000			04021ntm a2200373 i 4500
999			_c99755 _d99761
003			MY-KuUP
005			20251125110752.0
006			t||||fr|||| 000 0
007			ta
008			230718t20232023my a|||fr|||| 000 0 eng d
020			_aTHE0009479 (Local) _qHardback
040			_aUMP _beng _cUMP _erda
090			_aFTKEE .M37 2023 r Thesis
100	1		_aMas Ira Syafila Mohd Hilmi Tan, _eauthor.
245	1	0	_aTerahertz sensing analysis for early detection of ganoderma boninense disease using near infrared (NIR) spectrometer / _cMas Ira Syafila Mohd Hilmi Tan
264		1	_aPahang : _bUMP, _c2023
264		4	_c© 2023
300			_axix, 279 pages : _bIllustration ; _c30 cm.+ _e1 CD ROM
336			_2rdacontent _atext
336			_2rdacontent _atext
337			_2rdamedia _aunmediated
337			_2rdamedia _acomputer
338			_2rdacarrier _avolume
338			_2rdacarrier _acomputer disc
347			_2rda _atext file _bPDF
500			_aFaculty of Electrical & Electronics Engineering Technology
502			_aThesis (Doctor of Philosophy) -- Universiti Malaysia Pahang – 2023
504			_aIncludes bibliographical reference
520	3		_aGanoderma boninense (G. boninense) infection reduces the productivity of oil palms and causes a significant impact on the economic viability of oil palm plantations. Early detection is critical for the effective management of this disease since there is no effective treatment that can stop the spread of this disease. The proposed system uses integrated hand-held near-infrared spectroscopy (NIRS) for early detection of G. boninense on asymptomatic oil palm seedlings and classification of spectral data using machine learning (ML) techniques. The non-destructive method using NIRS with ML and predictive analytics has the potential to be a highly sensitive and reliable method for the early detection of G. boninense. Spectral data are collected from 6 samples of inoculated and non-inoculated oil palm samples at nursery stages using an integrated NIRS sensor. Chemometrics is performed by implementing principal component analysis (PCA), derivatives and partial least square (PLS) regression to extract the vital information of the spectra. The significant wavelengths are at 1310 nm and 1450 nm which are attributable to ergosterol and water content, respectively. Furthermore, the SG derivatives spectra peaks corresponded to specific functional groups that could be utilized for the detection of G. boninense. These functional groups encompass the third overtone of N-H stretching, the second overtone of C-H stretching, and a combination band involving both C-H stretching and O-H stretching. High-performance liquid chromatography (HPLC) analysis is performed to identify the ergosterol content in oil palm sample. Ergosterol can be used as a biomarker for the detection of G. boninense since it can only be found in the fungal-infested plant. In classification, four different ML algorithms: K-Nearest Neighbour (kNN), Naïve Bayes (NB), Support Vector Machine (SVM) and Decision Tree (DT) are tested to classify healthy and infected oil palm samples. DT algorithm on leaves spectra achieves a satisfactory overall performance compared to the other classifiers with high accuracy up to 93.1% and an F1-score of 92.6%. Therefore, a DT-based predictive analytic on leaves NIR spectral reference data is developed for real-time detection of G. boninense infection. A portable smart G. boninense detection system prototype is developed by implementing the Internet of Things (IoT) into the system which enables the integration of sensors and server to perform prediction of healthy or infected oil palm seedlings. This working prototype showed that this proposed approach is reliable and practical for the early detection of G. boninense in oil palm seedlings.
610	2	0	_aFaculty of Electrical & Electronics Engineering Technology _xDissertations
650		0	_aUniversities and colleges _xDissertations
650		0	_aThesis
942			_2lcc _cTHESIS