WalnutData: A UAV Remote Sensing Dataset of Green Walnuts and Model Evaluation

2025-02-27

AI Summary

To address the scarcity of high-quality remote sensing datasets for green walnuts in agricultural computer vision, this paper introduces WalnutData—the first dedicated UAV-based remote sensing dataset for green walnuts, comprising 30,240 images and 706,208 annotated instances. We propose a high-granularity annotation paradigm that systematically models the combined effects of illumination (front-lit vs. back-lit) and occlusion (occluded vs. non-occluded), yielding four distinct subsets. Furthermore, we establish the first authoritative benchmark and unified evaluation protocol for this task, conducting comprehensive experiments using state-of-the-art detectors including YOLOv5, YOLOv8, and Faster R-CNN. Both the full dataset and benchmark results are publicly released, significantly improving detection accuracy and robustness under complex field conditions. The resource has garnered substantial attention on GitHub.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Efficient ML / Green AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity

Abstract

The UAV technology is gradually maturing and can provide extremely powerful support for smart agriculture and precise monitoring. Currently, there is no dataset related to green walnuts in the field of agricultural computer vision. Thus, in order to promote the algorithm design in the field of agricultural computer vision, we used UAV to collect remote-sensing data from 8 walnut sample plots. Considering that green walnuts are subject to various lighting conditions and occlusion, we constructed a large-scale dataset with a higher-granularity of target features - WalnutData. This dataset contains a total of 30,240 images and 706,208 instances, and there are 4 target categories: being illuminated by frontal light and unoccluded (A1), being backlit and unoccluded (A2), being illuminated by frontal light and occluded (B1), and being backlit and occluded (B2). Subsequently, we evaluated many mainstream algorithms on WalnutData and used these evaluation results as the baseline standard. The dataset and all evaluation results can be obtained at https://github.com/1wuming/WalnutData.

Problem

• Lack of green walnut dataset in agricultural computer vision.
• UAV-collected dataset addresses varying lighting and occlusion.
• Evaluates mainstream algorithms on new WalnutData baseline.

Innovation

• UAV collected high-granularity walnut data
• Constructed WalnutData for algorithm evaluation
• Evaluated mainstream algorithms on green walnuts