Summary of Comparing Surface Landmine Object Detection Models on a New Drone Flyby Dataset, by Navin Agrawal-chung and Zohran Moin
Comparing Surface Landmine Object Detection Models on a New Drone Flyby Dataset
by Navin Agrawal-Chung, Zohran Moin
First submitted to arxiv on: 17 Oct 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper explores the application of deep learning-based object detection algorithms for detecting small, soda-can size surface landmines using drone videos. The traditional methods used for landmine detection are slow, dangerous, and expensive, making this approach promising but challenging due to the unique characteristics of the landmines. To overcome these challenges, a custom dataset was created comprising drone images of Russian surface landmines (POM-2 and POM-3). This dataset was then used to train, test, and compare four different computer vision foundation models: YOLOF, DETR, Sparse-RCNN, and VFNet. The results showed that all four detectors performed well, with YOLOF outperforming the others with a mean Average Precision (mAP) score of 0.89. Additionally, YOLOF was found to be quicker to train, consuming only 56 minutes on a Nvidia V100 compute cluster. This research contributes a custom dataset and model Jupyter notebooks to enable future research in surface landmine detection. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Landmines are dangerous and expensive to detect using traditional methods. Scientists are trying to find new ways to use computers to help spot them from the air. They took videos of small landmines with drones and used special computer programs called deep learning models. These models looked at the pictures and tried to find the landmines. Some models worked better than others, but one called YOLOF was the best. It found 89% of the landmines correctly. This is important because it could help make detecting landmines faster and cheaper. |
Keywords
» Artificial intelligence » Deep learning » Mean average precision » Object detection » Rcnn
