Study On Electronic Image Stabilizing Algorithm In Image Pickup System On Vehicle

In image pickup system on vehicle, the output of image sequence is unstable because thevideo camera shakes with the vehicle. This paper studies on each technical tache of electronicimage stabilizing, and puts forward a new stabilizing algorithm. The primary content andcontributions of this paper are as follows:First, it delves into the key technique of electronic image stabilizing algorithm——imageregistration, and brings forward a feature based registration method which is from coarse toprecise, and form local matching to global registration. In the coarse step, an improved KLT(Kanade-Lucas-Tomasi) algorithm is applied in feature extraction, and a new acceleratingalgorithm of template matching is presented to match features, and also a feasible feature validitytesting method is given. In the precise step, the"regional counting method based onjoint-histogram"is used for reference to make out its own improved method, which achieves"sub-pixel"precision as well as avoids heavy floating-point operation by clever definition ofcriterion function.Second, it gives the fundamental and realization of image motion compensation. It mainlystudies on the key technique——intentional motion parameter estimation. It analyzes thelimitation of traditional methods, and expatiates on the application in motion filtering ofrecursive Kalman filter.Third, it explains the arising of undefined regions, and uses mosaicking in reconstructingundefined regions. This reconstructing technique solves the common problems of informationdrop-out and dropping of image quality, and also maintains visual fluency of the imagesequence.The paper evaluates some indexes of performance of this electronic image stabilizingalgorithm by emulational experiments, such as accuracy, resolution and real-time. As experimentresults show, this algorithm is accurate and reliable; its resolution ration is obviously better thanone pixel; and it needs only 35.8ms to handle one frame; so it can meet the demand of imagepickup system on vehicle.