A strange request, but something that I thought I'd share with you.  Here is a simple console application that converts all pages in a hidden text behind image searchable PDF to a image-only non-searchable PDF.  It requires that each page be in memory at once.  If you have more than a few pages, you might want to save a temporary image to a file and using PdfEncoder with PdfImageCollection to save.

using System;
using System.Collections.Generic;
using System.Text;
using System.IO;
using Atalasoft.Imaging;
using Atalasoft.Imaging.Codec;
using Atalasoft.Imaging.Codec.Pdf;

namespace ConvertPdfToPdf
{
    class Program
    {
        //first arg is input, second arg is output
        static void Main(string[] args)
        {
            string file = args[0];
            ImageCollection col = new ImageCollection();
            using (FileStream fs = File.OpenRead(file))
            {
                using (Document doc = new Document(fs))
                {
                    for (int i = 0; i < doc.Pages.Count; i++)
                    {
                        ExtractedImageInfo[] info = doc.Pages[i].ExtractImages();
                        if (info.Length > 1)
                            throw new ApplicationException("contains more than one image");
                        col.Add(info[0].Image);
                    }
                }
            }
            col.Save(args[1], new PdfEncoder(), null);
        }
    }
}