Detecting drawn lines and dots on a notebook paper

0
votes

I have a picture of notebook (with squares) and lines and dots are drawn in it like in the description. Output should be a data structure which contains info about boundaries and dots. How one can accomplish that? If possible, program should process this dynamically (given a video).

example

1
image-processingcomputer-visionstraight-line-detection
Nothing yet, I'll let you know when it's done (got some other stuff to do). Btw. i'm going to implement it by myself, from scratch. My idea is to detect edges (Canny edge detector) and apply appropriate threshold . As for the circles, calculating average color in squares seems good to me. I was also thinking about detecting crosses (Harris corner detection??) and see if between two is thicker line or not. – Danijel Pavlek

1 Answers

0
votes

Yes this can be accomplished by various image processing techniques.

One famous technique that can help is called the Canny Edge Detector. It can detect all the defined edges within an image. More can be looked into it here. Various python and C# image processing libraries make this extremely easy. Take for example OpenCV

For detecting dots in the middle of the edges, that would be up to you to come up with, unless anyone knows of a library to make that easy as well. I suggest looking at each square that we detected by the canny edge detector and see if there are any dark color values around the middle.

For the data structure, that is also up to you.

Remember that a video is just a sequence of images. Just apply the same technique to all the images.