AV_Halo CORTEX, Telesto

In partnership with multiple satellite data providers, Telesto offers automated monitoring services for known facilities and transportation networks, and cost-effective broad area tasking and searching for discovery of unknown structures and sites, and tracking site changes.

Deep learning-based perceptual harmonization creates a temporal datacube of satellite images drawn from multiple satellite sources. Broad area search and characterization algorithms detect and classify heavy construction activities at global scale. The system distinguishes natural changes from anthropogenic activities for GEOINT analysis.

Leveraging the state of the art in AI/ML models, Telesto provides an extensive object-centric analysis suite including object detection and characterization, mensuration, change detection, movement analysis and custody maintenance.

The interplay between objects, structures, surroundings, environmental context and their relationships emerges in site-centric analysis. Telesto provides transparent site and asset monitoring capability using EO, SAR and non-imaging data sources.

Automated image geolocalization estimates the latitude, longitude and time of capture images taken from ground view (e.g., cellphone camera) or drone view without requiring any metadata.

SAR to EO conversion capability vastly improves image palatability to non-experts in SAR and enables rapid multi-modal analysis.

For a given location, the cross-view synthesis capability generates likely ground view images using only satellite image as its input. Using this capability, analysts can visualize likely ground camera views in places with missing street view imagery.

Using just a few images, the novel view renderer understands the 3D scene and generates views of the structure / scene from unseen directions/locations. The neural modeling capability has revolutionized image-based interpolation and 3D scene understanding.