Intro: Highlighted Borders Using OpenCV on DragonBoard [DRAGONBOARD][QUALCOMM]
This is an Instructable to show you how to do this type of "highlight" like on the image.
I made this to detect if, on an ambulance, a patient has any hemorrhage on his body, so i highlighted the red color. You'll be able to choose what color to highlight when you complete the tutorial :)
To learn more about OpenCV library you can access the library on http://docs.opencv.org/2.4.10/index.html; to install the lib on C++ on Debian based OS, in this Zip you'll find a script file called "opencv.sh", it'll download and install the lib; to have some nice examples, in this same Zip, you'll find a folder called "Códigos da Apresentação". Inside the folder, if you have the lib installed, you can use the "rodar.sh" script to compile your C++ code(just do "$./rodar.sh Filename" and it will compile and run); Also in this Zip, there is a workshop's PDF (It's only in PT-BR only, unfortunately) explaining some Image processing treats and, on the final slide, there is some util links. The code that I used to make this instructable is based on some examples of this Zip. I'd like to thank my friends from RoboCin team to that. :)
So... Let's do it!
Step 1: All in One
What did i do?
All that i do was study those 6, 5, 1_3 algorithms and put in one code, the IMaca.
What was my Idea?
I catch a frame from the camera, then I choose only the red color (segmentation part), then i map all the borders that is shown on the segmented frame to the original frame and i highlight the mapped borders.
Step 2: Segmentation
To do the segmentation of the image,i studied this algorithm.
When you understand it, you'll be able to select what color to highlight in the image.