RealSense Bike Fitting
Biking is a new hobby I have immersed myself into since the last 2 years. In my journey of becoming a novice to professional biker, I learned many things.
One of the most important things was that the bike should be perfectly fitted to you.
A properly fitted bike will not only improve your performance, but also avoid and long term injuries.
Bike Fitting is an expensive, time consuming and a complex process. Our bodies change and our bikes need to be refitted. A bike fitting session can cost around $200 to $600. The Sport professional fitting can cost upto $1500.
I had 2 bike fitting sessions which costed me $800 in total but one thing I noticed most of the process is manual. It is all about measuring angles manually and changing things on your bike to achieve the perfect fit.
My motivation for this project was to automate this measurement process and save some money and do it for cheap.
2 RealSense R200 cameras are used in the setup. One is kept to the left of the bike on the trainer and one is to the right.
The biker wears reflective markers on both sides of the body at ball of the foot, ankle, knee, hip, shoulder, elbow and hands.
The IR cameras track the markers and identify the joints automatically and and displays the joint angle measurements. The colors of the angles show you how off they are.
Depending on how off these measurements are, you can tweak the positions of your seat, handle, cleats, etc. till you get the most comfortable fit and close to perfect angles.
My Contribution
-
Interviewed bike fitting experts
-
Sketched Ideas and concepts
-
Created the application using 2 RealSense R200 depth cameras, OpenCV, C++, OpenFrameworks
-
Tested the application with biker friends by fitting their bike saving them a lot of money