Third Place Winner (tie) – Hackathon 2021
Enhanced AI/ML object detection.
Our soldiers frequently use tactical mixed reality headsets to gain situational awareness in the field. One such headset, the Black Hornet, has both thermal and color imaging sensors. The thermal video feed is particularly helpful at nighttime when the color feed does not deliver much detail due to lack of light. The military uses the video feeds to perform a variety of real-time Artificial Intelligence (AI)/Machine Learning (ML) operations to detect and track objects.
Currently, soldiers can only use one video feed at a time. There are instances, however, when using both feeds simultaneously could drastically increase the precision and recall of detections. That’s why we introduced a deep learning approach of fusing the two feeds together in the CV Pod hack.
To address this challenge, we introduced a deep learning approach of fusing the two feeds together to use both the RGB spectrum and the short wave infrared spectrum to perform object detection. Using DenseFuse, we align and fuse the images to create a new one that combines the most useful features from each feed. Every edge found in the image is calculated to deliver the minimum error between edges. Then, we apply pixel correction.
Team Name: CV Pod
- John Burrows (Team Lead)
- Jonathan Cotugno
- Josh Fowler
- Joseph Amato
- Aden O’Donoghue
The result is an image that is less blurry and allows for enhanced object detection and tracking regardless of lighting conditions.
Instead of Show and Tell, we’ll Listen and Show. We’ll listen to what challenges your agency is facing. Then we’ll show you our cutting-edge prototypes and collaborate to decide which provides the best solution and the greatest value.