What is a contrail?

What is a contrail?

Contrails, often overlooked, pose a significant challenge in our skies. Recent research reveals their surprising impact, they are responsible for 60% of aviation-related global warming and causing twice as much warming as direct CO2 emissions from aircraft. In the US alone, contrails blanket an area equivalent to Massachusetts and Connecticut combined.

About Our Project

Our mission is to help airlines and aircraft makers meet climate targets by reducing contrail creation.
We contribute to a quick, low-cost, and high impact solution of contrail avoidance in flight software planning tools by helping researchers improve their predictions.

satelite

What We Provide

We offer a trained deep learning algorithm for precise pixel-level contrail segmentation in satellite images.

Why We Do It

Research indicates that diverting Only 1.7% of flights could reduce contrail climate damage by up to 59%. Contrail reduction also creates a new $9 billion carbon market forecasted to rise tenfold by 2030. Airlines have been enthusiastic about the subject as the benefits outweight costs.

How To Use It

Data scientists can use our model to analyze curated historical satellite images at specific locations and timestamps to identify the presence of contrails. This empowers data scientists working in the area of contrail prediction and avoidance by providing automatic contrail detection.

Who We Serve

We serve the needs of data scientists at commercial aircraft manufacturers and flight software companies looking to predict contrail formation, evaluate and develop contrail avoidance systems.

Our Data

We are using data from the paper: OpenContrails: Benchmarking Contrail Detection on GOES-16 ABI by Joe Yue-Hei Ng et. al. taken from the Kaggle Competition.

Independent Variables

Satellite images from the GOES-16 ABI satellite which orbits the North and South American continent over April 2019-2020 (450GB):

  • More than 22,000 records each containing bands 8-16. Each band file contains 8 timesteps.
  • The data is 256x256x8x9 dimensions per record.
  • Each image represents an area 500km x 500km and each pixel is 2km x 2km.

Dependent Variables

Prediction labels containing binary labels identifying existence of contrails at pixel level from individual human labelers and the final ground truth labels:

  • Only 1.2% of mean positive pixels in training datas

Metadata

Metadata shows the associated projection parameters such as the central meridian and timestamps of the satellite images.

  • Images are taken at 10-min intervals.

Data Processing

The data has been processed by the researchers to downsample the negative instances by using:

  • Aircraft flight tracks, wind data, humidity and temperature information
  • The Mannstein contrail detection algorithm to find more positive samples
  • Google Street View images of locations with sky view containing a contrail

Here is a visual representation of one data sample and how we processed it:

False color image RGB channels were created by plotting brightness temperature contrasts using feature transformation of 3 bands.

Here is an example of a plot of one-single time step of an image (note there are 8 timesteps in the dataset):

Here is an example of contrails under one record moving and changing in shape over time:

Our Model

TransUNet built on a hybrid Transfomer and UNet Architecture.
Combines the best of both worlds in learning global context and local information.



We used Mean Intersection Over Union (IOU) as the main metric to evaluate the model performance.

A model with Mean IoU Score of 0.7 to 0.8 is considered to be Very Good
We aimed to reduce the False positives to reduce the impact of misleading predictions

Results

We Used UNet, Resnet101 and DeeplabV3 as baseline models, and a hybrid TransUnet using the original UNet architecture but with an additional Transformer Cycle

MVP vs Basline Model metrics summary

Model Validation Mean IoU Validation Dice Coefficient False Positive Ratio False Negative Ratio
MVP - TransUNet 0.6997 0.7718 0.0988 0.1653
Baseline - UNet 0.6951 0.7642 0.1550 0.1422

Demo

We hosted our model on Huggingface with test satellite images where we can detect contrails in real-time

Our team

We are a team of graduate students in UC Berkeley's Master of Information and Data Science (MIDS) Program.
We have expertise in computer vision, data engineering/processing, and user/domain research.

Ziling Huang

Research Lead

Prakash Krishnan

Project Lead

Rebecca Nissan

Product Lead

Sitao Chen

Data Engineering Lead

Pedro Melendez

Machine Learning Lead

Acknowledgments

We would like to express our sincere thanks to the following people who were integral to our project:

  • Thank you to our Professors Alberto Todeschini and Professor Joyce Shen for helping us through every step of the way and providing constructive feedback.
  • Thank you to experts we consulted - Dr. Sanjay Krishnan, Assistant Professor of Computer Science, University of Chicago and Dr. Namrata Anand, CEO Diffuse Bio.
  • Thank you to Mary Green (Tazzie) for offering us advice on Google Earth Engine usage.