Semi-automated plant growth monitoring system for cherry tomatoes (Solanum Lycopersicum var Cerasiforme)

• Main Authors: Dimaculangan, William Mitchell C., Hacinas, Eros Allan S., Que, Simon Justin C., Tendido, Ma. Isabel M.
• Format: text
• Language: English
• Published: Animo Repository 2023
• Subjects: Growth (Plants); Vegetation monitoring; Tomatoes > Monitoring; Computer Sciences; Databases and Information Systems
• Online Access: https://animorepository.dlsu.edu.ph/etdb_comtech/16; https://animorepository.dlsu.edu.ph/context/etdb_comtech/article/1014/viewcontent/Semi_automated_plant_growth_monitoring_system_for_cherry_tomatoes.pdf
• Institution: De La Salle University

Plant growth monitoring system (PGMS) is a platform that assists agriculturalists in monitoring the current state of the crops. However, commonly existing PGMS solely employ either sensor-based or image-based approaches. Given the equal importance of environmental and plant phenotype data to optimize the growing conditions of the crops, there is a need for a system that integrates both sensor and image-based approaches to enable agriculturalists in performing experimentations, eliciting knowledge, and making data-driven decisions. Thus, the developed system is a semi-automated PGMS with the objective of periodically collecting sensor and image data from the monitored cherry tomato crops, extrapolating soil moisture and nutrient solution pH, and determining plant productivity using deep learning. The reliability of the system was evaluated through precision, recall, f1-score, and mean absolute percentage error (MAPE) while, the acceptance of agriculturalists to the system was evaluated through a user acceptance form. Based on the results, with f1-scores of 0.83, 0.93, and 0.95 for leaves, owers, and tomatoes respectively, the system detection pipeline to acquire plant productivity information achieves competitive performance. Furthermore, the extrapolation system was able to achieve a MAPE of 11.96% for soil moisture using SVR methods, while a MAPE of 5.60% for nutrient solution pH was achieved using LR methods. Lastly, with a mean score of 4.58, 4.33, and 4.38 out of 5.00 for the usefulness of data, user interface, and system usability, the system was highly satisfactory and approved by the agriculturalist. In conclusion, the developed PGMS provides agriculturalists a reliable and quantitative platform for crop monitoring and experimentation.