Remote Sensing Task Taxonomy

A classification of machine learning tasks for Earth observation.

Task Categories

graph TB
    EO[Earth Observation ML]

    EO --> PIX[Pixel-Level]
    EO --> SCENE[Scene-Level]
    EO --> TEMP[Temporal]
    EO --> GEN[Generative]

    PIX --> SEG[Segmentation]
    PIX --> REG[Regression]

    SCENE --> CLS[Classification]
    SCENE --> DET[Detection]

    TEMP --> CD[Change Detection]

    GEN --> SR[Super-Resolution]
    GEN --> TRANS[Translation]

Semantic Segmentation

Assign a class label to every pixel.

Formulation:

f: \mathbb{R}^{H \times W \times C} \rightarrow \{1, \ldots, K\}^{H \times W}

Reference: Ronneberger, O., et al. (2015). U-Net: Convolutional Networks for Biomedical Image Segmentation. MICCAI. DOI: 10.1007/978-3-319-24574-4_28

Change Detection

Identify changes between temporal observations.

Formulation:

f: \mathbb{R}^{H \times W \times C} \times \mathbb{R}^{H \times W \times C} \rightarrow \{0, 1\}^{H \times W}

Reference: Daudt, R.C., et al. (2018). Fully Convolutional Siamese Networks for Change Detection. ICIP. DOI: 10.1109/ICIP.2018.8451652

Super-Resolution

Enhance spatial resolution.

f: \mathbb{R}^{H \times W \times C} \rightarrow \mathbb{R}^{sH \times sW \times C}

Reference: Wang, X., et al. (2018). ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks. ECCV Workshops. arXiv:1809.00219

Object Detection

Reference: Lin, T.Y., et al. (2017). Focal Loss for Dense Object Detection. ICCV. DOI: 10.1109/ICCV.2017.324

Benchmark Datasets

Dataset	Task	Resolution	Classes
EuroSAT	Classification	10m	10
LoveDA	Segmentation	30cm	7
LEVIR-CD	Change Detection	50cm	2