KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 21

Some of the metrics are blocked by your
consent settings
Filter-based short-wave infrared imaging combined with machine learning for non-destructive quality assessment of durian pulp
(2026-09-01)
Promnioy, Surasak
;
Phetpan, Kittisak
;
Riza, Dimas Firmanda Al
;
Sharma, Sneha
;
Sirisomboon, Panmanas
The development of affordable, real-time quality monitoring tools is essential for industrial applications involving high-value tropical fruits such as durian. This study presents a cost-effective short-wave infrared multispectral imaging (SWIR-MSI) system employing three discrete bandpass filters (880, 905, and 940 nm) integrated with machine learning algorithms for non-destructive evaluation of fresh-cut durian pulp. Compared with conventional point-based NIR spectroscopy and complex hyperspectral imaging systems, the proposed configuration markedly reduces system complexity and cost while maintaining sensitivity to key compositional variations. It accurately predicted dry matter content (DMC) and starch, demonstrating that limited spectral information within the 860–1100 nm range can effectively capture moisture- and carbohydrate-related features. These findings confirm the feasibility of implementing filter-based SWIR imaging as a practical and scalable alternative to hyperspectral systems for on-line fruit quality assessment. Practically, this approach enables rapid, non-destructive, and spatially adaptable evaluation of durian pulp quality, offering significant potential for on-site grading, ripeness classification, and process control in fresh-cut durian production and packaging operations, particularly for small- and medium-scale agro-processors.
Some of the metrics are blocked by your
consent settings
Systematic evaluation of spectral preprocessing and machine learning for near-infrared prediction of mechanical stability in complex colloidal systems
(2026-06-30)
Suttho, Pisit
;
Phetpan, Kittisak
;
Al Riza, Dimas Firmanda
;
Lim, Chin Hock
;
Kuson, Pramote
Natural rubber latex (NRL) is a critical industrial material, with concentrated rubber latex (CRL) serving as a major export product. Among its quality parameters, mechanical stability time (MST) is particularly important, reflecting colloidal stability and influencing downstream applications such as glove and balloon manufacturing. Conventional MST testing, however, relies on reagents, manual agitation, and visual assessment, making it labor-intensive, operator-dependent, and unsuitable for real-time quality monitoring. Since variations in proteins, lipids, and carbohydrates strongly govern MST, near-infrared (NIR) spectroscopy offers a promising non-destructive alternative by probing their molecular vibrations. This study developed a near-process NIR instrumentation system integrated with machine learning (ML) to predict MST in CRL. Spectral signals were preprocessed using eight techniques and modeled with five supervised regression algorithms. The best-performing configuration, Savitzky-Golay second derivative and orthogonal signal correction coupled with partial least squares regression, yielded high predictive accuracy, with coefficient of determination for prediction (R2<inf>p</inf>) of 0.94 and ratio of performance to deviation (RPD) of 4.2. This performance demonstrates the system's ability to extract chemically relevant information governing latex stability. The proposed NIR-ML framework provides a rapid, reagent-free, and scalable alternative to conventional MST testing, addressing the limitations of existing methods and supporting industrial quality monitoring. This approach is also transferable to the analysis of complex colloidal systems across diverse applications. Furthermore, the study provides mechanistic insight into how spectral preprocessing enhances the extraction of chemically meaningful information, establishing a physically interpretable framework for NIR-based analysis of such complex systems.
