Moving Object Detection And Tracking Under Rotating Camera

In recent years,owing to the development of computer vision,moving object detection and tracking technology has been widely applied in military,transportation,and environmental monitoring etc.The detection and tracking technology of moving object on fixed cameras has been gradually matured.However,in many practical applications,fixed cameras cannot achieve the gaze monitoring and continuous tracking of the moving object.Camera motion such as rotating-scan is required to increase the monitoring range and achieve gaze monitoring.Compared with the moving object detection and tracking under fixed camera conditions,camera's motion makes moving object detection and tracking more difficult.Under the condition of camera rotating-scan,nonlinear transformation occurs between the images.In addition,the background in of the image varies over time.The traditional method for detecting and tracking moving object under fixed cameras will be no longer applicable.The conmmen solution of this problem is to compensate the images to eliminate the background motion and the nonlinear transformation between images.The key step of image compensation is to find an accurate motion model to describe image transformations between frames.However,the traditional linear model can not solve the nonlinear transform that is generated by the rotating-scan movement of cameras.At the same time,the existing methods is not able to meet the application requirements in terms of calculation time and accuracy simultaneously.In addition,when the object is cross-frame and continuously tracked for a long time in the video image.A common coordinate system is required to describe the object motion trajectory.However,under the condition of camera rotating-scan,different shooting angles of the camera make the imaging planes different.So the relative coordinate data of the object cannot be obtained in the tracking process,and the motion trajectory of the object cannot be described.So the video image sequence needs to be normalized into a unified coordinate system.Therefore,there are still many difficulties and challenges in the research field of moving object detection and tracking under rotating-scan camera.In this paper,due to the limitations of the existing motion models in the image compensation methods,a robust image compensation method under the condition of camera rotating-scan is proposed,which can achieve background motion compensation and image nonlinear transform compensation simultaneously.In this method,according to the analyses of the camera imaging mechanism for rotating-scan,a nonlinear motion model is proposed.Based on the nonlinear motion model,a camera equation is established,which is transformed into a linear problem further by the parameter space conversion.Then the Hough transform is utilized to estimate the parameters of the global motion model by the matched point pairs.Through this image compensation method,the problem of background motion and nonlinear transformation between images can be eliminated.The problem of moving object detection and tracking under dynamic background can be converted into the problem of moving object detection and tracking under static background.In order to better detect and track moving object under the condition of camera rotating-scan,and to solve the problem of the imaging plane between frames is not uniform due to different camera shooting angles.In this paper,based on the theory of cylindrical model,a concept of panoramic cylindrical public coordinate system is proposed and combined with image compensation technology to complete the detection and tracking of moving object.The method performs frame difference detection on the basis of image compensation,and performs projection on the cylindrical common coordinate system of the image sequence during object detection,and normalizes the image sequence into a uniform cylindrical common coordinate system.At the same time,the relative coordinates of the moving object in the common coordinate system are dynamically solved to calculate the object motion trajectory,and the moving object is tracked in real time.