cross-scene hyperspectral image datasets

Citation Author(s):
Chenglong
Wang
Submitted by:
Chenglong Wang
Last updated:
Mon, 12/30/2024 - 03:15
DOI:
10.21227/khtn-3v61
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

1) RPaviaU-DPaviaC Dataset: The RPaviaU-DPaviaC dataset is constructed by amalgamating two publicly accessible HSI datasets: the ROSIS Pavia University (RPaviaU) scene and the DAIS Pavia Center (DPaviaC) scene. The RPaviaU dataset, featuring dimensions of 610 × 340 × 103, was acquired by the ROSIS HSI sensor over the terrain of the University of Pavia, Italy. Conversely, the DPaviaC dataset, with dimensions of 400 × 400 × 72, was collected using the DAIS sensor over the central area of Pavia City, Italy. These two scenes share a common set of seven land cover classes. 

2) EHangzhou-RPaviaHR Dataset: The EHangzhou-RPaviaHR dataset encompasses two distinctive scenes: the EO-1 Hangzhou (EHangzhou) scene and the ROSIS Pavia HR (RPaviaHR) scene. The EHangzhou scene, sourced from the EO-1 Hyperion hyperspectral sensor, was obtained over Hangzhou City, Zhejiang, China, and possesses dimensions of 590 × 230 × 198. In contrast, the target scene, RPaviaHR, was captured using the ROSIS HSI sensor over Pavia city, Italy, and is characterized by dimensions of 1400 × 512 × 102. These two scenes share a common subset of three land cover classes.

Instructions: 

You can read the details about these datasets in the article"Dual-Stream Discriminative Attention Network for Cross-Scene Hyperspectral Image Classification"

Comments

Please support our works

Submitted by Chenglong Wang on Mon, 12/30/2024 - 03:16