Skip to yearly menu bar Skip to main content

Workshop: Tackling Climate Change with Machine Learning

Closing the Domain Gap -- Blended Synthetic Imagery for Climate Object Detection

Caleb Kornfein · Frank Willard · Caroline Tang · Yuxi Long · Saksham Jain · Jordan Malof · Simiao Ren · Kyle Bradbury


Object detection models have great potential to increase both the frequency and cost-efficiency of assessing climate-relevant infrastructure in satellite imagery. However, model performance can suffer when models are applied to stylistically different geographies. We propose a technique to generate synthetic imagery using minimal labeled examples of the target object at a low computational cost. Our technique blends example objects onto unlabeled images of the target domain. We show that including these synthetic images improves the average precision of a YOLOv3 object detection model when compared to a baseline and other popular domain adaptation techniques.

Chat is not available.