Translator Disclaimer
17 February 2003 Optical sensed image fusion based on neural networks
Author Affiliations +
Proceedings Volume 4833, Applications of Photonic Technology 5; (2003)
Event: Applications of Photonic Technology 5, 2002, Quebec City, Canada
This paper proposes a neural network-based technique for improving the quality of the image fusion as required for the remote sensing (RS) imagery. This proposes to exploit information about the point spread fucntions of the corresonding RS imaging systems combining it with prior realistic knowledge about the properites of teh scene contained in the maximum entropy (ME) a priori image model. Applying the aggregate regularization method to solve the fusion tasks aimed to achieve the best resolution and noise suppression performances of the overall resulting image solves the problem. The proposed fusion method assuems the availability to control the design parameters, which influence the overall restoration performances. Computationally, the fusion method is implemented using the maximum entropy Hopfield-type neural network with adjustable parameters. Simulations illustrate the improved performances of the developed MENN-based fusion method.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

Back to Top