Home

Datasets

A list of image / video datasets used to train a military vehicle detection algorithm

• The Search 2 dataset consists of 44 high-resolution digital color images of different complex natural scenes. Each scene (image) contains a single military vehicle that serves as a search target.
• Sensor Data Management System (SDMS). The video sample sets #1 and #2 contain video clips from the DARPA VIVID Data Collection 1 collected at Eglin, AFB December 2nd and 3rd 2003. The clips are taken from three sensors: EO Daylight TV(DLTV), EO DLTV Spotter, and IR. The scenes contain military and civilian vehicles in many settings. All clips are approximately 60 seconds each and are in AVI format.
• Military vehicles datasets from Kaggle:
  • Military Tanks
  • Military Vehicles
  • Normal vs Military Vehicles
  • War Tank Images
• Open Images V7. Open Images is a dataset of ~9M images annotated with image-level labels, object bounding boxes, object segmentation masks, visual relationships, and localized narratives. It contains a total of 16M bounding boxes for 600 object classes on 1.9M images.
• ImageNet. ImageNet, also known as ILSVRC 2012, is an image dataset organized according to the WordNet hierarchy. Each meaningful concept in WordNet, possibly described by multiple words or word phrases, is called a “synonym set” or “synset”. There are more than 100,000 synsets in WordNet, majority of them are nouns (80,000+). ImageNet provides on average 1,000 images to illustrate each synset.
• Army Recognition Database of world army equipment and recognition information.

Projects & papers

• Identify Military Vehicles in Satellite Imagery with TensorFlow
• Military Vehicles Tracking
• Automated Military Vehicle Detection From Low-Altitude Aerial Images
• Military Vehicles CNN
• Fine-tune YOLOv8 models for custom use cases with the help of FiftyOne

Classes

• Vehicle or Non-Vehicle
• Vehicle: Military or Non-Military
• Vehicle - Military:
  • MILITARY ARMOURED
  • HEAVY EXPANDED MOBILITY TACTICAL TRUCK (HEMTT)
  • MILITARY TRUCK
  • HIGH MOBILITY MULTI-PURPOSE WHEELED VEHICLE
  • MILITARY CAR
  • MILITARY AMBULANCE
  • Self-propelled artillery
  • Self-propelled anti-aircraft
  • Amphibious armored scout car
  • Armored personnel carrier
  • Military cargo truck
  • Main battle tank

Data Augmentation

• PyTorch transform package: https://pytorch.org/vision/stable/transforms.html . Specifically, RandomZoomOut and ScaleJitter
• Yolov8 already applies random data augmentation when training : https://github.com/ultralytics/ultralytics/blob/main/ultralytics/data/augment.py (see https://github.com/ultralytics/ultralytics/issues/2701 and https://github.com/ultralytics/yolov5/issues/7572 on how to configure / manage data augmentations with yolo training)
• Albumentations lib