Agfdhyk

I am working with PDF/A style PDF documents that have a mixture of scanned in, full page size images and then a page or two after the image pages that have text in a ColumnText object.

Using Java, how do i detect which pages have an image?

The intent to detect which pages have either images or text is to determine where the first page with text appears. I need to either edit the text or replace the page(s) with text with updated text. The pages with images would remain untouched.

I'm using iText5 and don't currently have the option of upgrading to iText7.

Here's the solution I implemented with the solution provided by @mkl:

ImageDetector.java

package org.test.pdf;



import com.itextpdf.text.pdf.parser.ImageRenderInfo;

import com.itextpdf.text.pdf.parser.RenderListener;

import com.itextpdf.text.pdf.parser.TextRenderInfo;



public class ImageDetector implements RenderListener {

    public void beginTextBlock() { }

    public void endTextBlock() { }

    public void renderText(TextRenderInfo renderInfo) {

        textFound = true;

    }



    public void renderImage(ImageRenderInfo renderInfo) {

        imageFound = true;

    }



    boolean textFound = false;

    boolean imageFound = false;

}

PdfDocumentServiceTest.java

package org.test.pdf;



import com.itextpdf.text.pdf.PdfReader;

import com.itextpdf.text.pdf.parser.PdfReaderContentParser;

import com.test.PdfService;

import org.junit.Assert;

import org.junit.Test;

import org.junit.runner.RunWith;

import org.springframework.beans.factory.annotation.Autowired;

import org.springframework.boot.test.context.SpringBootTest;

import org.springframework.test.annotation.DirtiesContext;

import org.springframework.test.context.ActiveProfiles;

import org.springframework.test.context.junit4.SpringRunner;

import org.springframework.transaction.annotation.Transactional;



@ActiveProfiles({"local", "testing"})

@DirtiesContext

@Transactional

@RunWith(SpringRunner.class)

@SpringBootTest

public class PdfDocumentServiceTest {



    @Autowired

    private PdfService pdfService;



    @Test

    public void testFindImagesInPdf(Long pdfId)) {

        final byte resource = PdfService.getPdf(pdfId);

        int imagePageCount = 0;

        int textPageCount = 0;

        if (resource != null && resource.length > 0) {

            PdfReader reader = new PdfReader(resource);

            PdfReaderContentParser parser = new PdfReaderContentParser(reader);



            for (int pageNumber = 1; pageNumber <= reader.getNumberOfPages(); pageNumber++) {



                ImageDetector imageDetector = new ImageDetector();

                parser.processContent(pageNumber, imageDetector);



                if (imageDetector.imageFound) {

                    imagePageCount++;

                }

                if (imageDetector.textFound) {

                    textPageCount++;

                }

            }

            Assert.assertTrue(imagePageCount > 0);

            Assert.assertTrue(textPageCount > 0);

        }

    }

}

Using iText 5 you can find out whether images actually are shown on a page by parsing the page content into a custom RenderListener implementation. E.g.

class ImageDetector implements RenderListener {

    public void beginTextBlock() { }

    public void endTextBlock() { }

    public void renderText(TextRenderInfo renderInfo) { }



    public void renderImage(ImageRenderInfo renderInfo) {

        imageFound = true;

    }



    boolean imageFound = false;

}

used like this:

PdfReader reader = new PdfReader(resource);

PdfReaderContentParser parser = new PdfReaderContentParser(reader);

for (int pageNumber = 1; pageNumber <= reader.getNumberOfPages(); pageNumber++)

{

    ImageDetector imageDetector = new ImageDetector();

    parser.processContent(pageNumber, imageDetector);

    if (imageDetector.imageFound) {

        // There is at least one image rendered on page i

        // Thus, handle it as an image page

    } else {

        // There is no image rendered on page i

        // Thus, handle it as a no-image page

    }

}

As a possible improvement: In a comment you mention full-page-size images. Thus, in the ImageDetector method renderImage you might want to check the image size before setting imageFound to true. Via the ImageRenderInfo parameter you can retrieve both information on how large the image is displayed on the page and how large it actually is.

Try the below code example (Spire.PDF for Java library is needed), hopefully it works for you.

    PdfDocument doc = new PdfDocument();

    doc.loadFromFile("sample.pdf");



    for(int i = 0; i < doc.getPages().getCount(); i ++) {

        PdfPageBase page = doc.getPages().get(i);

        PdfImageInfo imageInfo = page.getImagesInfo();



        if (imageInfo != null && imageInfo.length > 0) {

            System.out.println("Page" + i + "contains image");

        } 

        else {

            System.out.print("Page" + i + "doesn't contain image");

        }

Disclaimer: I work for Spire.