Data fusion through fuzzy reasoning applied to feature extraction from multisensory images

Multi-sensor systems provide a purposeful description of the environment that a single sensor cannot offer. Fusing several types of data enhances the recognition capability of a robotic system and yields more meaningful information otherwise unavailable or difficult to acquire by a single sensory modality. Because observations provided by sensors are uncertain, incomplete, and/or imprecise, we adopted the use of the theory of fuzzy sets as a general framework to combine uncertain measurements. We developed a fusion formula based on the measure of fuzziness, and then tested it mathematically against several desirable properties of fusion operators. We established a fuzzification scheme by which different types of input data (images) would be modeled. This modeling process was essential in providing suitable predictions and explanations of a set of observations in a given environment. A defuzzification scheme was then carried out to recover crisp data from the combined fuzzy assessment. This approach was implemented and tested with real range and intensity images acquired by an Odetics Laser Scanner. The goal was to obtain better scene descriptions through a segmentation process of both images. Despite the low resolution of the images and the amount of noise level associated with the acquisition process, the segmented output picture should be suitable for recognition purposes. Other data fusion approaches such as the Super Bayesian Approach and Dempster's rule of combination were implemented and evaluated. A systematic method for evaluating and comparing segmentation results was presented. A level of noise was added to the real data and then segmentation results from all approaches were compared and evaluated. Limitations of methods which deal with combining evidence and managing uncertainty were outlined.