Design and Implementation of a Multimodal, Multiscale Non-Invasive Bioelectrical Measurement System for Integrative and Proactive Plant Assessment

Author

• Jonnel Dorado Alejandrino, De La Salle University, ManilaFollow

Date of Publication

2025

Document Type

Dissertation/Thesis

Degree Name

Doctor of Philosophy in Electronics and Communications Engineering

College

Gokongwei College of Engineering

Department/Unit

Electronics And Communications Engg

Honor/Award

Outstanding Dissertation Award

Thesis Advisor

Elmer P. Dadios

Defense Panel Chair

Edwin Sybingco

Defense Panel Member

Argel A. Bandala
Ryan Rhay P. Vicerra
Laurence A. Gan Lim
Raouf N.G. Naguib

Abstract (English)

Global crop losses of 20–40% persist because traditional plant characterization is either invasive, causing damage, or non-invasive, which relies on visual traits, making it reactive and symptom-based. Advanced bioelectrical approaches overcome these constraints; however, existing systems remain limited in terms of operational modes, frequency ranges, and organ-specific applicability. This study presents a multimodal, multiscale, and integrative bioelectrical system designed to address these limitations. The system incorporates a modified analog signal processing module (MASPM) and different electrode configurations, enabling both spectroscopic measurements and temporal conductivity imaging. Implementation and validation were conducted on the strawberry plant (Fragaria×ananassa) across four experiments: (1) leaf water-status classification, (2) market-based fruit categorization, (3) stem-productivity classification, and (4) temporal root imaging. For each organ, optimal frequencies were identified and used in feature extraction. These features served as inputs to classification models, where TabPFN outperformed CatBoost, LightGBM, and XGBoost, achieving 98.3%±1.0% accuracy for fruit categorization, 95.8%±1.6% for leaf water status, and 88.2%±1.9% for stem productivity. For root imaging, reconstruction algorithms produced a spatial resolution of 2.6–2.8 mm. Validation results showed strong correlations between optimal features and physiological reference metrics, ranging from r = 0.87 to 0.92 for fruits, r = 0.78 to 0.83 for leaves, and r = 0.66 to 0.71 for stems, confirming that the bioelectrical measurements accurately reflected physiological variations. This developed system addresses longstanding limitations in plant monitoring and establishes a foundation for proactive, non-invasive plant characterization, supporting long-term goals in sustainable agricultural management.

Abstract Format

html

Abstract (Filipino)

Patuloy ang 20–40% na pandaigdigang pagkalugi sa ani dahil ang tradisyunal na pagsusuri ng halaman ay maaaring invasive, na nakakasira ng tisyu, o non-invasive na nakabatay lamang sa panlabas na anyo, kaya nagiging reaktibo. Nalulutas ito sa pamamagitan ng bioelectrical na pagsukat, na nag-aalok ng maagang pisyolohikal na pagsusuri, ngunit limitado pa rin ang kasalukuyang sistema sa paraan ng operasyon, saklaw ng frequency, at aplikasyon. Ipinapakita sa disertasyon na ito ang isang multimodal, multiscale, at integratibong bioelectrical na sistema na idinisenyo upang tugunan ang mga limitasyon. Isinagawa ang implementasyon sa halamang presa (Fragaria × ananassa) sa apat na eksperimento: (1) klasipikasyon ng kalagayang tubig ng dahon, (2) pag-uuri ng prutas batay sa pamantayang pang-merkado, (3) klasipikasyon ng produktibidad ng tangkay, at (4) temporal na pag-iimahen ng ugat. Sa bawat parte, teknikal na tinukoy ang pinakamainam na frequency at ginamit sa pagkuha ng mga feature na ipinapasok sa mga modelong pang-klasipikasyon. Pinakamahusay ang TabPFN kumpara sa CatBoost, LightGBM, at XGBoost, na nagtamo ng 98.3%±1.0% na katumpakan para sa prutas, 95.8%±1.6% para sa dahon, at 88.2%±1.9% para sa tangkay. Sa pag-iimahen ng ugat, nakabuo ang mga reconstruction algorithm ng spatial resolution na 2.6–2.8 mm. Ipinakita din ng beripikasyon ang mataas na ugnayan sa pagitan ng pinakamainam na feature at pisyolohikal na metrika: r = 0.87–0.92 (prutas), r = 0.78–0.83 (dahon), at r = 0.66–0.71 (tangkay). Sa kabuuan, nagbibigay ang sistemang ito ng pundasyon para sa proaktibo at non-invasive na pagsusuri at pag-uuri ng halaman, na sumusuporta sa mga layunin ng napapanatiling agrikultura.

Abstract Format

html

Language

English

Recommended Citation

Alejandrino, J. D. (2025). Design and Implementation of a Multimodal, Multiscale Non-Invasive Bioelectrical Measurement System for Integrative and Proactive Plant Assessment. Retrieved from https://animorepository.dlsu.edu.ph/etdd_ece/13

Upload Full Text

wf_yes

Embargo Period

12-9-2025