Some of the metrics are blocked by your
consent settings
Low-Cost Near-Infrared Spectroscopy for Rapid Prediction of Biodiesel Properties: Acid Value, Density, Viscosity, and Water Content
(2026-03-31)
Phetpan, Kittisak
;
Thongphut, Chitwadee
;
Chungcharoen, Thatchapol
;
Limmun, Warunee
Partial least squares (PLS) regression, combined with various spectral pre-processing techniques, was employed to compare the performance of two diode array near-infrared (NIR) spectrometers in predicting key biodiesel quality parameters, including acidity, viscosity, density, and water content. An AvaSpec-Mini4096CL NIR spectrometer, operating within the 350–1100 nm wavelength range, was used as the representative shortwave near-infrared (SW-NIR) spectrometer, while a NIRQuest512 spectrometer, covering the 900–1700 nm range, was employed as the longwave near-infrared (LW-NIR) spectrometer. Both spectrometers were equipped with a transflection probe for spectral collection from oil palm-based biodiesel samples. The SW-NIR spectrometer outperformed the LW-NIR spectrometer. The optimal PLS models achieved root mean square errors of prediction (RMSEP) of 0.0037 mg KOH/g for acidity, 0.062 cSt (mm2/s) for viscosity, 2.67 kg/m3 for density, and 59.14 mg/kg for water content, highlighting the potential of compact SW-NIR spectrometers as effective, low-cost tools for rapid biodiesel quality monitoring.
Some of the metrics are blocked by your
consent settings
Investigation of physiological disorder classification in mangosteen fruit using visible and shortwave near-infrared spectroscopy combined with machine learning
(2025-12-01)
Ruttanadech, Nuttapong
;
Momin, Abdul
;
Phetpan, Kittisak
;
Chaichanyut, Montree
;
Thongphut, Chitwadee
Accurate classification of physiological disorders in mangosteen fruit is crucial for ensuring production quality, safety, sustainability, and economic viability. This study investigates the application of visible and shortwave near-infrared (Vis/SWNIR) reflectance spectroscopy, combined with machine learning algorithms, to classify three primary disorders: normal fruit (NF), translucent flesh disorder (TFD), and TFD with yellow gummy latex (TFD & YGL). The study specifically examines the effects of light intensity, spectral pretreatments, and machine learning models on classification performance. Spectral data were collected using two light intensities (50 % and 100 % of a 150 W light source) and processed with three pretreatments: standard normal variate (SNV), second derivative Savitzky-Golay (SGD2), and a combination of SNV and SGD2. Random forest (RF), support vector machine (SVM), and multi-layer perceptron (MLP) algorithms were used for classification. The SGD2 method improved differentiation, especially for the TFD & YGL class, in the 700–725 nm wavelength range, which is associated with xanthone content in the fruit's pericarp. Higher light intensity (100 %) significantly improved classification accuracy, achieving an overall accuracy of 0.71 and an average F1 score of 0.61 with the RF model. Despite these improvements, the model struggled to distinguish the TFD class from NF due to their similar spectral profiles. Overall, the Vis/SWNIR spectroscopy and machine learning combination shows strong potential for the non-destructive classification of mangosteen fruit disorders. Both light intensity and spectral pretreatments play critical roles in enhancing performance. Future studies should focus on improving spectral sensitivity to better capture internal fruit characteristics.
Some of the metrics are blocked by your
consent settings
Enhanced biodiesel purification using coffee husk bioadsorbents: The role of pyrolysis temperature, KOH activation, and adsorption efficiency
(2025-05-01)
Chungcharoen, Thatchapol
;
Limmun, Warunee
;
Srisang, Siriwan
;
Phetpan, Kittisak
;
Ruttanadech, Nuttapong
This study evaluates the performance of bioadsorbents derived from coffee husk pyrolyzed at temperatures of 600, 700, and 800 °C (CH600, CH700, and CH800), along with activated CH700 (ACH700), in biodiesel purification. The results indicate that CH700 significantly enhances biodiesel purity, with optimal purification conditions achieved at a dosage of 2 wt% CH700, a stirring rate of 400 rpm, and a contact time of 45 min. CH700 demonstrated modest performance, achieving approximately 20 % removal of methanol and water. However, after activation with potassium hydroxide (KOH), ACH700 demonstrated improved efficiency, achieving 96.92 % methanol removal and 39.46 % water removal. ACH700 also refined biodiesel quality to meet EN14214 standards and maintained a higher biodiesel yield compared to other adsorbents. The bioadsorption process is influenced by the chemical interactions between the surface functional groups of the bioadsorbent and the contaminants, which is further enhanced by the optimized pore structure of ACH700. The use of ACH700 represents a novel and highly effective approach to biodiesel purification, combining both technical efficiency and economic feasibility. Furthermore, the valorization of agricultural waste adds significant environmental benefits, reinforcing the potential of ACH700 for large-scale biodiesel production.
