Day046 — Switching to RealSense
It is a follow up post for Day043. After playing around with the librealsense 2.0 example that integrated with OpenCV. It is time to use the library for my need. I am going to write a program that detects hands in the frame.
I am a noobie for hand detection algorithm. It is best for me to see how other people do it. My program will be based on simena86's work.
In the tutorial he gave, a RGB camera is used to check the color profile of a hand and use the detected color for segmentation for the hand and background. However, we can extract the depth information from RealSense. I solely use the depth info from the RealSense device and use a fixed threshold for segmenting the foreground and background. I know that is the most unreliable way for segmentation, but I just want to keep everything else simple and focus on the RealSense integration.
You can see the change I made in code with below commit.