Question

我制作了一个简单的应用程序，用于读取图像并使用android检索数字图像作为文本。但问题是精度仅为60％左右，并且也会出现一些不需要的噪音。我确实认为准确度不可能达到100％，但是，我认为必须有一种方法来改进它。但是，既然我是业余爱好者，我觉得很难。我搜索了谷歌，但无法获得可靠的信息。

我想从东方幸运票中读取数字596,00和012345，如下图所示。

enter image description here

Answer 1

Tesseract-ocr最适合符合以下条件的角色图像：

输入图像应至少具有300 dpi
输入图像应为黑白
输入图像中应该有最小的噪音（即文字应该与背景清晰可辨）
文字行应该是直的
图像应以待检测文本为中心

(See the tesseract-ocr wiki for further details)

对于给定的输入图像，tesseract将尝试预处理和清洁图像以满足这些标准，但为了最大限度地提高检测准确度，最好自己进行预处理。

根据您提供的输入图像，主要问题是背景噪音太大。为了消除图像中文本的背景噪声，我发现应用具有阈值的笔画宽度变换（SWT）算法来消除噪声会产生有希望的结果。在libCCV library.中提供了具有许多可配置参数的SWT的快速实现。它清理图像的程度取决于许多因素，包括图像大小，笔划宽度的均匀性以及算法的其他输入参数。 A list of the configurable parameters is provided here.

然后将SWT的输出传递给tesseract以获取图像中字符的文本值。

如果传递给tesseract的图像仍然包含一些噪音，则可能会返回一些错误的检测，例如标点字符。鉴于您正在处理的图像可能只包含字母和数字a-z A-Z 0-9，您只需将正则表达式应用于输出即可删除任何最终的错误检测。

Answer 2

import { Component } from '@angular/core';
import { NavController } from 'ionic-angular';
import { StreamingMedia, StreamingAudioOptions } from '@ionic-native/streaming-media';
@Component({
  selector: 'page-home',
  templateUrl: 'home.html'
})
export class HomePage {
constructor(public navCtrl: NavController,private streamingMedia: StreamingMedia) {}

      startAudio(){
        let options: StreamingAudioOptions = {
          successCallback: () => { console.log('Video played') },
          errorCallback: (e) => { console.log(JSON.stringify(e,null,2)) },
          initFullscreen:false,
        };
          this.streamingMedia.playAudio('YOUR_URL', options);
      }

在app gradle中添加依赖项

you can use Vision for text detection.

添加Manifest.xml

compile 'com.google.android.gms:play-services-vision:10.0.0'

MainActivity.java

<meta-data
        android:name="com.google.android.gms.vision.DEPENDENCIES"
        android:value="ocr" />

activity_main.xml中

import android.app.AlertDialog;
import android.content.ContentValues;
import android.content.Intent;
import android.graphics.Bitmap;
import android.graphics.BitmapFactory;
import android.net.Uri;
import android.provider.MediaStore;
import android.support.v7.app.AppCompatActivity;
import android.os.Bundle;
import android.text.method.ScrollingMovementMethod;
import android.util.DisplayMetrics;
import android.util.Log;
import android.util.SparseArray;
import android.view.View;
import android.widget.TextView;

import com.google.android.gms.vision.Frame;
import com.google.android.gms.vision.text.TextBlock;
import com.google.android.gms.vision.text.TextRecognizer;

import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.InputStream;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;
import java.util.List;

public class MainActivity extends AppCompatActivity {
    private static final int REQUEST_GALLERY = 0;
    private static final int REQUEST_CAMERA = 1;

    private static final String TAG = MainActivity.class.getSimpleName();

    private Uri imageUri;
    private TextView detectedTextView;

    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        setContentView(R.layout.activity_main);

        findViewById(R.id.choose_from_gallery).setOnClickListener(new View.OnClickListener() {
            @Override
            public void onClick(View v) {
                Intent intent = new Intent();
                intent.setType("image/*");
                intent.setAction(Intent.ACTION_GET_CONTENT);
                startActivityForResult(intent, REQUEST_GALLERY);
            }
        });

        findViewById(R.id.take_a_photo).setOnClickListener(new View.OnClickListener() {
            @Override
            public void onClick(View v) {
                String filename = System.currentTimeMillis() + ".jpg";

