There is no size limitation on the hardware components to be used. This project is only to prove that the algorithm works. Of course, there are very advanced prosthetic legs available as well, but they are usually super expensive. Hence, this project focusses on the development of a simple algorithm that can switch this mode autonomously when the user This could be very inconvenient and also risky at times. So, when a patient or a user of these prosthetic legs is walking towards a set of stairs, he or she must manually switch overįrom 'walking mode' to 'stairs climbing' mode. Prosthetic legs which are available in the market are usually devoid of any battery or motors or other types of active actuators.

Project is focused towards incorporating this technology into prosthetic legs. The overall failure rate was 1% for the algorithm. This project only involves proving and testing out the vision-based software for the leg and does not involve any interfacing with the hardware. Since, the overall algorithm should be capable of running on some single-board micro-computer installed in the prosthetic leg, so no neural networks were used in the image processing. This project uses a Microsoft Kinect depth camera and Opencv libraries to detect stairs in real-time using the depth and RGB images from the camera. That means the person using this leg has to do some kind of manual movement to switch the leg from walking gait to stairs climbing move when he is about to go up or down a stair. Although there are several smart actuated prosthetic legs in the market, they mostly do not have any automated gait modification mechanism built into them. This project is about the detection of stairs using a 3D camera in real-time for a prosthetic leg application.

GitHub - abhanjac/stairs_detection_kinect_opencv: This project is about the detection of stairs using a 3D camera in real-time for a prosthetic leg application.

Solidworks with kinect