In developing countries, infectious diseases are a serious public health problem. Often times, these diseases are highly related to certain urban conditions found at poor neighborhoods, such as the informal (non-permitted) settlements. Remote sensing can be a valuable tool to study these phenomena, however, the complexity of these informal settlements is still a challenge for remote sensing analysis. For the present research, classification of urban image data with very high spatial resolution but low spectral resolution was considered The identification of which objects and features to look for in the images was done with the help of a leptospirosis contamination risk model Our remote sensing analysis included four levels of segmentation and an object-based classification process. Objects were classified as vegetation, shadow, roofs, streets, open area and other auxiliary classes with reasonable accuracy. © 2008 IEEE.