Abiotic Stress Prediction from RGB-T Images of Banana Plantlets

TL;DR

This paper introduces neural network methods for predicting abiotic stress in banana plantlets using RGB and thermal images, achieving over 90% accuracy in classifying water and fertilizer treatments.

Contribution

It presents novel strategies for abiotic stress prediction using multi-modal RGB-T images and demonstrates high accuracy on a new dataset of banana plantlets.

Findings

Neural networks achieved over 90% accuracy in stress classification.

Multi-modal RGB-T data improves prediction over single modalities.

Methods outperform expert assessments in distinguishing treatments.

Abstract

Prediction of stress conditions is important for monitoring plant growth stages, disease detection, and assessment of crop yields. Multi-modal data, acquired from a variety of sensors, offers diverse perspectives and is expected to benefit the prediction process. We present several methods and strategies for abiotic stress prediction in banana plantlets, on a dataset acquired during a two and a half weeks period, of plantlets subject to four separate water and fertilizer treatments. The dataset consists of RGB and thermal images, taken once daily of each plant. Results are encouraging, in the sense that neural networks exhibit high prediction rates (over $90\%$ amongst four classes), in cases where there are hardly any noticeable features distinguishing the treatments, much higher than field experts can supply.