Chinaunix首页 | 论坛 | 博客
  • 博客访问: 1071506
  • 博文数量: 403
  • 博客积分: 10272
  • 博客等级: 上将
  • 技术积分: 4407
  • 用 户 组: 普通用户
  • 注册时间: 2012-02-24 14:22
文章分类

全部博文(403)

文章存档

2012年(403)

分类: 嵌入式

2012-03-21 17:03:43

android应用中ocr的解决方案大致有两种,而采用最多的还是tesseract.小弟就在这里将我最近两天解决思路写下来,如有缺陷,欢迎拍砖:

  有两种解决方案,一种是采用tesseract cloud-service,这钟是把图片信息发送到云端,然后获得图片分析数据;第二种就是不用联网,本地化分析图片上信息。我就说说第二种,第一种我会在最后给大家一个链接(文章很不错)。

  搜先就是下载Tesseract native android library.这里有两个链接,你选哪个链接都可以:

  a.svn checkout tesseract-android-tools。(如果不能checkout到,废话别说就到官方上下:)

  b.可能上面一个下载后编译有些人会遇到一些问题,比如找不到jgep库,编译不成功。所以有了这个项目:git clone git://github.com/rmtheis/tess-two.git (这个包里面内容太多,不过也省得下那么多库了)

  这里先说采用第一个源下载:下载成功后,打开README文件,做下修改(如下):

git clone git://android.git.kernel.org/platform/external/jpeg.git libjpeg修改为:git clone https://android.googlesource.com/platform/external/jpeg libjpegndk-build //这个编译要到jni文件夹里面编译

  

  对于第二个源下载,由于里面没有README文件,操作命令如下:

cd /tess-two
export TESSERACT_PATH=${PWD}/external/tesseract-3.01
export LEPTONICA_PATH=${PWD}/external/leptonica-1.68
export LIBJPEG_PATH=${PWD}/external/libjpeg
ndk-build
android update project --path .
ant release

  最终两个都得到你想要的libs里面的so文件和src里面的对so文件的封装类。这个就是我们开发所用到的东东啦。

  然后新建工程,代码如下:

public class MainActivity extends Activity {
private static final String TAG = "MainActivity ...";

private static final String TESSBASE_PATH = "/mnt/sdcard/tesseract/";
private static final String DEFAULT_LANGUAGE = "eng";
private static final String IMAGE_PATH = "/mnt/sdcard/test1.jpg";
private static final String EXPECTED_FILE = TESSBASE_PATH + "tessdata/" + DEFAULT_LANGUAGE
+ ".traineddata";

private TessBaseAPI service;
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.main);
testOcr();

}

public void testOcr(){
mHandler.post(new Runnable() {

@Override
public void run() {
Log.d(TAG, "begin>>>>>>>");
ocr();
//test();
}
});

}
public void test(){
// First, make sure the eng.traineddata file exists.
/*assertTrue("Make sure that you've copied " + DEFAULT_LANGUAGE + ".traineddata to "
+ EXPECTED_FILE, new File(EXPECTED_FILE).exists());
*/
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
final Bitmap bmp = BitmapFactory.decodeResource(getResources(), R.drawable.test);
//digits is a .jpg image I found in one of the issues here.
ImageView img = (ImageView) findViewById(R.id.image);
img.setImageBitmap(bmp);//I can see the ImageView. So we know that it should work if I sent it to the setImage()
baseApi.setImage(bmp);
Log.v("Kishore","Kishore:Working");//This statement is never reached. Futhermore, on putting some more Log.v commands in the setImage function, I found out that the native function nativeSetImagePix is never accessed. I have attached the Logcat output below to show that it is not accessed.

String outputText = baseApi.getUTF8Text();
Log.v("Kishore","Kishore:"+outputText);
baseApi.end();
bmp.recycle();
}

protected void ocr() {

BitmapFactory.Options options = new BitmapFactory.Options();
options.inSampleSize = 2;
Bitmap bitmap = BitmapFactory.decodeFile(IMAGE_PATH, options);

try {
ExifInterface exif = new ExifInterface(IMAGE_PATH);
int exifOrientation = exif.getAttributeInt(ExifInterface.TAG_ORIENTATION, ExifInterface.ORIENTATION_NORMAL);

Log.v(TAG, "Orient: " + exifOrientation);

int rotate = 0;
switch (exifOrientation) {
case ExifInterface.ORIENTATION_ROTATE_90:
rotate = 90;
break;
case ExifInterface.ORIENTATION_ROTATE_180:
rotate = 180;
break;
case ExifInterface.ORIENTATION_ROTATE_270:
rotate = 270;
break;
}

Log.v(TAG, "Rotation: " + rotate);

if (rotate != 0) {

// Getting width & height of the given image.
int w = bitmap.getWidth();
int h = bitmap.getHeight();

// Setting pre rotate
Matrix mtx = new Matrix();
mtx.preRotate(rotate);

// Rotating Bitmap
bitmap = Bitmap.createBitmap(bitmap, 0, 0, w, h, mtx, false);
// tesseract req. ARGB_8888
bitmap = bitmap.copy(Bitmap.Config.ARGB_8888, true);
}

} catch (IOException e) {
Log.e(TAG, "Rotate or coversion failed: " + e.toString());
}

ImageView iv = (ImageView) findViewById(R.id.image);
iv.setImageBitmap(bitmap);
iv.setVisibility(View.VISIBLE);

Log.v(TAG, "Before baseApi");

TessBaseAPI baseApi = new TessBaseAPI();
baseApi.setDebug(true);
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setImage(bitmap);
String recognizedText = baseApi.getUTF8Text();
baseApi.end();

Log.v(TAG, "OCR Result: " + recognizedText);

// clean up and show
if (DEFAULT_LANGUAGE.equalsIgnoreCase("eng")) {
recognizedText = recognizedText.replaceAll("[^a-zA-Z0-9]+", " ");
}
if (recognizedText.length() != 0) {
((TextView) findViewById(R.id.field)).setText(recognizedText.trim());
}
}
private Handler mHandler = new Handler(){
public void handleMessage(android.os.Message msg) {

};
};
}


  当你很欢喜的运行程序的时候,发现事情没有你想象的那么简单。这个文件必须要用到一个语言包。不然你怎么匹配呢?想想也是:

adb shell mkdir /mnt/sdcard/tesseract
adb shell mkdir /mnt/sdcard/tesseract/tessdata
adb push eng.traineddata /mnt/sdcard/tesseract/tessdata/eng.traineddata
adb shell ls -l /mnt/sdcard/tesseract/tessdata
ls -l bin/tesseract-android-tools-test.apk
adb install -r -s bin/tesseract-android-tools-test.apk
adb shell am instrument -w -e class com.googlecode.tesseract.android.test.TessBaseAPITest com.googlecode.tesseract.android.test/android.test.InstrumentationTestRunner

  上面的额eng.traineddata这个东西。你可以搜下,网络有的。(囧,我还不知到怎么上传附件)

  最后效果如图(事实上解析结果是:44m><9。只不过那个字符不认识吧):

阅读(4654) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~