Deep Reality Simulation for Automated Poacher Detection | Spark Summit Europe 2018

We present a novel deep learning approach to create a robust object detection network for use in an infra-red, UAV-based, poacher recognition system. More specifically we have used Microsoft AirSim to generate thousands of hours of simulated drone footage in the African Savannah. We then used deep domain adaptation to translate our simulation into a form that is adversarially indistinguishable from real infrared drone footage. This yields a programmable data generator that can be used to dramatically improve the accuracy of algorithms without requiring expensive human curated annotations. Furthermore, we extend this work and contribute a photorealism extension to AirSim, automating much of the domain specific expertise needed for computer graphics work, and enabling the generation of limitless quantities of photorealistic data for use in reinforcement learning and autonomous vehicles.

Databricks provides a unified data analytics platform (opens in new tab), powered by Apache Spark™, that accelerates innovation by unifying data science, engineering and business.

Date:

October 10, 2018