An enhanced reasoning approach for multimodal inspection data fusion

Abstract

Multimodal inspection is widely used in the diagnosis and evaluation for complex engineering systems. By the use of multiple sensors, a reasonable reasoning approach is required to fuse multi-data acquired over conventional single modality sensors in inspection. In this paper, an enhanced evidential reasoning algorithm to effectively combine uncertain information is presented. We begin with a summary of the relative strength and weakness of vague sets and D-S theory, then the similarity between them is discussed, thereafter an enhanced reasoning algorithm is proposed by combining vague sets and D-S theory to realize the diagnosis and evaluation under the situation in which the presented information from various source is not only uncertain but also imprecise and vague. An example is finally given to illustrate the discussion.