Some of the metrics are blocked by your
consent settings
Hybrid machine learning models: A comprehensive, data-driven evaluation with diverse data partitioning strategies for net radiation estimation
(2025-04-30)
Bajao, Kristian Lorenz
;
Phetpan, Kittisak
;
Chophuk, Ponlawat
;
Suwalak, Rattapong
Surface net radiation (Rn) is crucial for climate modeling and agricultural management but is often not readily available, especially in regions like Thailand. Accurate prediction of Rn is essential for estimating evapotranspiration, which is vital for irrigation planning and agricultural productivity. This study develops a hybrid machine learning framework that incorporates K-Nearest Neighbors (KNN) for missing data imputation, Random Forest-Recursive Feature Elimination (RF-RFE) for feature selection, and machine learning models (Multi-layer Perceptron, K-Nearest Neighbors, and Random Forest) for prediction. The research evaluates various data partitioning methods, including hold-out split, K-fold cross-validation, and growing-window forward-validation (gwFV), alongside hyperparameter tuning using GridSearch to enhance model robustness and prevent overfitting. The primary objectives are to develop and evaluate the hybrid ML models for daily Rn estimation using basic meteorological inputs (temperature, relative humidity, and sunshine duration), assess the impact of different input combinations on prediction accuracy in Sawi, Chumphon, Thailand, and compare data partitioning techniques to determine the optimal model performance. Utilizing FAO56PM-calculated Rn as a reference, this study finds that the Random Forest model, with average temperature and sunshine duration (M2) as inputs evaluated under the gwFV method, achieves the highest stability and high accuracy (R² of 0.972, RMSE of 0.457 MJ m-2 day-1, and MAPE of 3.50%). The Random Forest demonstrates strong generalization capabilities, making it a reliable choice. Even models using only sunshine duration (M3) perform adequately, offering a solution when data availability is scarce. This study concludes that hybrid machine learning models, combined with careful data partitioning, significantly improve Rn estimation. These advancements provide valuable insights for climate modeling, agricultural management, and irrigation scheduling, particularly in data-scarce regions.
Some of the metrics are blocked by your
consent settings
The improvement of germination method for producing the germinated brown rice using a water spraying system with a revolved sieve
(2024-11-01)
Chungcharoen, Thatchapol
;
Sansiribhan, Sansanee
;
Munsin, Ronnachart
;
Phetpan, Kittisak
;
Fonghiransiri, Surasak
Water soaking is an important method in germinated brown rice (GBR) production that causes fermentation, leading to an unpleasant smell of GBR. In this research, a water spraying system with a revolved sieve is applied to produce the GBR. The increased speed and time of spray break led to higher moisture content and water absorption. The spray break of 30 min and revolved speed of 15 rpm provided the shortest time to obtain the paddy with a moisture content of 30% (w.b.). The incubation pattern with a revolved sieve and water spray provided the shortest incubation time for 90% germination. When producing the GBR with a water spraying system with a revolved sieve (GBR-WSSRS), it had a lower number of microorganisms compared to the GBR with a water soaking (GBR-WS), leading to higher scores of overall acceptability. However, the GBR-WSSRS had a lower GABA content than the GBR-WS.
