THRESHOLD IMPROVEMENT USING THE OTSU METHOD WITH HISTOGRAM ON PADDY IMAGES
DOI:
https://doi.org/10.35631/JISTM.1142010Keywords:
Image Segmentation, Otsu Method, Paddy Grain ClassificationAbstract
Accurate image segmentation methods play a very important role in improving paddy grain classification performance. One of the segmentation stages is determining the threshold, usually using classic Otsu. This method is widely used in the segmentation process, and the instability of the threshold in classic Otsu in uneven lighting and complex histogram distribution. This study proposes an improvement to the Otsu-based segmentation method by combining histogram normalization and trigonometric variance modulation, namely Otsu with normalization, Otsu Sine, and Otsu Tangent. The proposed methods were evaluated using Random Forest as a classifier and texture features with GLCM. Experimental results show that the Otsu Sine method achieves the best performance among the other methods, with an accuracy of 0.92, precision of 0.93, recall of 0.93, and F1 score of 0.93. Five-fold cross-validation yields superior results for Otsu-Sine, with the highest average accuracy (94.83%) among all methods. Further pairwise tests showed that the improvement in performance compared to classic Otsu was statistically significant (p = 0.04). Threshold stability analysis showed that Otsu-Sine maintained low variance while adapting effectively to changes in intensity distribution, whereas Otsu-Tangent showed high instability.
Downloads
References
Al-Johania, N. A., & Elrefaei, L. A. (2019). Dorsal hand vein recognition by convolutional neural networks: Feature learning and transfer learning approaches. International Journal of Intelligent Engineering and Systems, 12(3). https://doi.org/10.22266/IJIES2019.0630.19
B, S., K, S., & M, C. D. (2023). a Comprehensive Review on Diagnosis and Classification of Paddy Leaf Diseases Using Advanced Computer Vision Technologies. ICTACT Journal on Image and Video Processing, 13(4), 2973–2986. https://doi.org/10.21917/ijivp.2023.0424
Bhupendra, Kriz Moses, Ankur Miglani, P. K. K. (2022). Deep CNN-based damage classification of milled rice grains using a high-magnification image dataset. Computers and Electronics in Agriculture, 195(April). https://doi.org/https://doi.org/10.1016/j.compag.2022.106811
Chang, H. C., Yu, L. W., Liu, B. Y., & Chang, P. C. (2024). Classification of the implant-ridge relationship utilizing the MobileNet architecture. Journal of Dental Sciences, 19(1), 411–418. https://doi.org/10.1016/j.jds.2023.08.002
Dahiri, & Tineke, R. (2021). Tinjauan Kritis Produksi Padi Nasional. Pusat Kajian Anggaran, 4, 763–773.
Faurina, R., Rahma, S., Vatresia, A., & Susanto, A. (2024). Comparison of Convolutional Neural Networks Transfer Learning Models for Disease Classification of Food Crops. International Journal on Informatics Visualization, 8(4), 2020–2032. https://doi.org/10.62527/joiv.8.4.1936
Jing, Z., & Tang, B. (2024). Improved image segmentation method based on Otsu thresholding and level set techniques. Journal of Physics: Conference Series, 2813(1). https://doi.org/10.1088/1742-6596/2813/1/012017
Kalyani, R., Sathya, P. D., & Sakthivel, V. P. (2021). Multilevel Thresholding for Medical Image Segmentation Using Teaching-Learning Based Optimization Algorithm. International Journal of Intelligent Engineering and Systems, 14(2), 11–21. https://doi.org/10.22266/ijies2021.0430.02
Kantor Staf Presiden. (2025). Presiden Prabowo Tegaskan Komitmen Swasembada Pangan dan Harga Gabah. https://www.ksp.go.id/presiden-prabowo-tegaskan-komitmen-swasembada-pangan-dan-harga-gabah.html
Mamoriska, S. (2024). Pengelolaan Beras sebagai Cadangan Pangan dari Perspektif Word Trade Organization (WTO) Rice Management as a Food Reserve from the Perspective of the World Trade Organization (WTO). 57–80.
Meinam, S., Nongmeikapam, K., & Singh, N. B. (2025). Techniques of image segmentation: a review. Indonesian Journal of Electrical Engineering and Computer Science, 38(2), 830. https://doi.org/10.11591/ijeecs.v38.i2.pp830-844
Nugroho, H., Pramudito, W. A., & Laksono, H. S. (2024). Gray Level Co-Occurrence Matrix (GLCM)-based Feature Extraction for Rice Leaf Diseases Classification. Buletin Ilmiah Sarjana Teknik Elektro, 6(4), 392–400. https://doi.org/10.12928/biste.v6i4.9286
Otekunrin, O. A. (2024). Assessing the Prevalence and Severity of Global Hunger and Food Insecurity: Recent Dynamics and Sub-Saharan Africa’s Burden. Sustainability (Switzerland), 16(12). https://doi.org/10.3390/su16124877
Rajalakshmi, R., Faizal, S., Sivasankaran, S., & Geetha, R. (2024). RiceSeedNet: Rice seed variety identification using deep neural network. Journal of Agriculture and Food Research, 16(September 2023), 101062. https://doi.org/10.1016/j.jafr.2024.101062
Sadhana, K. R., & Mangalwede, S. R. (2020). Analysis of Segmentation and Classification Approach Using Image Processing over Rice Samples. 7(2), 409–413.
Sang, G., Wang, X., & Zhang, J. (2024). Improved OTSU Theory of Image Multi-Threshold Segmentation by Incorporating Ant Colony Algorithm. Informatica (Slovenia), 48(9), 65–74. https://doi.org/10.31449/inf.v48i9.5872
Setiawan, A., Adi, K., & Edi Widodo, C. (2023). Identification Texture of Rice Varieties by Feature Extraction using GLCM. E3S Web of Conferences, 448. https://doi.org/10.1051/e3sconf/202344802063
Suarti, B. (2024). Sifat Fisikokimia Beras Pecah Kulit dan Beras Sosoh pada Beberapa Varietas. Jurnal Pangan, 33(1), 1–6. https://doi.org/10.33964/jp.v33i1.756
Unajan, M. C., Gerardo, B. D., & Medina, R. P. (2019). A modified otsu-based image segmentation algorithm (OBISA). Lecture Notes in Engineering and Computer Science, 2239(1), 363–366.
Xu, D., Li, X., & Zhang, X. (2025). Brightness-based Threshold-Weighted OTSU Method for Image Segmentation. Journal of Image Processing Theory and Applications, 8(1), 8–16. https://doi.org/10.23977/jipta.2025.080102
Yu, Y., Wang, C., Fu, Q., Kou, R., Huang, F., Yang, B., Yang, T., & Gao, M. (2023). Techniques and Challenges of Image Segmentation: A Review. Electronics (Switzerland), 12(5).
