1 1 1 2
D. Cheng , R. E. Mercer , J. L. Barron and P. Joe
Tp = 0:85). We choose a subset of these tracks, with to plot the tracks, with one or more applications of 3-
the condition that no storm lies upon more than one point averaging of the fuzzy storm centers before the B-
chosen track. spline approximation is performed. By using the fuzzy
We have changed the implementation of the algo- storm center concept to obtain the tracks and the B-
rithm from the processing of a complete image sequence spline approximation with 3-point averaging for their
of known length to an incremental process. Given n display, we can now obtain storm tracks that are long
images the hypothesized storm disparities are relaxed. and smooth and which closely match human perception
When an (n + 1)th image is added, the relaxation is of a \motion picture" of a storm image sequence.
restarted using the results for the rst n images plus the
hypothesized storms of the (n + 1)th image. We have 6. REFERENCES
observed empirically that the result is always the same
as if all n +1 images had been initially relaxed. The dif- [1] D. Cheng. \Tracking fuzzy storm centers in Doppler
ference in computation speed between the two methods radar images". Master's thesis, The University of
is insignicant since relaxation converges within ten it- Western Ontario, Dept. of Computer Science, 1996.
erations in either mode. The incremental algorithm [2] S. L. Horowitz and T. Pavlidis. \Picture segmenta-
allows us to view the current storm tracks as the data tion by a tree traversal algorithm". Journal of the
become available. ACM, 23(2):368{388, 1976.
[3] D. Krezeski, R. E. Mercer, J. L. Barron, P. Joe,
5. EXPERIMENTAL RESULTS and H. Zhang. \Storm tracking in Doppler radar
images". In Proc. International Conf. on Image
Due to space limitations, we only show test results Processing (ICIP), volume III, pages 226{230, 1994.
of our algorithms on two images from two image se-
quences: the 77-series and the 18-series. The storms [4] A. Rosenfeld. \Fuzzy geometry: An overview". In
are moving from left to right in the 77-series and the Proc. First IEEE Conf. Fuzzy Systems (San Diego),
results obtained from our algorithms agree with this pages 113{117, 1992.
fact, as shown in Figure 1. Storm movements are more [5] A. Rosenfeld. \Fuzzy plane geometry: Trian-
complicated in the 18-series. Some of the storms are gles". Pattern Recognition Letters, 15(12):1261{
moving from left to right, and some are moving from 1264, 1994.
top to bottom. Our algorithms are capable of track-
ing these storm movements, as shown in Figure 2. To [6] L. A. Zadeh. \Fuzzy sets". Information and Con-
ensure the display of smooth tracks, we use B-splines trol, 8:338{353, 1965.