유로샛

  • 설명 :

EuroSAT 데이터 세트는 13개의 스펙트럼 대역을 포괄하고 27,000개의 레이블이 지정되고 지리 참조된 샘플이 있는 10개의 클래스로 구성된 Sentinel-2 위성 이미지를 기반으로 합니다.

두 가지 데이터 세트가 제공됩니다. - rgb: JPEG 이미지로 인코딩된 광학 R, G, B 주파수 대역만 포함합니다. - all: 원래 값 범위(float32)의 13개 밴드를 모두 포함합니다.

URL: https://github.com/phelber/eurosat

나뉘다
'train' 27,000
  • 인용 :
@misc{helber2017eurosat,
    title={EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification},
    author={Patrick Helber and Benjamin Bischke and Andreas Dengel and Damian Borth},
    year={2017},
    eprint={1709.00029},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}

eurosat/rgb(기본 구성)

  • 구성 설명 : Sentinel-2 RGB 채널

  • 다운로드 크기 : 89.91 MiB

  • 데이터 세트 크기 : 89.50 MiB

  • 자동 캐시 ( 문서 ): 예

  • 기능 구조 :

FeaturesDict({
    'filename': Text(shape=(), dtype=string),
    'image': Image(shape=(64, 64, 3), dtype=uint8),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=10),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
파일 이름 텍스트
영상 영상 (64, 64, 3) uint8
상표 클래스 레이블 int64

심상

유로위성/모두

  • 구성 설명 : 13 Sentinel-2 채널

  • 다운로드 크기 : 1.93 GiB

  • 데이터세트 크기 : 5.36 GiB

  • 자동 캐시 ( 문서 ): 아니요

  • 기능 구조 :

FeaturesDict({
    'filename': Text(shape=(), dtype=string),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=10),
    'sentinel2': Tensor(shape=(64, 64, 13), dtype=float32),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
파일 이름 텍스트
상표 클래스 레이블 int64
센티넬2 텐서 (64, 64, 13) float32