Typhoon Cloud Tracking by Kalman Filter
Typhoon cloud has its changeability, so it is difficult to track and predict compared with the rigid targets. Region of interest (ROI) and reference region were selected by using interactive methods. Bezier curve is used to smooth the gray level histogram of ROI and obtain Bezier histogram. The gray level value which is corresponding to the valley of the Bezier histogram is used to segment the ROI in order to get the tracking target. And target parameters could be predicted by using Kalman filter, then getting the moving track of the target. Experimental results show that the proposed algorithm has nice real-time ability and adaptability.
Y. Lv "Typhoon Cloud Tracking by Kalman Filter", Applied Mechanics and Materials, Vols. 58-60, pp. 2487-2492, 2011