Airbus Computer Vision Image Annotation Case Study

“ We are pleased with the high quality and speed of the delivered results and are happy to have engaged in this collaboration with Appen.” – Aleix Canet, Product Owner and Team Lead at Airbus

Overview: Data for Flight Forecasting

AirSense, an advanced analytics solution enabled by the defense and space division of Airbus, aims to increase the accuracy of their estimated flight times and time of arrivals leveraging historical flight path data.

Challenge: Annotating Data with Accuracy

To reach a high level of accuracy, Airbus needed to annotate thousands of flight paths based on historical GPS data. Each flight path consists of a sequence of coordinates called a ‘cluster,’ and those coordinates needed to be labeled in the correct order to ensure accuracy and reduce the margin of error. A cluster also has several variations due to the various flight paths; short-haul, long-haul, diverted routes, etc. The project required speed, scale and quality to deliver success.

Solution: Quality Data Annotation with Speed

Our Appen Data Annotation Platform provided Airbus with the quality and speed needed to scale their computer vision project. Airbus trained our onsite team of annotators through video and constant feedback to ensure high-accuracy of image annotations. The tasks required annotators to map the beginning, middle, and end of the flight path. They also needed to coordinate a position on the image with a GPS coordinate. The platform’s quality assurance tools allowed the Airbus team to review the data quickly and accurately.

Result: Successful Pilot Launch

Using Appen’s data annotation with our onsite annotation services, Airbus was able to securely and accurately annotate thousands of images. due to the collaboration with Airbus and Appen teams, the pilot project successfully launched with 98% accuracy.