# 4D Crop Monitoring: Spatio-Temporal Reconstruction for Agriculture

arXiv: https://arxiv.org/abs/1610.02482 · HTML: https://mlsanity.com/paper/1610.02482/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Oct 2016.

## Built on by (most cited first, 8 of 8)

- [CanFields: Consolidating Diffeomorphic Flows for Non-Rigid 4D Interpolation from Arbitrary-Length Sequences](https://mlsanity.com/paper/2406.18582/) (2024)
- [GrowSplat: Constructing Temporal Digital Twins of Plants with Gaussian Splats](https://mlsanity.com/paper/2505.10923/) (2025)
- [Statistical shape representations for temporal registration of plant components in 3D](https://mlsanity.com/paper/2209.11526/) (2022)
- [Grow with the Flow: 4D Reconstruction of Growing Plants with Gaussian Flow Fields](https://mlsanity.com/paper/2602.08958/) (2026)
- [Transformer-Based Spatio-Temporal Association of Apple Fruitlets](https://mlsanity.com/paper/2503.03200/) (2025)
- [Advances in 4D Generation: A Survey](https://mlsanity.com/paper/2503.14501/) (2025)
- [How to rewrite the stars: Mapping your orchard over time through constellations of fruits](https://mlsanity.com/paper/2602.04722/) (2026)
- [GrowFields: Compositional 4D Neural Fields for Topology-Changing Plant Growth](https://mlsanity.com/paper/2607.03330/) (2026)

## Abstract (arXiv)

Autonomous crop monitoring at high spatial and temporal resolution is a critical problem in precision agriculture. While Structure from Motion and Multi-View Stereo algorithms can finely reconstruct the 3D structure of a field with low-cost image sensors, these algorithms fail to capture the dynamic nature of continuously growing crops. In this paper we propose a 4D reconstruction approach to crop monitoring, which employs a spatio-temporal model of dynamic scenes that is useful for precision agriculture applications. Additionally, we provide a robust data association algorithm to address the problem of large appearance changes due to scenes being viewed from different angles at different points in time, which is critical to achieving 4D reconstruction. Finally, we collected a high quality dataset with ground truth statistics to evaluate the performance of our method. We demonstrate that our 4D reconstruction approach provides models that are qualitatively correct with respect to visual appearance and quantitatively accurate when measured against the ground truth geometric properties of the monitored crops.