A hyperspectral data set can be used for testing binary and multi-class change detection techniques.
All the files are in .mat format that can be loaded in Matlab.
PreImg_2004: Pre-processed Hyperion image acquired on May 1, 2004;
PostImg_2007: Pre-processed Hyperion image acquired on May 8, 2007;
Reference_Map_Binary: A binary reference map for evaluting the binary change detection performance (Two classes: change and no-change).
Reference_Map_Multiclass: A multiclass reference map for evaluting the multiclass change detection performance (Seven classes: six change classes and no-change).
This dataset is made up of a pair of bitemporal hyperspectral images acquired by the Hyperion sensor mounted on board the EO-1 satellite on May 1, 2004 and May 8, 2007. The study area is an irrigated agricultural land of Benton County, Oregon, USA, which has a size of 180 ×225 pixels.
Preprocessing operations were made (i.e., repairing bad stripes, removal of uncalibrated and noisiest bands, atmospheric correction, co-registration) on two images, where 159 bands (i.e., bands 8–57, 82–119, 131–164, 182–184, and 187–220) out of original 242 bands were selected.
The major land-cover changes in this scenario are due to the transitions among different kinds of crops, soil and other land-cover types.
(a) May 1, 2004; (b) May 8, 2007; (c) Binary CD Reference Map; (d) Multiclass CD Reference Map
Change Class 1 (C1): 1034 pixels
Change Class 2 (C2): 1048 pixels
Change Class 3 (C3): 5111 pixels
Change Class 4 (C4): 1261 pixels
Change Class 5 (C5): 479 pixels
Change Class 6 (C6): 988 pixels
No-change Class (NC): 30579 pixels
Total: 40500 pixels
If you use this data set for your research, please cite the following papers.
[1] S. Liu, D. Marinelli, L. Bruzzone and F. Bovolo, "A Review of Change Detection in Multitemporal Hyperspectral Images: Current Techniques, Applications, and Challenges," IEEE Geoscience and Remote Sensing Magazine, vol. 7, no. 2, pp:140-158, 2019. DOI: 10.1109/MGRS.2019.2898520
[2] S. Liu, Q. Du, X. Tong, A. Samat, H. Pan , X. Ma, “Band Selection based Dimensionality Reduction for Change Detection in Multitemporal Hyperspectral Images,” Remote Sensing, vol. 9, no.10, pp:1008, 2017. DOI: 10.3390/rs9101008
[3] S. Liu, L. Bruzzone, F. Bovolo, P. Du., “Unsupervised Multitemporal Spectral Unmixing for Detecting Multiple Changes in Hyperspectral Images,” IEEE Transactions on Geoscience and Remote Sensing, vol.54, no. 5, pp:2733-2748, 2016. DOI: 10.1109/TGRS.2015.2505183