Other Research Areas

Multispectral Image Fusion and Classification

Image Fusion

Data fusion is an image compression problem in which two or more data sets of a related observation are combined to produce a composite result that possesses the salient characteristic of each component. How to fuse information depends on the given application as well as the number of data sets available for fusion.

At its most basic, data fusion can be achieved by simple weighted averaging of information (e.g., averaging of noisy measurements of the same observation). A a high level it can involve artificial intelligence and expert systems that must take somewhat diverse data of a related situation and infer complex knowledge.

In our research we focus on multispectral and hyperspectral data fusion of images for the purpose of improved classification in the presence of blurring and noise. Each frame of the multispectral images contain redundant and new information that can be effectively combined to overcome non-idealities of image acquisition and aid in accurate classification.

Related Publications

Sorry, no publications matched your criteria.