cs.ROAug 26, 2025

AgriChrono: A Longitudinal Dataset and Benchmark for Novel-View Synthesis of Field-Grown Crops

Authors: Jaehwan Jeong, Tuan-Anh Vu, Mohammad Jony, Shahab Ahmad, Md. Mukhlesur Rahman, Sangpil Kim, M. Khalid Jawed

Organizations: Korea University · University of California, Los Angeles · North Dakota State University

Abstract

Reliable 3D reconstruction of field-grown crops is hindered by thin, overlapping structures, wind-driven foliage motion, and changing illumination. Yet longitudinal field datasets and standardized novel-view synthesis (NVS) benchmarks remain scarce. We present an integrated framework for robotic acquisition, dataset construction, NVS evaluation, and agricultural scene reconstruction. AgriChrobot enables web-based teleoperation and repeated multi-view capture along crop rows. Using this platform, we introduce AgriChrono, a longitudinal canola dataset collected over approximately one month, comprising 160 recording sessions, four RGB streams, two depth maps, and estimated camera poses. We define seven NVS scenarios spanning four times of day and three acquisition dates, and benchmark 15 NeRF- and Gaussian Splatting-based methods. In addition to whole-image metrics, we assess fine-detail reconstruction within manually annotated canola regions using gradient- and edge-based metrics. Our results show that strong whole-image scores do not guarantee faithful reconstruction of thin, overlapping crop structures. To address this gap, we propose AgriSplat, which relaxes visibility filtering during initialization and reallocates model capacity through contribution-based pruning and budgeted reinvestment in sub-threshold candidates ranked by existing gradient statistics. After reallocation, asymmetric directional supervision penalizes insufficient contrast along reference image-gradient directions. Relative to Octree-GS, AgriSplat improves mean crop-region gradient-energy ratio from 0.506 to 0.847 and edge F-score from 0.712 to 0.797 across seven scenarios, while also improving whole-image PSNR, SSIM, and LPIPS. These contributions provide a dataset, benchmark, and method for systematic crop reconstruction under field conditions.

Figures & tables

Explore similar work

Aug 3, 2026

UAV3DCrop: Benchmarking 3D Reconstruction in Repeated Multi-Angle UAV Crop Surveys

Accurate 3D crop monitoring underpins data-driven precision agriculture by enabling field-scale analysis of plant structure, growth dynamics, and management response. Modern 3D reconstruction methods perform strongly on generic benchmarks, but rendered appearance may not translate into metrically and agronomically useful geometry in crop fields. We introduce UAV3DCrop, a public benchmark of repeated multi-angle unmanned aerial vehicle (UAV) crop surveys. It contains 88,830 RGB images at 5280×39565280 \times 3956 pixels, with a ground sampling distance of 3.6-5.8 mm, from 91 scenes spanning corn, soybean, wheat, and oat. Track A evaluates seven scene-optimized methods -- Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) variants -- on held-out views, photogrammetry-referenced depth, and canopy-height recovery. Track B tests four pretrained feed-forward models on zero-shot camera-pose and geometry estimation. The scene-optimized methods rank differently across the three targets: Splatfacto-big leads appearance, whereas Scaffold-GS leads depth and is statistically tied with Splatfacto for canopy height. Among feed-forward models, MapAnything leads on seven of the eight metrics, while the remaining models vary more across crops and fail severely on absolute scale in a way that alignment conceals. Repeated acquisitions reveal further sensitivities that differ by output type and by model, associated with position within the acquisition sequence and with tie-point multiplicity. Current 3D reconstruction methods are therefore not yet interchangeable for agronomic use: no single method wins on appearance, geometry, and canopy height at once, and only one of four feed-forward models recovers usable metric scale. The dataset is publicly available at https://link-dev.github.io/UAV3DCrop/
Sep 22, 2026

AgriGen: Large-Scale Scene Generation Framework for Photorealistic Agricultural Robotics Simulation

Agricultural robotics is advancing rapidly, yet progress remains constrained by limited field access, lack of control over field conditions, geographic variability, and seasonal crop cycles. These factors make it difficult and costly to acquire diverse agricultural datasets, resulting in limited evaluation and reduced system robustness. While other robotics domains have scaled learning and evaluation through high-fidelity simulation, agricultural robotics still lacks comparably capable tools. In this paper, we present a ROS-integrated framework, built on Isaac Sim, for large-scale procedural generation of agricultural environments. The framework supports photorealistic rendering, physics simulation, and domain randomization at scales relevant to robotics research, with built-in support for row crops, orchards, and vineyards and straightforward extensibility to additional crop categories. Project Page: https://baj31415.github.io/agrigen/
Sep 15, 2026

Evaluating Mesh Reconstruction Methods for Crop Phenotyping

Phenotyping an agricultural crop is crucial for studying its entire life cycle, as it provides vital insights to improve yield and, ultimately, food production. Doing the same for crops grown on remote sites is a challenge for the specialists who cannot be available on-site. 3D reconstruction techniques offer a promising solution to this problem by enabling crop digitization, allowing specialists to access the resulting 3D crop models from anywhere at any time. In this work, we evaluate recent 3D reconstruction pipelines for crop phenotyping. We focus on 7 mesh reconstruction pipelines and measure the fidelity and consistency of their outputs qualitatively and quantitatively. Our results suggest that the meshes produced by the GGGS, PGSR, and 2DGS are preferable to the other pipelines, owing to their quantitative metrics and visually pleasing outputs. The GGGS pipeline is better than the second-best pipeline (2DGS) by about 27% on the radar chart with 5 dimensions, namely, User ratings, Chamfer distance, LPIPS, PSNR, and SSIM.