P. Ghamisi; R. Souza; J. A. Benediktsson; X. X. Zhu; L. Rittner; R. A. Lotufo, "Extinction Profiles for the Classification of Remote Sensing Data," in IEEE TGRS, vol.54, no.10, pp.5631 - 5645, 2016 [The most popular paper published by IEEE TGRS in July, August, and September 2016]. [Code]Description: In order to use the software, please cite:
P. Ghamisi, J. A. Benediktsson, M. O. Ulfarsson, Spectral-Spatial Classification of Hyperspectral Images Based on Hidden Markov Random Fields, IEEE TGRS. 52(5): 2565-2574, 2014. [Code]Description:
In this work, a new fully automatic framework for the classification of hyperspectral images was implemented. The new method is based on combining Hidden Markov Random Field (HMRF) segmentation with Support Vector Machine (SVM) classifier. In order to preserve edges in the final classification map, a gradient step is taken into account. Experiments confirm that the new spectral and spatial classification approach is able to improve results significantly in terms of classification accuracies compared to the standard SVM method and also outperforms other studied methods. The HMRF was implemented by Dr. Quan Wang and the spectral-spatial classification approach was implemented by me. In order to use the software, please refer to the following articles:
P. Ghamisi, M. S. Couceiro and J. A. Benediktsson, "A Novel Feature Selection Approach Based on FODPSO and SVM," IEEE TGRS, vol. 53, no. 5, pp. 2935-2947, May 2015. [Code]
Description: This code implements a fast and accurate feature selection approach based on FODPSO and SVM dealing with HYPERSPECTRAL remote sensing images. In order to use the MATLAB code, please cite the following paper. For a detailed description on the use of the aforementioned feature selection approach, please refer to [J13].
1. P. Ghamisi, M. S. Couceiro and J. A. Benediktsson, "A Novel Feature Selection Approach Based on FODPSO and SVM," IEEE TGRS, vol. 53, no. 5, pp. 2935-2947, May 2015.
This software contains: PSO, DPSO and FODPSO based segmentation techniques. For a detailed description on the use of the aforementioned segmentation techniques, please refer to [J4] and [J6].
For those of you who are interested in the fusion of LiDAR and hyperspectral data or the classification of hyperspectral images, we made our dataset public. The dataset was captured over Samford Ecological Research Facility (SERF), Queensland, Australia. The dataset is composed of hyperspectral and LiDAR data as well as their corresponding training and test samples. You may download the data from the following address:
Copyright © www.Pedram-Ghamisi.com - All Rights Reserved.