A general computational framework is proposed to address the problem o
f labeling oceanographic images. The essential ideas stem from fitting
a bicubic polynomial to each pixel's neighborhood and assigning topol
ogical labels based on the first and second directional derivatives of
the polynomial surface. The topographic approach has been successfull
y applied by numerous previous investigators to other problems. The mo
tivation of our work is to build a complete autonomous and self-contai
ned module that segments and labels oceanographic images. In this pape
r, the relationship between the oceanographic features in infrared sat
ellite imagery and the topographic structures is presented. Algorithms
are developed that demonstrate the ability to locate and identify the
North and South Walls of the Gulf Stream and to find approximate cent
ers of warm and cold eddies. Experimental results on detecting these o
ceanographic features are provided.