Roboflow | Which Image Augmentation Steps Should You Use with Aerial Data? @Roboflow | Uploaded 3 years ago | Updated 6 hours ago
Image augmentation is an effective way to get more out of your data with computer vision. When working with aerial data - like from satellites or drones - there's a few common augmentation steps you can use that will increase the size of your dataset and improve model generalizability (performance) in the process.
Using Roboflow: roboflow.com
More about augmentation/preprocessing: blog.roboflow.com/why-preprocess-augment
Tackling the small object problem in computer vision: blog.roboflow.com/detect-small-objects
✅ Subscribe: bit.ly/rf-yt-sub
Image augmentation is an effective way to get more out of your data with computer vision. When working with aerial data - like from satellites or drones - there's a few common augmentation steps you can use that will increase the size of your dataset and improve model generalizability (performance) in the process.
Using Roboflow: roboflow.com
More about augmentation/preprocessing: blog.roboflow.com/why-preprocess-augment
Tackling the small object problem in computer vision: blog.roboflow.com/detect-small-objects
✅ Subscribe: bit.ly/rf-yt-sub
![[LIVE CODING] Visualizing Football Plays with Computer Vision (Part 1)](https://i.ytimg.com/vi/QI1dQF1iw0c/hqdefault.jpg "[LIVE CODING] Visualizing Football Plays with Computer Vision (Part 1)
In this live session well use the Football Player Detection dataset[1] by Scott Bronkema[2] shared on Roboflow Universe to create a program that tracks players movements using zero-shot object detection[3] in a football play and visualizes the result.
Tools well use/cover include: Python, Roboflow, YOLOv4, YOLOv5, CLIP, zero shot object tracking, ffmpeg, and OpenCV.
Update: we sent an improved dataset to training at the end of the video; well pick up here once the improved model is finished training with part 2. Subscribe and ding the bell to be notified!
[1] https://universe.roboflow.com/bronkscottema/football-player-detection
[2] https://www.linkedin.com/in/scott-bronkema-09389824/
[3] https://blog.roboflow.com/zero-shot-object-tracking/")
