问题描述

我正在尝试查找图像中文本的边界框，目前正在使用这种方法:

I am trying to find the bounding boxes of text in an image and am currently using this approach:

// calculate the local variances of the grayscale image
Mat t_mean, t_mean_2;
Mat grayF;
outImg_gray.convertTo(grayF, CV_32F);
int winSize = 35;
blur(grayF, t_mean, cv::Size(winSize,winSize));
blur(grayF.mul(grayF), t_mean_2, cv::Size(winSize,winSize));
Mat varMat = t_mean_2 - t_mean.mul(t_mean);
varMat.convertTo(varMat, CV_8U);

// threshold the high variance regions
Mat varMatRegions = varMat > 100;

当给出这样的图像时:

然后当我显示 varMatRegions 我得到这个图像:

Then when I show varMatRegions I get this image:

正如您所看到的，它在某种程度上将左侧的文本块与卡片的标题结合在一起，对于大多数卡片来说，这种方法效果很好，但在较繁忙的卡片上，它可能会导致问题.

As you can see it somewhat combines the left block of text with the header of the card, for most cards this method works great but on busier cards it can cause problems.

那些轮廓连接不好的原因是它使轮廓的边界框几乎占据了整张卡片.

The reason it is bad for those contours to connect is that it makes the bounding box of the contour nearly take up the entire card.

谁能建议我找到文本的不同方式以确保正确检测文本?

Can anyone suggest a different way I can find the text to ensure proper detection of text?

200 分给能在卡片中找到这两个上面的文字的人.

推荐答案

您可以通过查找关闭边缘元素来检测文本(灵感来自 LPD):

You can detect text by finding close edge elements (inspired from a LPD):

#include "opencv2/opencv.hpp"

std::vector<cv::Rect> detectLetters(cv::Mat img)
{
    std::vector<cv::Rect> boundRect;
    cv::Mat img_gray, img_sobel, img_threshold, element;
    cvtColor(img, img_gray, CV_BGR2GRAY);
    cv::Sobel(img_gray, img_sobel, CV_8U, 1, 0, 3, 1, 0, cv::BORDER_DEFAULT);
    cv::threshold(img_sobel, img_threshold, 0, 255, CV_THRESH_OTSU+CV_THRESH_BINARY);
    element = getStructuringElement(cv::MORPH_RECT, cv::Size(17, 3) );
    cv::morphologyEx(img_threshold, img_threshold, CV_MOP_CLOSE, element); //Does the trick
    std::vector< std::vector< cv::Point> > contours;
    cv::findContours(img_threshold, contours, 0, 1); 
    std::vector<std::vector<cv::Point> > contours_poly( contours.size() );
    for( int i = 0; i < contours.size(); i++ )
        if (contours[i].size()>100)
        { 
            cv::approxPolyDP( cv::Mat(contours[i]), contours_poly[i], 3, true );
            cv::Rect appRect( boundingRect( cv::Mat(contours_poly[i]) ));
            if (appRect.width>appRect.height) 
                boundRect.push_back(appRect);
        }
    return boundRect;
}

用法:

int main(int argc,char** argv)
{
    //Read
    cv::Mat img1=cv::imread("side_1.jpg");
    cv::Mat img2=cv::imread("side_2.jpg");
    //Detect
    std::vector<cv::Rect> letterBBoxes1=detectLetters(img1);
    std::vector<cv::Rect> letterBBoxes2=detectLetters(img2);
    //Display
    for(int i=0; i< letterBBoxes1.size(); i++)
        cv::rectangle(img1,letterBBoxes1[i],cv::Scalar(0,255,0),3,8,0);
    cv::imwrite( "imgOut1.jpg", img1);  
    for(int i=0; i< letterBBoxes2.size(); i++)
        cv::rectangle(img2,letterBBoxes2[i],cv::Scalar(0,255,0),3,8,0);
    cv::imwrite( "imgOut2.jpg", img2);  
    return 0;
}

结果:

一个.元素 = getStructuringElement(cv::MORPH_RECT, cv::Size(17, 3));

a. element = getStructuringElement(cv::MORPH_RECT, cv::Size(17, 3) );

B.元素 = getStructuringElement(cv::MORPH_RECT, cv::Size(30, 30));

b. element = getStructuringElement(cv::MORPH_RECT, cv::Size(30, 30) );

提到的其他图像的结果相似.

Results are similar for the other image mentioned.

这篇关于提取文本 OpenCV的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持跟版网！