2021-06-21-论文笔记-Interpretable Hyperspectral Artificial IntelligenceWhen nonconvex: meets hyperspectral remote sensing

发表于 分类于

本文从高光谱图像恢复，降维，数据融合与增强，光谱解混，为大规模地表覆盖制图服务的多模学习技术（cross-modality learning）5个方面论述了非凸方法在高光谱智能解译中的应用。

0. 基本信息

  • 引用信息

D. Hong et al., "Interpretable Hyperspectral Artificial Intelligence: When nonconvex modeling meets hyperspectral remote sensing," in IEEE Geoscience and Remote Sensing Magazine, vol. 9, no. 2, pp. 52-87, June 2021, doi: 10.1109/MGRS.2021.3064051.

  • bibtex
1
2
3
4
5
6
7
8
9
10
@article{Hong2021,
  author  = {Hong, Danfeng and He, Wei and Yokoya, Naoto and Yao, Jing and Gao, Lianru and Zhang, Liangpei and Chanussot, Jocelyn and Zhu, Xiaoxiang},
  journal = {IEEE Geoscience and Remote Sensing Magazine},
  title   = {Interpretable Hyperspectral Artificial Intelligence: When nonconvex modeling meets hyperspectral remote sensing},
  year    = {2021},
  volume  = {9},
  number  = {2},
  pages   = {52-87},
  doi     = {10.1109/MGRS.2021.3064051}
}

1. 研究背景

  1. 本文主要介绍了以下几方面：

    1. 高光谱图像恢复（restoration）
    高光谱图像恢复
    1. 降维（dimensionnality reduction ）
    高光谱降维，避免维度灾难
    1. 数据融合与增强
    多/高光谱数据融合
    1. 光谱解混
    光谱解混
    1. 为大规模地表覆盖制图服务的多模学习技术（cross-modality learning）
    为大规模地表覆盖制图服务的多模学习技术（CML）

  2. 高光谱存在广泛应用：As a result, HS RS has been significantly advanced and widely applied in many challenging tasks of Earth observa-tion [5], such as fine-grained land cover classification, min-eral mapping, water-quality assessment, precision farming, urban planning and monitoring, disaster management and prediction, and concealed-target detection

  3. 高光谱3D数据立方体相对于1D和2D数据的优势：

    1. 1D：有更多的语音和结构信息可以辅助区分
    2. 2D：有更精细的光谱信息，更好的区分材料的细小差异，可以分辨十分详细的材料，进行理化参数的定量反演，食品安全保护等

  4. 高光谱遥感任务（卫星）：

    1. MODIS
    2. Hyperspectral Satallite for Earch Observation(HypSEO)
    3. German Aerospace Center Earth Sensing Imaging Spectrometer (DESIS)
    4. Gaofen-5
    5. EnMap
    6. HyspIRI

  5. 在大数据时代，以专家为系统核心的数据处理流程存在绝大的瓶颈，无法满足需求（With the ever-growing availability of RS data sources from both satellite and airborne sensors on a large and even global scale, expert system-centric data processing and the analysis mode have run into bottlenecks and cannot meet the demands of the big data era）