VIDEO STABILIZATION BY MOBILE SENSOR FUSION

摘要

As processing capacity is increasingly usable, the robotics system switches away from active sensors for the perception of the world and turns into passive vision sensors. With the advent of teleportation and real time video monitoring of complex environments, the need for onboard mobile robots to stabilize video has become more prevalent. Many of today’s handheld video capture systems uses a rotating shutter camera, such as a smartphone. The footage is taken from a handheld camera, resulting in moving point or targeted image blur as well as frame to frame jitters. The typical way to protect the footage is with the use of additional hardware including tripod stand. This paper introduces an algorithm of optical stabilization that combines inertial fusion adaptive low pass filters. Using picture processing with rolling shutter camera correction, the suggested method uses accelerometer and gyroscope to approximate camera orientation by picture distortion caused by video recording system motion. Adaptive filter algorithms can be used to create a smooth camera movement, and frame to frame jitters can be minimized using smoothing camera movement. For every android smartphone that has been built in accelerometer and gyroscope, the suggested program can be introduced.