I happened to take about 300 photos with my phone of a packet, from a professor, I needed to read. After doing so, I wanted to use an application for perspective transformation and so I began searching for an application that would allow me to do so by selecting four corners of my quadrangle. I was unable to find a manual QUAD to QUAD image perception tranformation application for sets of images so in turn...
- Input a folder path filled with raw photo images
- Input a specified output path (path must already be created, it will not create folders)
- Process each image manually by traversing through each image and allow you to choose four points of the page (click the "c" key to cancel early and keep processed images)
- See what photos you have "fixed" in a small list in GUI
- Use the specified output path to filter the images into a PDF File. (I have specified a width that happened to fit my photos in the page most appropriately, you may want to switch it to cater to your needs)
- You may view the PDF by going to the main directory and opening "output.pdf"
I hope that this code can help someone else accomplish similiar tasks. Please continue to credit. I do know about using OpenCV's contouring and boundry boxes, but due to the fact that I was taking hundreds of photos quickly, the shape of each page became too distorted and skewed to rely on computer vision.
I continue to believe that taking photos of all the pages and creating this entire program has taken less time than it would have had I turned and scanned each individial page in the book in a flatbed scanner.
