Research On Integrated Image Stabilization System

Facing the complex and changing sea environment,marine traffic safety has always been a hot topic related to social economy and people’s livelihood.The sea surface traffic video monitoring system based on running ships and shore bases has become an important means of marine traffic safety information collection and supervision.It provides a strong guarantee for marine traffic safety.Due to the influence of uncontrollable natural factors such as sea surface wind and waves,the ship will appear six-dimensional posture movement.And the ship-borne video surveillance platform will inevitably follow the movement of the hull,which reduces the usability of the surveillance images it generates.To ensure the stability of the imaging equipment visual axis to ensure the quality of the captured video.Under this requirement,this thesis adopts a comprehensive image stabilization scheme to deal with the camera system carried by the platform such as the ship with large turbulence.The first means of mechanical image stabilization is adopted to isolate the carrier jitter from the imaging equipment.On this basis,the video jitter that cannot be handled by mechanical image stabilization is corrected by electronic image stabilization to further improve video quality.This thesis studies the image stabilization problem of ship-borne video at sea as the background.The main research contents are as follows:Based on the analysis of the functional requirements of the ship-borne video stabilization system and the onboard video working environment,an overall system scheme of comprehensive image stabilization is designed.A structural design of a mechanical image stabilization platform that maintains the camera to acquire stable images under the ship motion is proposed in view of the changing law of ship motion.And the design of the execution unit of the stable platform is carried out.The thesis analyzes the attitude angle solution method and determines the quadratic method for attitude solution based on the angular velocity information.For the case that the angular velocity sensed by the gyroscope has drift error,the angular velocity error is corrected by using the acceleration information.Inertial sensors containing accelerometers and gyroscopes are selected based on performance and resolution,and finally the attitude angle is solved based on the selected inertial sensor sensing acceleration information and angular velocity information.The motor control algorithm in the execution unit is studied.According to the characteristics of the controllable speed and angle of the stepping motor,the double-loop control algorithm of the motor is improved,and a time-shortened Proportional Integral Derivative(TS-PID)control algorithm is proposed.The simulation experiment proves that the TS-PID control algorithm proposed in this thesis can effectively shorten the system error compensation time.The problem of residual jitter in the image caused by the control error of the stabilization platform is analyzed,and the correction method of electronic image stabilization is studied and designed.The motion estimation method of Harris feature points,the motion smoothing method of Kalman filter and the motion compensation method of bilinear interpolation are used to electronically stabilize the video after mechanical stabilization.Finally,a STM32-based hardware experimental system is built to experiment the attitude information acquisition and decoding and attitude angle compensation to verify the performance improvement of the TS-PID algorithm in speeding up the compensation speed.An electronic image stabilization experimental system was established to electronically stabilize shipboard video with similar video jitter caused by stabilization platform compensation errors.Finally,the before and after video stabilization is evaluated based on PSNR and difference maps.The feasibility of the comprehensive image stabilization system is verified by combining the results of the stable platform attitude compensation experiment and the electronic image stabilization experiment.