Zhang et al. (2025) RiceStageSeg: A Multimodal Benchmark Dataset for Semantic Segmentation of Rice Growth Stages

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2025
• Date: 2025-08-16
• Authors: Jianping Zhang, Tailai Chen, Qi Meng, Yanying Chen, Jie Deng, E Sun
• DOI: 10.3390/rs17162858

Research Groups

Not specified

Short Summary

The paper introduces RiceStageSeg, a multimodal UAV-based benchmark dataset of RGB and multispectral imagery for rice growth stage segmentation, demonstrating that multimodal fusion improves identification accuracy.

Objective

• To address the lack of publicly available multimodal datasets for rice growth stage identification by creating a standardized benchmark (RiceStageSeg) to support the development and evaluation of semantic segmentation models.

Study Configuration

• Spatial Scale: Centimeter-level resolution (UAV-captured imagery).
• Temporal Scale: Critical rice growth stages, specifically jointing and heading.

Methodology and Data

• Models used: State-of-the-art semantic segmentation models evaluated under unimodal (RGB-only, MS-only) and multimodal (RGB + MS feature-level fusion) configurations.
• Data sources: UAV-acquired paired RGB and 10-band multispectral (MS) images with pixel-level annotations.

Main Results

• Multimodal feature-level fusion (combining RGB and MS data) outperforms unimodal approaches in terms of segmentation accuracy for rice growth stage identification.

Contributions

• Provision of the RiceStageSeg dataset, a new standardized benchmark for the agricultural remote sensing community to advance multimodal semantic segmentation for crop monitoring.

Funding

Not specified

Citation

@article{Zhang2025RiceStageSeg,
  author = {Zhang, Jianping and Chen, Tailai and Li, Yizhe and Meng, Qi and Chen, Yanying and Deng, Jie and Sun, E},
  title = {RiceStageSeg: A Multimodal Benchmark Dataset for Semantic Segmentation of Rice Growth Stages},
  journal = {Remote Sensing},
  year = {2025},
  doi = {10.3390/rs17162858},
  url = {https://doi.org/10.3390/rs17162858}
}