Tesseract在Android上的使用

上一篇文章介绍了Tesseract在Mac平台上的使用，现在看看怎么实现在Android上进行OCR识别。这里基于个人之前自己做的一个APP上集成，首先，修改app/build.gradle，增加对tess-two的支持，见76行：

把上一篇文章最后训练过的normal.trainneddata放到assert文件夹下:

编写测试界面如下:

编写页面对应java文件：

package com.linjk.ihouse.activity;import android.app.ActionBar;import android.app.Activity;import android.graphics.Bitmap;import android.graphics.BitmapFactory;import android.os.Bundle;import android.os.Environment;import android.util.Log;import android.view.MenuItem;import android.widget.TextView;import com.googlecode.tesseract.android.TessBaseAPI;import com.linjk.ihouse.R;import java.io.File;import java.io.FileOutputStream;import java.io.InputStream;import java.io.OutputStream;import butterknife.Bind;import butterknife.ButterKnife;/** * Created by LinJK on 07/06/2017. */public class ActivityTesseractOCR extends Activity {    @Bind(R.id.ocr_result)     TextView tvOCRResult;    @Override    protected void onCreate(Bundle savedInstanceState) {        super.onCreate(savedInstanceState);        setContentView(R.layout.activity_tesseract_ocr);        setTitle("TesseractOCR");        ButterKnife.bind(this);        ActionBar actionBar = getActionBar();        actionBar.setDisplayHomeAsUpEnabled(true);        try {            copyLanguagePackageToSDCard();            TessBaseAPI lvBaseAPI = new TessBaseAPI();            lvBaseAPI.init(Environment.getExternalStorageDirectory().getPath(), "normal");            lvBaseAPI.setPageSegMode(TessBaseAPI.PageSegMode.PSM_AUTO);            Bitmap lvBitmap = BitmapFactory.decodeResource(getResources(), R.drawable.name);            lvBaseAPI.setImage(lvBitmap);            String result = lvBaseAPI.getUTF8Text();            tvOCRResult.setText(result);            lvBaseAPI.clear();            lvBaseAPI.end();        }        catch (Exception e) {            Log.e("OCR", e.getMessage());        }    }    // 把字库文件拷贝到SD卡，要求SD卡根目录有tessdata文件夹，字库在该文件夹下    private void copyLanguagePackageToSDCard() {        try {            String filePath = Environment.getExternalStorageDirectory().getPath() + "/tessdata/normal.traineddata";            File lvFile = new File(filePath);            if (!lvFile.exists()) {                Log.i("file", "创建路径");                if (lvFile.mkdir()) {                    Log.i("file", "创建成功");                }                else {                    Log.i("file", "创建失败");                }            }            InputStream lvInputStream;            OutputStream lvOutputStream = new FileOutputStream(filePath);            // 拷贝文件            lvInputStream = this.getAssets().open("normal.traineddata");            byte[] buffer = new byte[1024];            int length = lvInputStream.read(buffer);            while (length > 0) {                lvOutputStream.write(buffer, 0, length);                length = lvInputStream.read(buffer);            }            lvOutputStream.flush();            lvInputStream.close();            lvOutputStream.close();        }        catch (Exception e) {            Log.e("OCR<copy>", e.getMessage());        }    }    @Override    public boolean onOptionsItemSelected(MenuItem item) {        switch (item.getItemId()){            case android.R.id.home:                finish();                return true;            default:                return super.onOptionsItemSelected(item);        }    }}

代码那里首先把assets文件的字库拷贝到sd卡根目录下的tessdata文件夹下，然后初始化OCR引擎，接着读取照片进行解析。这里还有很多可以改进的，识别应放到Handler来处理，这样就不会阻塞Ui线程，这只是测试功能而已就这么写了。