Add DOTA dataset

docs/api/datasets.rst

@@ -289,6 +289,10 @@ DL4GAM
 ^^^^^^
 .. autoclass:: DL4GAMAlps
 
+DOTA
+^^^^
+.. autoclass:: DOTA
+
 ETCI2021 Flood Detection
 ^^^^^^^^^^^^^^^^^^^^^^^^
 

docs/api/datasets/non_geo_datasets.csv

@@ -15,6 +15,7 @@ Dataset,Task,Source,License,# Samples,# Classes,Size (px),Resolution (m),Bands
 `DFC2022`_,S,Aerial,"CC-BY-4.0","3,981",15,"2,000x2,000",0.5,RGB
 `Digital Typhoon`_,"C, R",Himawari,"CC-BY-4.0","189,364",8,512,5000,Infrared
 `DL4GAM`_,S,"Sentinel-2","CC-BY-4.0","2,251 or 11,440","2","256x256","10","MSI"
+`DOTA`_,OD,"Google Earth, Gaofen-2, Jilin-1, CycloMedia B.V.","CC-BY-NC-4.0","5,229",15,"1000-5000",RGB
 `ETCI2021 Flood Detection`_,S,Sentinel-1,-,"66,810",2,256x256,5--20,SAR
 `EuroSAT`_,C,Sentinel-2,"MIT","27,000",10,64x64,10,MSI
 `EverWatch`_,OD,Aerial,"CC0-1.0","5,325",7,"1,500x1500p",0.01,RGB

tests/data/dota/data.py

@@ -0,0 +1,159 @@
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+import hashlib
+import os
+import shutil
+import tarfile
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+from PIL import Image
+
+
+def create_dummy_image(path: Path, size: tuple[int, int] = (64, 64)) -> None:
+    """Create small dummy image."""
+    img = np.random.randint(0, 255, (*size, 3), dtype=np.uint8)
+    Image.fromarray(img).save(path)
+
+
+def create_annotation_file(
+    path: Path, is_hbb: bool = False, no_boxes: bool = False
+) -> None:
+    """Create dummy annotation file with scaled coordinates."""
+    if is_hbb:
+        # Horizontal boxes scaled for 64x64
+        boxes = [
+            '10.0 10.0 20.0 10.0 20.0 20.0 10.0 20.0 plane 0\n',
+            '30.0 30.0 40.0 30.0 40.0 40.0 30.0 40.0 ship 0\n',
+        ]
+    else:
+        # Oriented boxes scaled for 64x64
+        boxes = [
+            '10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0\n',
+            '30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0\n',
+        ]
+
+    if no_boxes:
+        boxes = []
+
+    with open(path, 'w') as f:
+        f.write('imagesource:dummy\n')
+        f.write('gsd:1.0\n')
+        f.writelines(boxes)
+
+
+def create_test_data(root: Path) -> None:
+    """Create DOTA test dataset."""
+    splits = ['train', 'val']
+    versions = ['1.0', '1.5', '2.0']
+
+    # Create directory structure
+    for split in splits:
+        num_samples = 3 if split == 'train' else 2
+
+        if os.path.exists(root / split):
+            shutil.rmtree(root / split)
+        for version in versions:
+            # Create images and annotations
+            for i in range(num_samples):
+                img_name = f'P{i:04d}.png'
+                ann_name = f'P{i:04d}.txt'
+
+                # Create directories
+                (root / split / 'images').mkdir(parents=True, exist_ok=True)
+                (root / split / 'annotations' / f'version{version}').mkdir(
+                    parents=True, exist_ok=True
+                )
+
+                # Create files
+                if i == 0:
+                    no_boxes = True
+                else:
+                    no_boxes = False
+                create_dummy_image(root / split / 'images' / img_name)
+                create_annotation_file(
+                    root / split / 'annotations' / f'version{version}' / ann_name,
+                    False,
+                    no_boxes,
+                )
+
+            # Create tar archives
+            for type_ in ['images', 'annotations']:
+                src_dir = root / split / type_
+                if src_dir.exists():
+                    tar_name = f'dotav{version}_{type_}_{split}.tar.gz'
+                    with tarfile.open(root / tar_name, 'w:gz') as tar:
+                        tar.add(src_dir, arcname=f'{split}/{type_}')
+
+            # print md5sums
+            def md5(fname: str) -> str:
+                hash_md5 = hashlib.md5()
+                with open(fname, 'rb') as f:
+                    for chunk in iter(lambda: f.read(4096), b''):
+                        hash_md5.update(chunk)
+                return hash_md5.hexdigest()
+
+    print('file_info = {')
+    for split in splits:
+        print(f"    '{split}': {{")
+
+        for type_ in ['images', 'annotations']:
+            print(f"        '{type_}': {{")
+
+            for version in versions:
+                tar_name = f'dotav{version}_{type_}_{split}.tar.gz'
+                checksum = md5(tar_name)
+
+                # version 1.0 and 1.5 have the same images
+                if version == '1.5' and type_ == 'images':
+                    version_filename = '1.0'
+                else:
+                    version_filename = version
+
+                print(f"            '{version}': {{")
+                print(
+                    f"                'filename': 'dotav{version_filename}_{type_}_{split}.tar.gz',"
+                )
+                print(f"                'md5': '{checksum}',")
+                print('            },')
+
+            print('        },')
+
+        print('    },')
+    print('}')
+
+
+def create_sample_df(root: Path) -> pd.DataFrame:
+    """Create sample DataFrame for test data."""
+    rows = []
+    splits = ['train', 'val']
+    versions = ['1.0', '1.5', '2.0']
+
+    for split in splits:
+        num_samples = 3 if split == 'train' else 2
+        for version in versions:
+            for i in range(num_samples):
+                img_name = f'P{i:04d}.png'
+                ann_name = f'P{i:04d}.txt'
+
+                row = {
+                    'image_path': str(Path(split) / 'images' / img_name),
+                    'annotation_path': str(
+                        Path(split) / 'annotations' / f'version{version}' / ann_name
+                    ),
+                    'split': split,
+                    'version': version,
+                }
+                rows.append(row)
+
+    df = pd.DataFrame(rows)
+    df.to_csv(root / 'samples.csv')
+    return df
+
+
+if __name__ == '__main__':
+    root = Path('.')
+    create_test_data(root)
+    df = create_sample_df(root)

