Sitting Gesture Detection Algorithm Research And Design Based On Intel RealSense

In recent years,more and more people use computers all the day when they are working at office.The incidence of disease shoulders,lumbar spine and other parts of their body is increasingly high,while at the mean time their eye sight are not as good as before.How to avoid disease problems has become an important issue.The most appropriate and most effective way to real daily work is to maintain good posture,and how to detect posture has become the key to solving the problem.In this paper,we conducted extensive research and exploration,and ultimately designed a new algorithm which is for efficient,accurate and practical using depth camera.According to the principle it can be divided into smart hardware detection and image recognition detection.After working through a variety of the latest smart hardware available to experience and use,many problems were found.Through the study of image recognition detection scheme we found that the way using a single camera flawed in terms of accuracy and robustness,the use of multi-camera programs in real-time and operability are also some drawbacks,therefore,combined with a variety of programs and according to the latest technological developments,innovative use of Intel RealSense 3D camera becomes the solution.After the realization of the entire posture detection algorithm has been fully experimental verification,proven,the proposed method has high robustness,accuracy and practical significance.