Some of the metrics are blocked by your
consent settings
Enhancing biodiesel yield and purification with a recently developed centrifuge machine: A response surface methodology approach
(2024-04-15)
Limmun, Warunee
;
Chungcharoen, Thatchapol
;
Rattanamechaiskul, Chaiwat
;
Phetpan, Kittisak
;
Limmun, Wanida
Biodiesel production processes, such as gravity settling, have limitations in terms of biodiesel yield, purification efficiency, operating time in the separation process, and more extensive equipment. Therefore, this study has focused on using a recently developed centrifuge machine for biodiesel separation to address these challenges due to its compact design, high efficiency, and simplicity. Additionally, this study aimed to optimize the separation efficiency of glycerol from biodiesel using a centrifuge machine, employing response surface methodology (RSM) with central composite design (CCD). The optimum conditions for separating glycerol from biodiesel via centrifuge machine are a rotation speed of 1800 rpm, a mixture flow rate of 192.25 ml/min, and a temperature of 55 °C, respectively. In optimum conditions, 94.52% separation efficiency was achieved. Biodiesel production can be improved, leading to higher yields and greater purity. The utilization of RSM proved valuable in determining the optimum conditions for separation. Furthermore, the machine successfully separated the biodiesel to meet ASTM D6751 and EN 14,214 standards. The results highlight the potential of the centrifuge machine for efficient and reliable biodiesel production, contributing to the advancement of the biodiesel industry.
Some of the metrics are blocked by your
consent settings
Classification of the Crosslink Density Level of Para Rubber Thick Film of Medical Glove by Using Near-Infrared Spectral Data
(2024-01-01)
Jongyingcharoen, Jiraporn Sripinyowanich
;
Howimanporn, Suppakit
;
Sitorus, Agustami
;
Phanomsophon, Thitima
;
Posom, Jetsada
Classification of the crosslink density level of para rubber medical gloves by using near-infrared spectral data combined with machine learning is the first time reported in this paper. The spectra of medical glove samples with different crosslink densities acquired by an ultra-compact portable MicroNIR spectrometer were correlated with their crosslink density levels, which were referencely evaluated by the toluene swell index (TSI). The machine learning protocols used to classify the 3 groups of TSI were specified as less than 80% TSI, 80–88% TSI, and more than 88% TSI. The 80–88% TSI group was the group in which the compounded latex was suitable for medical glove production, which made the glove specification comply with the requirements of customers as indicated by the tensile test. The results show that when comparing the algorithms used for modeling, the linear discriminant analysis (LDA) developed by 2nd derivative spectra with 15 k-best selected wavelengths fairly accurately predicted the class but was most reliable among other algorithms, i.e., artificial neural networks (ANN), support vector machines (SVM), and k-nearest neighbors (kNN), due to higher prediction accuracy, precision, recall, and F1-score of the same value of 0.76 and no overfitting or underfitting prediction. This developed model can be implemented in the glove factory for screening purposes in the production line. However, deep learning modeling should be explored with a larger sample number required for better model performance.
Some of the metrics are blocked by your
consent settings
Near-infrared hyperspectral imaging combined with machine learning for physicochemical-based quality evaluation of durian pulp
(2023-06-01)
Sharma, Sneha
;
Sirisomboon, Panmanas
;
K.C, Sumesh
;
Terdwongworakul, Anupun
;
Phetpan, Kittisak
This research reports on the application of near-infrared hyperspectral imaging (NIR-HSI) system for predicting the physicochemical properties; dry matter (DM), total soluble solids (TSS), and fat content (FC) of durian. Partial least squares regression (PLSR), support vector machine (SVM), random forest (RF), and 1D convolution neural network (CNN) models: custom, U-Net, and VGG19; were developed to predict DM, TSS, and FC of durian pulp. Feature wavelengths were selected using a genetic algorithm (GA) and successive projection algorithm (SPA). The selected wavelengths were then validated based on the algorithms for regression model development. GA-PLSR model was compelling to predict the DM and FC in durian pulp, which obtained the coefficient of determination for the test set (r2) and root mean square error of prediction (RMSEP) of 0.97 and 1.12% for DM and 0.86 and 0.64% for FC, respectively. The GA-PLSR model provided the best result for the TSS prediction with r2, and RMSEP of 0.90 and 1.40%, respectively, whereas the SPA-PLSR model based on only thirteen wavelengths attained fair result with the r2 and RMSEP of 0.79 and 2.03%, respectively. The above results show that the pushbroom NIR-HSI system achieved promising results for estimating DM, TSS, and FC in durian pulp. This research identified the featured wavelengths that can be used to develop a portable and reliable HSI or multispectral system to be installed at durian packaging firms for quality inspection and grading.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results