Files and links (1)
CC BY V4.0, Open Access
Dissertation
Edge-intelligence enabled infield sensing system for heat stress mitigation in apple orchards
Washington State University
Doctor of Philosophy (PhD), Washington State University
2023
DOI:
https://doi.org/10.7273/000005350
Abstract
Heat and water stress can adversely affect apple (Malus domestica Borkh.) tree physiology and cause several fruit disorders (e.g., sunburn). Twining water and heat stress management in summer months is a key challenge to producers in the Pacific Northwest and globally. This is due to lack of reliable stress monitoring technologies and their tie up with management techniques. This dissertation addresses above challenge in-part through development of an improved crop physiology sensing system (CPSS) for heat and water stressors monitoring. The CPSS captures and computes localized weather and thermal-RGB imagery data. In objective 1, a cultivar-, color-, and size-independent mask region based convolutional neural network model was developed for edge deployment on CPSS to achieve season-long fruit surface temperature (FST) monitoring. Algorithm was effective in segmenting fruits (dice coefficient = 0.89) and estimating FST (error <0.5 ℃). FST was estimated within 37s of image capture, about 22%-time improvement over previous method. In objective 2, algorithm was further amended to estimate crop water stress index (CWSI) using hottest and coolest 2% pixels of shaded canopy temperature. Estimated CWSI effectively captured tree water stress and had negative correlation with corresponding stem water potential (r = -0.80). Algorithm’s robustness was evaluated in four heat stress mitigation techniques in commercial ‘Honeycrisp’ production, viz., cyclic conventional evaporative cooling (CEC), fogging, netting, and fognet¬—fogging and netting combined. Treatments had varying effects on CWSI estimates. Developed method found to be more reliable under fogging (r = -0.76) and less reliable under netting (r = -0.65) treatment.
Thereafter, objective 3 evaluated efficacy of all four heat stress mitigation techniques for two seasons (2021–2022) using CPSS. Localized weather and FST data indicated that CEC, fogging, and fognet techniques were effective in limiting apple FST below threshold (45 ℃). Fixed cycle frequency of CEC was unreliable in extreme stress events. Fogging had relatively higher, not significant different, sunburn percentage at harvest compared to other treatments. Post-harvest, soft scald and bitter pit incidence were highest, in CEC and fognet, respectively, compared to others. Overall, developed understanding using localized sensing data can increase operational efficacy of mitigation techniques.
Metrics
7 File views/ downloads
51 Record Views
Details
- Title
- Edge-intelligence enabled infield sensing system for heat stress mitigation in apple orchards
- Creators
- Basavaraj Rajkumar Amogi
- Contributors
- Lav Khot (Advisor)Sindhuja Sankaran (Committee Member)David Brown (Committee Member)Qin Zhang (Committee Member)
- Awarding Institution
- Washington State University
- Academic Unit
- Biological Systems Engineering, Department of
- Theses and Dissertations
- Doctor of Philosophy (PhD), Washington State University
- Publisher
- Washington State University
- Number of pages
- 191
- Identifiers
- 99901031240701842
- Language
- English
- Resource Type
- Dissertation