Khwantri Saengprachatanarug
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Khwantri Saengprachatanarug has a total of 61 co-authors. Here are the top co-authors by number of shared publications:
Jetsada Posom
13 shared publications- Jetsada Posom
Department of Agricultural Engineering, Faculty of Engineering, Khon Kaen University, Khon Kaen 40002, Thailand
9 shared publications Seree Wongpichet
9 shared publicationsArthit Phuphaphud
7 shared publications- Arthit Phuphaphud5 shared publications
Eizo Taira
5 shared publications
Latest Publications
- Open Access
- 12 May 2026
- Published by Elsevier in SSRN Electronic Journal
Attribute-dependent modeling for NIR-based quality grading of durian pulp
Effective grading of durian pulp is essential for the export market. While dry matter (DM) and total soluble solids (TSS) serve as key quality indices.This study aims to classify durian pulp samples according to their DM and TSS levels utilising NIR spectroscopy combined with machine learning (ML). The samples were scanned using reflectance spectra between 500 and 1100 nm, and then were categorised into different classes using DM, TSS, and combined DM–TSS criteria. The models were developed by integrating spectral preprocessing, dimensionality reduction, and class imbalance handling, followed by comparison of multiple ML algorithms. Using DM-based grading, index regression to classification method outperformed direct classification; Ridge regression with autoscaling pretreatment achieved 70% classification overall accuracy after threshold-based grouping. In contrast, TSS classification was more effectively addressed using a Gradient Boosting classifier with standard scaling, reaching 93.8% overall accuracy. For combined DM–TSS grading, an SNV–PLS–MLP model with random oversampling mitigated class imbalance and achieved 81.3% accuracy. These results demonstrate that optimal NIR–ML strategies are attribute-dependent: regression frameworks are more suitable for DM prediction, whereas direct classification performs better for sugar-related attributes. The proposed approach supports rapid, non-destructive quality grading of durian pulp for export applications.
- 12 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAttribute-dependent modeling for NIR-based quality grading of durian pulp
Effective grading of durian pulp is essential for the export market. While dry matter (DM) and total soluble solids (TSS) serve as key quality indices.This study aims to classify durian pulp samples according to their DM and TSS levels utilising NIR spectroscopy combined with machine learning (ML). The samples were scanned using reflectance spectra between 500 and 1100 nm, and then were categorised into different classes using DM, TSS, and combined DM–TSS criteria. The models were developed by integrating spectral preprocessing, dimensionality reduction, and class imbalance handling, followed by comparison of multiple ML algorithms. Using DM-based grading, index regression to classification method outperformed direct classification; Ridge regression with autoscaling pretreatment achieved 70% classification overall accuracy after threshold-based grouping. In contrast, TSS classification was more effectively addressed using a Gradient Boosting classifier with standard scaling, reaching 93.8% overall accuracy. For combined DM–TSS grading, an SNV–PLS–MLP model with random oversampling mitigated class imbalance and achieved 81.3% accuracy. These results demonstrate that optimal NIR–ML strategies are attribute-dependent: regression frameworks are more suitable for DM prediction, whereas direct classification performs better for sugar-related attributes. The proposed approach supports rapid, non-destructive quality grading of durian pulp for export applications.
- 23 March 2026
- Published by Springer Nature in Sugar Tech
- Vol. 28 (2) , 686-699
Site-Specific UAV Spraying Versus Conventional Weed Control in Sugarcane: Effects on Herbicide Use, Efficiency, and Environmental Impact
- 23 March 2026
- Published by Springer Nature in Sugar Tech
- Vol. 28 (2) , 686-699
Site-Specific UAV Spraying Versus Conventional Weed Control in Sugarcane: Effects on Herbicide Use, Efficiency, and Environmental Impact
- Open Access
- 3 March 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 13
A parameter-by-structure segmentation framework for enhanced CMD detection using UAV imagery and machine learning
- 3 March 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 13
Open AccessA parameter-by-structure segmentation framework for enhanced CMD detection using UAV imagery and machine learning
- Open Access
- 25 May 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 11
Comparison of sample preparation methods for higher heating values in various sugarcane varieties using near-infrared spectroscopy
- 25 May 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 11
Open AccessComparison of sample preparation methods for higher heating values in various sugarcane varieties using near-infrared spectroscopy
- Open Access
- 14 January 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 10
Forecasting gaps in sugarcane fields containing weeds using low-resolution UAV imagery based on a machine-learning approach
- 14 January 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 10
Open AccessForecasting gaps in sugarcane fields containing weeds using low-resolution UAV imagery based on a machine-learning approach
- Open Access
- 1 December 2024
- Published by Elsevier in Smart Agricultural Technology
- Vol. 9
Design, development, and testing of a cassava storage root-cutting robot utilizing a Stewart platform and mask R-CNN for precision agriculture
- 1 December 2024
- Published by Elsevier in Smart Agricultural Technology
- Vol. 9
