https://deeplearning.earth/categories/archive/Recent content in archive on DeepLearning.EarthHugo -- gohugo.ioen-usThu, 16 Nov 2023 15:11:48 +0000https://deeplearning.earth/posts/2023-11-16_detecting_ships_in_satellite_imagery_five_years_later/Thu, 16 Nov 2023 15:11:48 +0000https://deeplearning.earth/posts/2023-11-16_detecting_ships_in_satellite_imagery_five_years_later/Context Link to heading In 2018, when I was still working at Airbus Defence and Space, I organised a challenge on Kaggle to detect ships in Airbus SPOT satellite imagery (@ 1.5 meters resolution). Home page of the Airbus Ship Detection Challenge on Kaggle Check my article about the successes and issues that we encountered. That’s a whole story in itself! One of the interesting characteristics of the challenge was the oriented bounding boxes annotations.https://deeplearning.earth/posts/2023-02-08_is_yolo8_suitable_for_satellite_imagery/Wed, 08 Feb 2023 15:11:48 +0000https://deeplearning.earth/posts/2023-02-08_is_yolo8_suitable_for_satellite_imagery/The latest YOLO version has been published Link to heading YOLOv8 is the latest version of the YOLO object detection and image segmentation models developed by Ultralytics. YOLOv8 is a cutting-edge, state-of-the-art (SOTA) model that builds upon the success of previous YOLO versions and introduces new features and improvements to further boost performance and flexibility. Performances of the YOLO series (source) YOLOv8 is designed to be fast, accurate and user-friendly, making it a popular choice among researchers and practitioners in computer vision and AI.https://deeplearning.earth/posts/2023-01-02_how-to-choose-a-deep-learning-architecture-to-detect-aircrafts-in-satellite-imagery/Mon, 02 Jan 2023 15:11:48 +0000https://deeplearning.earth/posts/2023-01-02_how-to-choose-a-deep-learning-architecture-to-detect-aircrafts-in-satellite-imagery/Context Link to heading In recent years, artificial intelligence has made great strides in the field of computer vision. One area that has seen particularly impressive progress is object detection, with a variety of deep learning models achieving high levels of accuracy. However, this abundance of choice can be overwhelming for practitioners who are looking to implement an object detection system. On top of this, most public models and academic research are benchmarked on COCO which are dataset made of photographs.https://deeplearning.earth/posts/2022-06-15_oil-storage-detection-on-airbus-imagery-with-yolox/Wed, 15 Jun 2022 19:17:31 +0000https://deeplearning.earth/posts/2022-06-15_oil-storage-detection-on-airbus-imagery-with-yolox/Introduction Link to heading Last year, Airbus Intelligence published a few Machine Learning Datasets on the Kaggle platform. These datasets are samples from much larger and more comprehensive datasets provided by Airbus. Nevertheless, they are good datasets to start with and build upon if you wish to learn more about Earth Observation imagery and Deep Learning. In this article, we will analyse the Airbus Oil Storage dataset. It contains one hundred SPOT images and a little over 13,500 annotated POL (Petroleum, Oil and Lubricant) storage.https://deeplearning.earth/posts/2021-09-16_detecting-aircraft-on-airbus-pleiades-imagery-with-yolov5/Thu, 16 Sep 2021 15:11:48 +0000https://deeplearning.earth/posts/2021-09-16_detecting-aircraft-on-airbus-pleiades-imagery-with-yolov5/Introduction Link to heading Recently Airbus Intelligence has published a few Machine Learning Datasets on the Kaggle platform. These datasets are samples from much larger and more comprehensive datasets provided by Airbus. Nevertheless, they are good datasets to start with and build upon if you wish to learn more about Earth Observation imagery and Deep Learning. In this article, we will analyse the Airbus aircraft dataset. It contains one hundred civilian airports and a little over 3,000 annotated commercial aircrafts.https://deeplearning.earth/posts/2020-03-08_important-things-you-should-know-before-organizing-a-kaggle-competition/Sun, 08 Mar 2020 20:45:30 +0000https://deeplearning.earth/posts/2020-03-08_important-things-you-should-know-before-organizing-a-kaggle-competition/Lessons Learned from the Airbus Ship Detection Challenge Link to heading In 2018, my team at Airbus Defence and Space Intelligence and I organized a machine learning challenge on the well-known Kaggle platform (see https://www.kaggle.com/c/airbus-ship-detection). At the time, there were discussions internally about the interest of such an initiative, the concrete benefits that we could gain from it, and the potential risks of open-sourcing our imagery. One year later, we have no doubt that it was worth it.