Places in the Wild: Ecologically-sampled RAW photographs

Places in the Wild comprises over 67,000 RAW-format images, each captured with a 45-megapixel Canon EOS R5 full-frame mirrorless camera at 5-degree intervals, providing 360-degree coverage across over 800 unique locations. These locations span 260 basic-level scene categories, including both indoor and outdoor environments such as bedrooms, train stations, forests, and parking garages. The dataset is hosted on Academic Torrents: https://academictorrents.com/details/a1d810fbb54c21dfe8a68538d917c668e48663de Working with RAW files Places in the Wild is in Canon RAW format (.CR3). These can easily be converted to JPG. Method 1: rawpy + OpenCV (Python) import rawpy import cv2 with rawpy.imread("input.cr3") as raw: rgb = raw.postprocess(use_camera_wb=True, output_bps=8) bgr = cv2.cvtColor(rgb, cv2.COLOR_RGB2BGR) cv2.imwrite("output.jpg", bgr, [cv2.IMWRITE_JPEG_QUALITY, 95]) Method 2: rawpy + Pillow (Python) import rawpy from PIL import Image with rawpy.imread("input.cr3") as raw: rgb = raw.postprocess( use_camera_wb=True, output_bps=8, no_auto_bright=False ) img = Image.fromarray(rgb) img.save("output.jpg", quality=95) Method 3: ImageMagick (requires dcraw or ufraw as delegate) # Convert a single file magick input.cr3 output.jpg # Batch convert an entire directory magick mogrify -format jpg -quality 95 *.cr3