Multi-Class Aircraft Detections

Using multiple sources of geospatial data and artificial intelligence, we deliver timely, objective and transparent analytics on a global scale. The Orbital Insight GO platform is versatile and designed to answer your specific question about the physical world. Within the GO Platform, a user can choose from Object Detection and Land Use algorithms.

Object detection algorithms include:

• Cars
• Trucks
• Ships
• Multi-Class Aircraft

Land use algorithms include:

• Land use aggregation classification
• Land use change detection

Orbital Insight’s Multi-Class Aircraft Detection algorithm detects, classifies and quantifies aircraft from wide-area satellite detection at scale. By leveraging large datasets of commercial imagery and pre-trained AI algorithms, Orbital Insight is able to establish baseline order of battle, anomalies, and indications and warnings on a global scale.

Orbital Insight's Multi-Class Aircraft Detector currently operates on Digital Globe High-Resolution imagery. It produces bounding boxes corresponding to the location of aircraft within a specified input image (GeoTiff). The bounding box is rectangular and vertically oriented. This algorithm currently classifies Aircraft into four aircraft classes: Fighter, Bomber, Commercial/Passenger, and Other.

This listing consists of two datasets.

• OI_AOI - Military and Dual-Use Airfields comprises of 342 Russian, Chinese, and Middle Eastern airfield/dual-use airport polygons in KML format.
• OI_Results - Multi-Class Aircraft Detections 2017 comprises of Go-derived Multi-Class Aircraft detections for those airfields for 2017.

GO-derived data exists in CSV format for this product offering. With a GO subscription, data is also available via API. Metadata is structured for enabling further analysis in both statistical and geographical information system (GIS) workflows.

A selection of field values include:

• Area of Interest (AOI) Name
• AOI Scene Intersect (km2)
• Measurement Timestamp (UTC)
• Local Timestamp
• Imagery Scene ID
• Off Nadir Angle
• Imagery Provider ID
• Object Count
• Object Geometry (in Well-known-Text format)
• Aircraft Class
• Cloud Score
• Time Zone ID