Small ape search, homework to help similar effects. It is implemented based on Google tesseract-ocr. Since it is developed based on C++, it cannot be directly used in Android. Therefore, this project uses Tess -two as a branch of Android.
The preparatory work
Android Studio gradle dependencies
// The compiled SO library and jar package
compile 'com. Rmtheis: Tess - two: 6.1.1'
// Image clipping
compile 'com. Edmodo: cropper: 1.0.1'Copy the codeIn fact, you can also download github.com/rmtheis/tes… The source code is compiled in a Linux environment.
2) The second way to compile the source code (!!)
- Install the VMware VM
- Install Ubuntu for Linux
- Install necessary Tools
sudo apt-get update
sudo apt-get install gitCopy the code- Configure JDK, NDK, SDK environment (step 10,000 holes)
- Download the Tess – Two code
git clone https://github.com/rmtheis/tess-two.git tessCopy the code- Begin to compile
cd /tess/tess-two/jni
ndk-buildCopy the code- Copy the compiled Tess – Two directory to your project’s libraries and associate the app with the project’s libraries.
Tess – two use
-
Download the data dictionary, create an Assets directory in app/main, and go to github.com/tesseract-o… Download the dictionary library for the corresponding language.
-
Mainactivity. Java basically writes dictionary files from assets to the phone’s file directory
/** * Copy files from assets to **@param path
*/
public void deepFile(String path) {
String newPath = getExternalFilesDir(null) + "/";
try {
String str[] = getAssets().list(path);
if (str.length > 0) {// If it is a directory
File file = new File(newPath + path);
file.mkdirs();
for (String string : str) {
path = path + "/" + string;
deepFile(path);
path = path.substring(0, path.lastIndexOf('/'));// Return to the original path}}else {// If it is a file
InputStream is = getAssets().open(path);
FileOutputStream fos = new FileOutputStream(new File(newPath + path));
byte[] buffer = new byte[1024];
int count = 0;
while (true) {
count++;
int len = is.read(buffer);
if (len == -1) {
break;
}
fos.write(buffer, 0, len); } is.close(); fos.close(); }}catch(IOException e) { e.printStackTrace(); }}Copy the codeClick to jump to the photo taking interface
public void onClick(View view) {
switch (view.getId()) {
case R.id.btn_camera:
Intent intent = new Intent(this, TakePhoteActivity.class);
startActivity(intent);
break; }}Copy the code- Takephoteactivity. Java mainly performs photo taking and cropping operations.
The following is the back interface after a successful photo taking, and the clipping interface is displayed after a successful photo taking
/** * Callback * store the picture and display the screenshot interface **@param data
*/
@Override
public void onCameraStopped(byte[] data) {
Log.i("TAG"."==onCameraStopped==");
// Create an image
Bitmap bitmap = BitmapFactory.decodeByteArray(data, 0, data.length);
// System time
long dateTaken = System.currentTimeMillis();
// Image name
String filename = DateFormat.format("yyyy-MM-dd kk.mm.ss", dateTaken).toString() + ".jpg";
// Store images (PATH directory)
Uri source = insertImage(getContentResolver(), filename, dateTaken, PATH, filename, bitmap, data);
// Prepare screenshots
try {
mCropImageView.setImageBitmap(MediaStore.Images.Media.getBitmap(this.getContentResolver(), source));
} catch (IOException e) {
Log.e(TAG, e.getMessage());
}
showCropperLayout();
}Copy the codeClipping operation, jump to the recognition interface after completion
private View.OnClickListener cropcper = new View.OnClickListener() {
@Override
public void onClick(View view) {
switch (view.getId()) {
case R.id.btn_closecropper:
showTakePhotoLayout();
break;
case R.id.btn_startcropper:
// Get the screenshot and rotate it 90 degrees
Bitmap cropperBitmap = mCropImageView.getCroppedImage();
Bitmap bitmap = Utils.rotate(cropperBitmap, -90);
// System time
long dateTaken = System.currentTimeMillis();
// Image name
String filename = DateFormat.format("yyyy-MM-dd kk.mm.ss", dateTaken).toString() + ".jpg";
Uri uri = insertImage(getContentResolver(), filename, dateTaken, PATH, filename, bitmap, null);
Intent intent = new Intent(TakePhoteActivity.this, ShowCropperedActivity.class);
intent.setData(uri);
intent.putExtra("path", PATH + filename);
intent.putExtra("width", bitmap.getWidth());
intent.putExtra("height", bitmap.getHeight());
// intent.putExtra("cropperImage", bitmap);
startActivity(intent);
bitmap.recycle();
finish();
break; }}};Copy the code- ShowCropperedActivity. Java is mainly operating
Initialize the recognizer
/ / sd card path
private static String LANGUAGE_PATH = "";
// Recognize the language
private static final String LANGUAGE = "chi_sim";//chi_sim | eng
private TessBaseAPI baseApi = new TessBaseAPI();
@Override
protected void onCreate(Bundle savedInstanceState) {
LANGUAGE_PATH = getExternalFilesDir("") + "/";
baseApi.init(LANGUAGE_PATH, LANGUAGE);
// Set the mode
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_AUTO);
}Copy the codeGray image processing to remove noise can improve accuracy
/** * grayscale processing **@param bitmap3
* @return* /
public Bitmap convertGray(Bitmap bitmap3) {
colorMatrix = new ColorMatrix();
colorMatrix.setSaturation(0);
ColorMatrixColorFilter filter = new ColorMatrixColorFilter(colorMatrix);
Paint paint = new Paint();
paint.setColorFilter(filter);
Bitmap result = Bitmap.createBitmap(bitmap3.getWidth(), bitmap3.getHeight(), Bitmap.Config.ARGB_8888);
Canvas canvas = new Canvas(result);
canvas.drawBitmap(bitmap3, 0.0, paint);
return result;
}Copy the codeBegin to identify
// Pass in the image
baseApi.setImage(bitmap);
// Get the result of recognition
String result = baseApi.getUTF8Text();
// End the identification
baseApi.end();Copy the codeThe results of
Demo has been uploaded to Github. If necessary, download github.com/wangtaoT/An…