Internet of things-based soil sensing platform for Ganoderma boninense infection detection in oil palm seedlings using machine learning
Basal stem rot (BSR), caused by a white-rot fungus Ganoderma boninense is a destructive disease that causes tremendous losses in the oil palm industry. The primary route of the disease infection is through root that has contact with Ganoderma boninense inoculum in the soil. The use of planting mater...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/104208/1/MOHD%20HAMIM%20BIN%20ABDUL%20AZIZ%20-%20IR.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Basal stem rot (BSR), caused by a white-rot fungus Ganoderma boninense is a destructive disease that causes tremendous losses in the oil palm industry. The primary route of the disease infection is through root that has contact with Ganoderma boninense inoculum in the soil. The use of planting materials (seedlings) that are resistant to Ganoderma boninense could prevent the spread of BSR disease in the plantation. A manual census is used commonly by nurseries to monitor the progress of the disease development associated with various treatments. This common nursery practice is usually conducted every two to four weeks. An irregular monitoring leads to delays in detecting the disease occurrence. This study, therefore, is focused on the use of a sensor network to obtain soil data to diagnose the Ganoderma boninense infection using the internet of things (IoT) platform. This approach could lead to a possible early infection detection methodology since rapid monitoring can avoid missing data. The objectives of the research include studying the potential use of soil properties as the indicators for BSR disease, analyzing temporal changes of infected seedlings, and developing the Ganoderma boninense disease detection model using soil properties. A total of 40 oil palm seedlings aged five months old were used in the study. They consisted of 20 healthy and 20 infected seedlings. The infected seedlings were prepared by artificially inoculating the tree roots with the Ganoderma boninense rubber woodblock. The seedlings were placed in the greenhouse with controlled environmental temperature and humidity. Three soil sensors were buried at 8 cm depth in each seedling's growth medium to measure the amount of soil moisture content (MC) in volumetric water content (in %), soil electrical conductivity (EC) (in μS/cm), and soil temperature (T) (in °C). The soil parameters data was collected every hour daily for 24 weeks (six months). These data were stored in the cloud (ThingSpeak) and available for real-time monitoring and data extraction for further analysis. The results of soil analysis revealed that more than 80% of monitored weeks in all parameters yielded |
|---|