This package is a pip-installable version of the support/inference code for MegaDetector, an object detection model that helps conservation biologists spend less time doing boring things with camera trap images. Complete documentation for this Python package is available at megadetector.readthedocs.io.
If you aren't looking for the Python package specifically, and you just want to learn more about what MegaDetector is all about, head over to the MegaDetector repo.
If you don't want to run MegaDetector, and you just want to use the utilities in this package - postprocessing, manipulating large volumes of camera trap images, etc. - you may want to check out the megadetector-utils package, which is identical to this one, but excludes all of the PyTorch/YOLO dependencies, and is thus approximately one zillion times smaller.
Install with:
pip install megadetector
MegaDetector model weights aren't downloaded at the time you install the package, but they will be (optionally) automatically downloaded the first time you run the model.
See megadetector.readthedocs.io.
from megadetector.utils import url_utils
from megadetector.visualization import visualization_utils as vis_utils
from megadetector.detection import run_detector
# This is the image at the bottom of this page, it has one animal in it
image_url = 'https://github.com/agentmorris/MegaDetector/raw/main/images/orinoquia-thumb-web.jpg'
temporary_filename = url_utils.download_url(image_url)
image = vis_utils.load_image(temporary_filename)
# This will automatically download MDv5a; you can also specify a filename.
model = run_detector.load_detector('MDV5A')
result = model.generate_detections_one_image(image)
detections_above_threshold = [d for d in result['detections'] if d['conf'] > 0.2]
print('Found {} detections above threshold'.format(len(detections_above_threshold)))
from megadetector.detection.run_detector_batch import \
load_and_run_detector_batch, write_results_to_file
# Choose a folder to run MD on recursively, and an output file
image_folder = '/path/to/megadetector_test_images'
output_file = '/path/to/output/file.json'
# The package will automatically download whichever model you request; you
# can also specify a filename.
model_name = 'MDV5A'
# Run the model on all images in the folder
results = load_and_run_detector_batch(model_name, image_folder)
# Write results to a format that Timelapse and other downstream tools like
write_results_to_file(results,
output_file,
relative_path_base=image_folder,
detector_file=model_name)
Contact cameratraps@lila.science with questions.
Image credit University of Minnesota, from the Orinoquía Camera Traps data set.