Thesis
DEVELOPMENT AND EVALUATION OF A SMART CROP STRESS MONITORING AND MANAGEMENT SYSTEM FOR APPLICATIONS IN CONTROLLED ENVIRONMENTS
Master of Science (MS), Washington State University
2025
Abstract
Crop stress management is important in controlled environment agriculture as stress can lead to lower crop productivity and potential yield loss if not detected and addressed early. Current strategies to manage stress often need highly trained workers and control procedures and rely on intuition and expertise. This study introduces a prototype of a real-time and automated decision-support system for mist-based chemical applications to manage plant stress at the plant level in controlled environment. The system consists of three modules: an RGB imaging sensor module, a control module to operate and process the data in the cloud, and a sprayer module to apply the chemical solution. The imaging sensor module captures images of a group of plants, which are sent to the cloud for segmentation, mapping, and stress analysis. We performed an experiment on peas (Pisum sativum L.) by inducing nutrient-related stress, referred to as Dataset 1, and contains 386 healthy and 124 stressed crops. We used YOLOv11 algorithm to generate individual masked images of plants and a neural network classifier as the decision-making model for each image. If stress is detected, the microcontroller initiates a spray application. Our segmentation and detection model has achieved a high mean average precision (mAP) of about 93% while the decision-making model’s accuracy reached around 81% on Dataset 1. We also trained a classifier to explore applicability of the developed system in other crops using a publicly available dataset. We achieved an accuracy of 91% using a dataset of images (Dataset 2) of lettuce grown in vertical farming, comprising 1162 images of healthy crops and 768 images of stressed plants. The abiotic stress conditions in Dataset 2 included chlorosis, tip burn, and wilting. These results demonstrate a viable and adaptable approach for early, precise, and automatic plant stress control, which can be integrated into controlled environment management to optimize operations and, consequently, crop productivity and profitability
Metrics
1 Record Views
Details
- Title
- DEVELOPMENT AND EVALUATION OF A SMART CROP STRESS MONITORING AND MANAGEMENT SYSTEM FOR APPLICATIONS IN CONTROLLED ENVIRONMENTS
- Creators
- Toky Fanantenana Andriamihajasoa
- Contributors
- Sindhuja Sankaran (Advisor)Sindhuja Sankaran (Committee Member)Rebecca McGee (Committee Member)Manoj Karkee (Committee Member)
- Awarding Institution
- Washington State University
- Academic Unit
- College of Agricultural, Human, and Natural Resource Sciences
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Number of pages
- 80
- Identifiers
- 99901297094701842
- Language
- English
- Resource Type
- Thesis