Accenture/AIR
A deep learning object detector framework written in Python for supporting Land Search and Rescue Missions.
Built on RetinaNet architecture with ResNet152 backbone, AIR implements custom bounding box aggregation (MOB algorithm) and SAR-APD evaluation metrics tailored for aerial person detection rather than generic object detection. The framework processes both high-resolution images and drone video feeds with integrated Kalman filter tracking, achieving 90% precision on the HERIDAL benchmark while maintaining real-time inference speed. Deployable via Docker/Singularity containers or native Python installation with optional GPU acceleration for production SAR operations.
No commits in the last 6 months.
Stars
37
Forks
13
Language
Python
License
Apache-2.0
Category
Last pushed
Mar 15, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/Accenture/AIR"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
jonasrenault/orion
Orion - Automated Target Recognition of Military Vehicles in Video Data
roboflow/dji-aerial-georeferencing
Detect objects in drone videos and plot them on a map
atulyakumar97/fire-and-gun-detection
Fire and Gun detection using yolov3 in videos as well as images. Training code, dataset and...
hmakelin/gisnav
Estimates airborne drone global position by matching video to map retrieved from onboard GIS server.
Orange-OpenSource/marine-detect
Object detection models for identifying species in marine environments.