Summary of Weakly Supervised Framework Considering Multi-temporal Information For Large-scale Cropland Mapping with Satellite Imagery, by Yuze Wang et al.
Weakly Supervised Framework Considering Multi-temporal Information for Large-scale Cropland Mapping with Satellite Imagery
by Yuze Wang, Aoran Hu, Ji Qi, Yang Liu, Chao Tao
First submitted to arxiv on: 27 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 A weakly supervised framework for large-scale cropland mapping is proposed, utilizing multi-temporal information to reduce label costs. The approach combines high-quality labels from global land cover products with unsupervised learning signals from satellite image time series (SITS) and dense satellite images. This regularization term constrains the supervised part, alleviating overfitting issues. Additionally, the framework incorporates unsupervised learning signals in samples without high-quality labels, enriching feature space diversity. The method is experimentally validated for adaptability across three study areas: Hunan Province, Southeast France, and Kansas. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large-scale cropland mapping is crucial for agricultural production management and planning. This paper presents a new way to do this using satellite images and machine learning. It uses special labels that are easy to get from global land cover products to train the model. The approach also includes information from multiple time points in the same area, which helps to reduce mistakes and improves results. The method is tested on three different areas and shows good performance. |
Keywords
» Artificial intelligence » Machine learning » Overfitting » Regularization » Supervised » Time series » Unsupervised
