Title: Computational enhancement of large scale environmental imagery: aggregation of robust numerical regularisation, neural computing and digital dynamic filtering
Authors: Yuriy Shkvarko, Ivan Villalon
Addresses: Department of Telecommunications, CINVESTAV Campus Guadalajara, Av. Cientifica 1145, Col. El Bajio, Guadalajara, Jalisco C.P. 45015, Mexico. ' Department of Telecommunications, CINVESTAV Campus Guadalajara, Av. Cientifica 1145, Col. El Bajio, Guadalajara, Jalisco C.P. 45015, Mexico
Abstract: We address a new efficient robust optimisation approach to large-scale environmental image reconstruction/enhancement as required for remote sensing imaging with multi-spectral array sensors/SAR. First, the problem-oriented robustification of the previously proposed Fused Bayesian-Regularization (FBR) enhanced imaging method is performed to alleviate its ill-poseness due to system-level and model-model uncertainties. Second, the modification of the Hopfield-type Maximum Entropy Neural Network (MENN) is proposed that enables such MENN to perform numerically the robustified FBR technique via computationally efficient iterative scheme. The efficiency of the aggregated robust regularised MENN technique is verified through simulation studies of enhancement of the real-world environmental images.
Keywords: nonlinear regularisation; image enhancement; numerical inverse problems; entropy; neural networks; large scale imagery; optimisation; remote sensing; environmental images; simulation.
International Journal of Computational Science and Engineering, 2007 Vol.3 No.3, pp.219 - 231
Published online: 18 Apr 2008 *Full-text access for editors Access for subscribers Purchase this article Comment on this article