT_BARCODE) {
                barcode = ((BarcodeResultItem) item).getText();
            }
            if (item.getType() == EnumCapturedResultItemType.CRIT_TEXT_LINE) {
                textLine = ((TextLineResultItem) item).getText();
            }
        }
    }
});

加载模板并开始识别文本。

String template = readTemplate(R.raw.vin_text_template);
mRouter.initSettings(template);
mRouter.startCapturing("ReadVINText");

同时识别条形码和文本

我们可以合并这两个JSON模板，并创建一个新的CaptureVisionTemplate对象，以同时运行条形码和文本检测。

{
  "CaptureVisionTemplates": [
    {
      "Name": "ReadVINBarcodeAndText",
      "ImageROIProcessingNameArray": [
        "roi-read-vin-barcodes","roi-read-vin-text"
      ]
    }
  ]
}

然后根据它的名字，通过router来使用它。

mRouter.startCapturing("ReadVINBarcodeAndText");

过滤误读

在实时扫描场景中，可能会发生误读。我们可以启用结果过滤器来过滤误读。它的工作原理是检查多次读取结果。

MultiFrameResultCrossFilter filter = new MultiFrameResultCrossFilter();
filter.enableResultCrossVerification(EnumCapturedResultItemType.CRIT_TEXT_LINE, true);
filter.enableResultCrossVerification(EnumCapturedResultItemType.CRIT_BARCODE, true);
mRouter.addResultFilter(filter);

处理生命周期

当应用程序置于后台时停止扫描，当应用程序置于前台时恢复扫描。

@Override
public void onResume() {
    super.onResume();
    startScanning();
}

@Override
public void onPause() {
    super.onPause();
    stopScanning();
}

源代码

查看demo的源代码并尝试使用：https://github.com/tony-xlh/Android-VIN-Scanner