tests/data/dota/samples.csv

@@ -0,0 +1,16 @@
+,image_path,annotation_path,split,version
+0,train/images/P0000.png,train/annotations/version1.0/P0000.txt,train,1.0
+1,train/images/P0001.png,train/annotations/version1.0/P0001.txt,train,1.0
+2,train/images/P0002.png,train/annotations/version1.0/P0002.txt,train,1.0
+3,train/images/P0000.png,train/annotations/version1.5/P0000.txt,train,1.5
+4,train/images/P0001.png,train/annotations/version1.5/P0001.txt,train,1.5
+5,train/images/P0002.png,train/annotations/version1.5/P0002.txt,train,1.5
+6,train/images/P0000.png,train/annotations/version2.0/P0000.txt,train,2.0
+7,train/images/P0001.png,train/annotations/version2.0/P0001.txt,train,2.0
+8,train/images/P0002.png,train/annotations/version2.0/P0002.txt,train,2.0
+9,val/images/P0000.png,val/annotations/version1.0/P0000.txt,val,1.0
+10,val/images/P0001.png,val/annotations/version1.0/P0001.txt,val,1.0
+11,val/images/P0000.png,val/annotations/version1.5/P0000.txt,val,1.5
+12,val/images/P0001.png,val/annotations/version1.5/P0001.txt,val,1.5
+13,val/images/P0000.png,val/annotations/version2.0/P0000.txt,val,2.0
+14,val/images/P0001.png,val/annotations/version2.0/P0001.txt,val,2.0

tests/data/dota/train/annotations/version1.0/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/train/annotations/version1.0/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/train/annotations/version1.0/P0002.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/train/annotations/version1.5/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/train/annotations/version1.5/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/train/annotations/version1.5/P0002.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/train/annotations/version2.0/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/train/annotations/version2.0/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/train/annotations/version2.0/P0002.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/val/annotations/version1.0/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/val/annotations/version1.0/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/val/annotations/version1.5/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/val/annotations/version1.5/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0

tests/data/dota/val/annotations/version2.0/P0000.txt

@@ -0,0 +1,2 @@
+imagesource:dummy
+gsd:1.0

tests/data/dota/val/annotations/version2.0/P0001.txt

@@ -0,0 +1,4 @@
+imagesource:dummy
+gsd:1.0
+10.0 10.0 20.0 12.0 18.0 20.0 8.0 18.0 plane 0
+30.0 30.0 42.0 32.0 40.0 40.0 28.0 38.0 ship 0