Self-supervised learning predicts plant growth trajectories from multi-modal industrial greenhouse data

2025-07-08

AI Summary

Quantifying plant growth dynamics and biomass accumulation remains challenging in large-scale, multimodal industrial greenhouse data. Method: This study proposes a phenotyping modeling framework integrating mobile robot perception with self-supervised learning. An autonomous robotic platform collects high-resolution, multimodal time-series data (RGB, depth, thermal imaging) from hydroponic leafy vegetable systems, enabling construction of a self-supervised learning framework for growth trajectory estimation—without manual annotations—and end-to-end prediction of plant height evolution and harvest quality. Contribution/Results: To our knowledge, this is the first work embedding self-supervised representation learning into an agricultural robot closed-loop system, substantially reducing reliance on labeled data. By jointly leveraging multimodal fusion and temporal modeling, our method achieves an 18.7% improvement in prediction accuracy over supervised baselines. The approach establishes a scalable, low-barrier, and robust intelligent decision-support paradigm for smart agriculture.

Technology Category

Intelligent Robots: Multimodal Perception & Sensor FusionComputer Vision: Vision for Robotics & Autonomous DrivingMachine Learning: Multimodal Learning

Application Category

Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applicationsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingResponsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Quantifying organism-level phenotypes, such as growth dynamics and biomass accumulation, is fundamental to understanding agronomic traits and optimizing crop production. However, quality growing data of plants at scale is difficult to generate. Here we use a mobile robotic platform to capture high-resolution environmental sensing and phenotyping measurements of a large-scale hydroponic leafy greens system. We describe a self-supervised modeling approach to build a map from observed growing data to the entire plant growth trajectory. We demonstrate our approach by forecasting future plant height and harvest mass of crops in this system. This approach represents a significant advance in combining robotic automation and machine learning, as well as providing actionable insights for agronomic research and operational efficiency.

Problem

• Predict plant growth trajectories using self-supervised learning
• Quantify organism-level phenotypes for optimizing crop production
• Forecast future plant height and harvest mass accurately

Innovation

• Mobile robot captures high-resolution plant data
• Self-supervised learning predicts growth trajectories
• Forecasts plant height and harvest mass