Perbandingan Fitur Handcrafted dan Non-Handcrafted untuk Klasifikasi Kelembapan Tanah
DOI:
https://doi.org/10.54367/means.v11i1.6234Keywords:
soil moisture classification, handcrafted features, Random Forest, MobileNetV2, image processing, transfer learningAbstract
Soil moisture is a crucial parameter in precision agriculture because it impacts irrigation management and crop productivity. Conventional measurement methods using physical sensors have limitations in terms of cost, maintenance, and coverage area. This study aims to compare the performance of handcrafted and non-handcrafted feature approaches in digital image-based soil moisture classification. The handcrafted approach extracts 60 color features consisting of RGB and HSV statistics and HSV histograms, then classifies them using the Random Forest algorithm. The non-handcrafted approach uses the transfer learning- based MobileNetV2 architecture to automatically learn feature representations from raw images. Both methods were tested using the same dataset and experimental protocol to ensure an objective comparison. The experimental results showed that Random Forest achieved an accuracy of 91.05%, higher than MobileNetV2's 87.65%. Confusion matrix analysis indicated that the dominant misclassification occurred between the moderate and wet classes due to the similarity in color distribution. The results show that for limited datasets and problems dominated by color characteristics, handcrafted features can provide more stable and efficient performance than deep learning models. This study emphasizes the importance of selecting a classification method based on data characteristics and computational requirements, rather than solely on model complexity.References
X. Zhang, G. Feng, and X. Sun, “Advanced technologies of soil moisture monitoring in precision agriculture: A Review,” Dec. 01, 2024, Elsevier B.V.
S. M. Shawon, N. I. Neha, A. N. Jui, N. Dey, and H. T. Zubair, “Advances in soil moisture measurement techniques and prediction using artificial intelligence: An extensive and systematic review,” Dec. 01, 2025, Elsevier B.V.
S. Gupta et al., “Smart agriculture using IoT for automated irrigation, water and energy efficiency,” Smart Agricultural Technology, vol. 12, Dec. 2025
M. Naeimi, P. Daggupati, and A. Biswas, “Image-based soil characterization: A review on smartphone applications,” Dec. 01, 2024, Elsevier B.V.
M. Naeimi et al., “Environmental factors in image-based soil analysis: Interaction effects of moisture, texture, and illumination,” Geoderma, vol. 463, Nov. 2025
D. Roy, T. Ghosh, B. Das, R. Jatav, and D. Chakraborty, “Smartphone-based image analysis and interpretable machine learning for soil moisture estimation across diverse Indian soils,” Remote Sens. Appl., vol. 39, Aug. 2025
E. O. Babalola, M. H. Asad, and A. Bais, “Soil Surface Texture Classification Using RGB Images Acquired Under Uncontrolled Field Conditions,” IEEE Access, vol. 11, pp. 67140–67155, 2023
C. Suneetha, L. S. Kumar, K. Sreenivas, and T. Mitran, “Deep learning-driven soil texture classifier using Landsat 8 images,” Remote Sens. Appl., vol. 38, Apr. 2025,
H. Elmannai and A. D. Al-Garni, “Classification using semantic feature and machine learning: Land-use case application,” Telkomnika (Telecommunication Computing Electronics and Control), vol. 19, no. 4, pp. 1242–1250, 2021
M. Latha, M. Vasavi, C. K. Kumar, R. Balamanigandan, J. B. Guttikonda, and R. T. Kumar, “Machine Learning Based Precision Agriculture using Ensemble Classification with TPE Model,” Journal of Machine and Computing, vol. 4, no. 1, pp. 261–268, Jan. 2024
[1] X. Zhang, G. Feng, and X. Sun, “Advanced technologies of soil moisture monitoring in precision agriculture: A Review,” Dec. 01, 2024, Elsevier B.V.
S. M. Shawon, N. I. Neha, A. N. Jui, N. Dey, and H. T. Zubair, “Advances in soil moisture measurement techniques and prediction using artificial intelligence: An extensive and systematic review,” Dec. 01, 2025, Elsevier B.V.
S. Gupta et al., “Smart agriculture using IoT for automated irrigation, water and energy efficiency,” Smart Agricultural Technology, vol. 12, Dec. 2025
M. Naeimi, P. Daggupati, and A. Biswas, “Image-based soil characterization: A review on smartphone applications,” Dec. 01, 2024, Elsevier B.V.
M. Naeimi et al., “Environmental factors in image-based soil analysis: Interaction effects of moisture, texture, and illumination,” Geoderma, vol. 463, Nov. 2025
D. Roy, T. Ghosh, B. Das, R. Jatav, and D. Chakraborty, “Smartphone-based image analysis and interpretable machine learning for soil moisture estimation across diverse Indian soils,” Remote Sens. Appl., vol. 39, Aug. 2025
E. O. Babalola, M. H. Asad, and A. Bais, “Soil Surface Texture Classification Using RGB Images Acquired Under Uncontrolled Field Conditions,” IEEE Access, vol. 11, pp. 67140–67155, 2023
C. Suneetha, L. S. Kumar, K. Sreenivas, and T. Mitran, “Deep learning-driven soil texture classifier using Landsat 8 images,” Remote Sens. Appl., vol. 38, Apr. 2025,
H. Elmannai and A. D. Al-Garni, “Classification using semantic feature and machine learning: Land-use case application,” Telkomnika (Telecommunication Computing Electronics and Control), vol. 19, no. 4, pp. 1242–1250, 2021
M. Latha, M. Vasavi, C. K. Kumar, R. Balamanigandan, J. B. Guttikonda, and R. T. Kumar, “Machine Learning Based Precision Agriculture using Ensemble Classification with TPE Model,” Journal of Machine and Computing, vol. 4, no. 1, pp. 261–268, Jan. 2024
A. Verma and P. Patel, “Comparative analysis of handcrafted and deep features for agricultural image classification,” IEEE Access, vol. 10, pp. 98765–98778, 2022.
F. Ahmed, T. Hasan, and M. Uddin, “Performance evaluation of ensemble learning for small agricultural datasets,” Expert Systems with Applications, vol. 195, p. 116594, 2022.
L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001.
Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, pp. 436–444, 2015.
M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted residuals and linear bottlenecks,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2018, pp. 4510–4520.
A. Howard et al., “Searching for MobileNetV3,” in Proc. IEEE/CVF Int. Conf. Computer Vision (ICCV), 2019, pp. 1314–1324.
T. Nguyen and D. Lee, “Soil image classification using convolutional neural networks and transfer learning,” Remote Sensing, vol. 13, no. 7, p. 1234, 2021.
P. Rahman, M. Islam, and S. Rahman, “Deep learning-based agricultural image classification: A comparative study,” Sensors, vol. 23, no. 5, p. 2456, 2023.
S. Yadav and K. Mehta, “Comparative study of deep learning and traditional machine learning for environmental image classification,” Ecological Informatics, vol. 75, p. 102012, 2024.
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Diah Septiani, Faisal Dharma Adhinata
This work is licensed under a Creative Commons Attribution 4.0 International License.
