Visual AI on the Edge and in the Cloud Easy Access to Earth - PowerPoint PPT Presentation

Visual AI on the Edge and in the Cloud

Easy Access to Earth Observation Data

Problems We Solve What vehicle arrived at the location? Was the patient anesthetized? How many cells are on this slide? Should we give this person access?

Machine Learning Process • Chooch is easy to train • Models for each type of visual data. • Classify objects within an image, an image itself, actions, anomalies, faces, etc. • Patented machine learning system.

Cloud AI Features • Dashboard • API • Deployable on edge • Object, Image, Facial Recognition Training • AI Training As A Service • ~200,000 public classes

Use Case: Synthetic Aperture Radar Identifying objects on the ground with 99%+ accuracy.

Use Case: Geospatial Imagery Real time wildfire alerts.

SkyWatch EarthCache™ THE EASIEST WAY TO GET SATELLITE DATA FOR YOUR APPLICATION • Aggregates data from low, medium, and high resolution sources. • Pay-as-you-go data model makes it affordable. • With one API call, you can go through past data archives or set up monitoring over your Area of Interest (AOI) going forward. EarthCache returns the best available data for each specified interval. • Request access: www.skywatch.com.

SkyWatch EarthCache™

SkyWatch EarthCache™ WHAT RESOLUTION DO I NEED? • 10 m resolution means each pixel represents (roughly) a 10 x 10 m area on the ground (33 x 33 ft). • For large scale events, such as wildfires, floods, and city growth, we recommend low resolution, open data (10 m or 15 m). • For construction monitoring, forestry, or mining, medium resolution (3 – 5m) is better suited. • For object detection, high resolution (50 – 80 cm) is recommended.

DEMO DEMO Geo eospatial Training & Det etec ection Thank You

Q&A Q&A Thank You

Thank You chooch.ai/demo