Open AccessDesign, development, and testing of a cassava storage root-cutting robot utilizing a Stewart platform and mask R-CNN for precision agriculture
- 11 June 2024
- Published by Elsevier in Crop Protection
- Vol. 184, 106805
Time-efficient low-resolution RGB aerial imaging for precision mapping of weed types in site-specific herbicide application
- 11 June 2024
- Published by Elsevier in Crop Protection
- Vol. 184, 106805
Time-efficient low-resolution RGB aerial imaging for precision mapping of weed types in site-specific herbicide application
- Open Access
- 1 August 2023
- Published by PAGEPress Publications in Journal of Agricultural Engineering
- Vol. 54 (2)
Cassava stalk detection for a cassava harvesting robot based on YOLO v4 and Mask R-CNN
The quality of fresh cassava roots can be increased through the use of precision equipment. As a first step towards developing an automatic cassava root cutting system, this study demonstrates the use of a computer vision system with deep learning for cassava stalk detection. An RGB image of a cassava tree mounted on a cassava-pulling machine was captured, and the YOLO v4 model and two Mask R-CNN models with ResNet 101 and ResNet 50 base architectures were employed to train the weights to predict the position of the cassava stalk. One hundred test images of stalks of various shapes and sizes were used to determine the grasping point and inclination, and the results from manual annotation were compared with the predicted results. Regarding localisation, Mask R-CNN with ResNet 101 gave a significantly higher performance than the other models, with an F1 score and a mean IoU of 0.81 and 0.70, respectively. YOLO v4 showed the highest correlation for the x- and y-coordinates for the prediction of the grasping point, with values for R2 of 0.89 and 0.53, respectively. For inclination prediction, Mask R-CNN with ResNet 101 and Mask R-CNN with ResNet 50 gave the same level of correlation, with values for R2 of 0.50 and 0.61, respectively. These results were acceptable for use as design criteria for developing a cassava rootcutting robot.
- 1 August 2023
- Published by PAGEPress Publications in Journal of Agricultural Engineering
- Vol. 54 (2)
Open AccessCassava stalk detection for a cassava harvesting robot based on YOLO v4 and Mask R-CNN
The quality of fresh cassava roots can be increased through the use of precision equipment. As a first step towards developing an automatic cassava root cutting system, this study demonstrates the use of a computer vision system with deep learning for cassava stalk detection. An RGB image of a cassava tree mounted on a cassava-pulling machine was captured, and the YOLO v4 model and two Mask R-CNN models with ResNet 101 and ResNet 50 base architectures were employed to train the weights to predict the position of the cassava stalk. One hundred test images of stalks of various shapes and sizes were used to determine the grasping point and inclination, and the results from manual annotation were compared with the predicted results. Regarding localisation, Mask R-CNN with ResNet 101 gave a significantly higher performance than the other models, with an F1 score and a mean IoU of 0.81 and 0.70, respectively. YOLO v4 showed the highest correlation for the x- and y-coordinates for the prediction of the grasping point, with values for R2 of 0.89 and 0.53, respectively. For inclination prediction, Mask R-CNN with ResNet 101 and Mask R-CNN with ResNet 50 gave the same level of correlation, with values for R2 of 0.50 and 0.61, respectively. These results were acceptable for use as design criteria for developing a cassava rootcutting robot.
- 8 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 62-67
Measurement of Trash Percentage for Mechanical Harvested Sugarcane Using Multispectral Image
This research aimed to develop a new technique for estimating the percentage of trash in mechanically harvested sugarcane samples using multispectral imaging and machine learning. The research found that the Random Forest model with a Classification Tree of 150 was the most effective classification model, with a Quality Percentage of 96.05%. In addition, the best prediction models for estimating the weight of sugarcane billet using Multiple Linear Regression had an R2 value of 0.77 and a Root Mean Square Error of calibration (RMSEC) value of 21.03 g. The results of the research also showed that the new technique was capable of accurately estimating the percentage of trash in sugarcane samples, with a Root Mean Square Error of prediction (RMSEP) of 2.09% when compared to traditional methods. This suggests that the new technique may be a useful, rapid method for assessing the percent of sugarcane billet and trash in production processes.
- 8 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 62-67
Measurement of Trash Percentage for Mechanical Harvested Sugarcane Using Multispectral Image
This research aimed to develop a new technique for estimating the percentage of trash in mechanically harvested sugarcane samples using multispectral imaging and machine learning. The research found that the Random Forest model with a Classification Tree of 150 was the most effective classification model, with a Quality Percentage of 96.05%. In addition, the best prediction models for estimating the weight of sugarcane billet using Multiple Linear Regression had an R2 value of 0.77 and a Root Mean Square Error of calibration (RMSEC) value of 21.03 g. The results of the research also showed that the new technique was capable of accurately estimating the percentage of trash in sugarcane samples, with a Root Mean Square Error of prediction (RMSEP) of 2.09% when compared to traditional methods. This suggests that the new technique may be a useful, rapid method for assessing the percent of sugarcane billet and trash in production processes.