Video Image Stabilization System Design And Target Recognition

With the increase of requirement of military reconnaissance, observation systems, and civil popularity of television surveillance and tracking systems, the requirements of stability of video image are becoming increasingly improved. Based on the analysis of inertial image stabilization and digital image stabilization technology, the real-time image stabilization system of moving vector (such as automotive, ship-borne or airborne) adopted the architecture that process the second digital image stabilization and target identification based on the low-cost gyro inertial image stabilization platform. This system architecture can effectively improve the accuracy and speed of image stabilization, and provide a low-cost and efficient performance solution of real-time video image stabilization system.This article mainly discusses the domestic and foreign development of video image stabilization,working principle and key technologies of video stabilization system design, do research on setting up moving image mathematical model and experiment on system simulation. Based on the practical application of object and the requirements of application environment, do the classified research of existing video image stabilization technology and target recognition algorithm,this system adopted a practical steady-state method of treatment for video images– the RANSAC (Random Sample Consensus) in the block-matching algorithm. And do software simulation of correlation algorithm to estimate of velocity,direction, target distance and other aspects of moving object, and do exploratory research on the relationship between object and image, mapping trajectory map, target detection, identification and tracking.