                ContentValues values = new ContentValues();
                values.put(MediaStore.Images.Media.TITLE, filename);
                values.put(MediaStore.Images.Media.MIME_TYPE, "image/jpeg");
                imageUri = getContentResolver().insert(MediaStore.Images.Media.EXTERNAL_CONTENT_URI, values);

                Intent intent = new Intent();
                intent.setAction(MediaStore.ACTION_IMAGE_CAPTURE);
                intent.putExtra(MediaStore.EXTRA_OUTPUT, imageUri);
                startActivityForResult(intent, REQUEST_CAMERA);
            }
        });

        detectedTextView = (TextView) findViewById(R.id.detected_text);
        detectedTextView.setMovementMethod(new ScrollingMovementMethod());
    }

    private void inspectFromBitmap(Bitmap bitmap) {
        TextRecognizer textRecognizer = new TextRecognizer.Builder(this).build();
        try {
            if (!textRecognizer.isOperational()) {
                new AlertDialog.
                        Builder(this).
                        setMessage("Text recognizer could not be set up on your device").show();
                return;
            }

            Frame frame = new Frame.Builder().setBitmap(bitmap).build();
            SparseArray<TextBlock> origTextBlocks = textRecognizer.detect(frame);
            List<TextBlock> textBlocks = new ArrayList<>();
            for (int i = 0; i < origTextBlocks.size(); i++) {
                TextBlock textBlock = origTextBlocks.valueAt(i);
                textBlocks.add(textBlock);
            }
            Collections.sort(textBlocks, new Comparator<TextBlock>() {
                @Override
                public int compare(TextBlock o1, TextBlock o2) {
                    int diffOfTops = o1.getBoundingBox().top - o2.getBoundingBox().top;
                    int diffOfLefts = o1.getBoundingBox().left - o2.getBoundingBox().left;
                    if (diffOfTops != 0) {
                        return diffOfTops;
                    }
                    return diffOfLefts;
                }
            });

            StringBuilder detectedText = new StringBuilder();
            for (TextBlock textBlock : textBlocks) {
                if (textBlock != null && textBlock.getValue() != null) {
                    detectedText.append(textBlock.getValue());
                    detectedText.append("\n");
                }
            }

            detectedTextView.setText(detectedText);
        }
        finally {
            textRecognizer.release();
        }
    }

    private void inspect(Uri uri) {
        InputStream is = null;
        Bitmap bitmap = null;
        try {
            is = getContentResolver().openInputStream(uri);
            BitmapFactory.Options options = new BitmapFactory.Options();
            options.inPreferredConfig = Bitmap.Config.ARGB_8888;
            options.inSampleSize = 2;
            options.inScreenDensity = DisplayMetrics.DENSITY_LOW;
            bitmap = BitmapFactory.decodeStream(is, null, options);
            inspectFromBitmap(bitmap);
        } catch (FileNotFoundException e) {
            Log.w(TAG, "Failed to find the file: " + uri, e);
        } finally {
            if (bitmap != null) {
                bitmap.recycle();
            }
            if (is != null) {
                try {
                    is.close();
                } catch (IOException e) {
                    Log.w(TAG, "Failed to close InputStream", e);
                }
            }
        }
    }

    @Override
    protected void onActivityResult(int requestCode, int resultCode, Intent data) {
        switch (requestCode) {
            case REQUEST_GALLERY:
                if (resultCode == RESULT_OK) {
                    inspect(data.getData());
                }
                break;
            case REQUEST_CAMERA:
                if (resultCode == RESULT_OK) {
                    if (imageUri != null) {
                        inspect(imageUri);
                    }
                }
                break;
            default:
                super.onActivityResult(requestCode, resultCode, data);
                break;
        }
    }
}

Android：如何改善tesseract ocr检索到的图像中的数字？

2 个答案: