Objective Detection of Center of Tropical Cyclone in Remotely Sensed Infrared Images

In the present work, an objective technique has been presented to fix the center position of TC in the satellite generated infrared images. The basis of the technique is to determine the point around which the fluxes of the gradient vectors of brightness temperature (BT) are converging. First, variance of brightness temperature at each pixel from its neighboring pixels is computed and then flux of the gradient of variance values is computed. Next, a line parallel to the gradient vector at each pixel is drawn across the image, and the locations where these lines intersect each other are stored in a density matrix. The score values accumulated in the density matrix are averaged and location with the highest score is identified. This position is considered to be the center location of the cyclone. The technique has been tested over the Kalpana satellite generated (approximately 1000) IR images of the cyclones that formed during the period 2009-2010. The technique has been used in fully automated mode for the four cyclones viz., Phyan, Ward, Laila, and Phet. The half hourly sequential IR images during the life period of each cyclone is analyzed and the center position is determined. The track of cyclone obtained by the automatically determined center position is compared with the observed track obtained from Joint Typhoon Warning Centre (JTWC). The mean track error with respect to JTWC observations for four cyclones Phyan, Ward, Laila, and Phet was computed and found as 42, 82, 58, and 42.5 km, respectively.