Iranian Journal of Biosystem Engineering

Iranian Journal of Biosystem Engineering

A new acoustic sensing approach for predicting the percentage of filled rice grains based on the acoustic absorption spectrum using the Deep Spectra

Document Type : Research Paper

Authors
Majid Fathi Ghalemiri 1
Ali Maleki 2
Majid Lashgari 3
Ali Loghmani 4
1 Ph.D. Student, Department of Mechanical Engineering of Biosystems, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran
2 Associate professor of Mechanical Engineering of Biosystems Department, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran
3 Associate Professor, Department of Mechanical Engineering of Biosystems, Faculty of Agriculture, Arak University, Arak, Iran
4 Associate Professor, Department of Mechanical Engineering, Isfahan University of Technology, Isfahan, Iran
10.22059/ijbse.2024.376335.665548
Abstract
Rice is recognized as one of the main cereals in the world, feeding two-thirds of the global population, especially in Asian countries. Accurate assessment of the percentage of filled grains (PFG) is crucial for the efficiency and quality of rice harvesting. Traditional methods of measuring PFG are practical and based on personal judgment. This study introduces an innovative and non-destructive approach based on an acoustic sensor alongside deep learning models to predict PFG based on the acoustic spectrum of rice grains. Using an advanced deep learning architecture, the Deep Spectrum, which works directly on raw spectral data, eliminates the need for preprocessing and enhances prediction accuracy. A modified impedance tube was used to measure the acoustic spectrum, which was then analyzed using the Deep Spectrum model to predict PFG. Results indicated that this approach significantly improves the quantitative analysis of spectral data and provides a reliable prediction of rice grain filling. The prediction accuracy of the Deep Spectrum model was significantly higher compared to traditional methods, with a low root mean square error of prediction (RMSEP) of 0.24 ± 0.05 and a coefficient of determination (R²) of 0.95 ± 0.02. This prediction is vital for assessing rice quality, breeding, and genetic research. This study introduces new perspectives and methods in the field of grain quality prediction and classification using acoustic spectrum analysis and deep learning, which could be beneficial for future research in this area.
Keywords
Acoustic Absorption Coefficient
Deep Spectra
Non-destructive Measurement
Rice Acoustic Sensor
Spectral Analysis
Subjects

New acoustic sensing approach for predicting the percentage of filled rice grains based on the acoustic absorption spectrum using the Deep Spectra

 

EXTENDED ABSTRACT

Introduction

Rice serves as a staple food for a substantial portion of the global population, particularly in Asian countries. Traditional methods of determining the percentage of filled grains (PFG) in rice are labor-intensive and often rely on subjective judgments. This study introduces an innovative and non-destructive approach that leverages a new acoustic sensing technology combined with deep learning models. The primary innovation of this approach is its use of the acoustic absorption spectrum to predict the PFG without the need for pre-processing the spectral data. This method holds significant promise for enhancing the efficiency and accuracy of rice quality assessments, crucial for agricultural productivity and research in rice breeding and genetics. By employing advanced deep learning architectures, this study aims to overcome the limitations of traditional grain analysis methods, offering a more reliable and scalable solution.

Material and Methods

The research utilizes an altered impedance tube to capture the acoustic spectrum of rice grains. This method measures how sound waves are absorbed by the grains, providing a detailed spectral profile. The acoustic data collected is then analyzed using a deep learning model, specifically designed for this study. The model, referred to as DeepSpectra, operates on the raw spectral data directly, eliminating the need for the usual pre-processing steps that can degrade data quality. This model employs a convolutional neural network (CNN) architecture, which is highly effective in pattern recognition tasks. CNNs utilize layers of filters to process spatial hierarchies in data, making them ideal for handling complex spectral data. This approach not only enhances the prediction accuracy but also significantly speeds up the analysis process by leveraging the model's ability to learn and generalize from the data presented.

Results and Discussion

The results of this study were highly promising, demonstrating that the deep learning model, DeepSpectra, significantly outperformed traditional methods in predicting the percentage of filled grains in rice. The model achieved high accuracy, showcasing its capability to handle the raw spectral data effectively without the need for manual preprocessing. This is a considerable advancement over traditional grain analysis methods, which are slower and often less reliable. The discussion also highlights how the model's deep learning architecture, which includes multiple convolutional and pooling layers, contributes to its robust performance by effectively extracting and learning important features from complex datasets. The study further discusses the potential applications of this technology in the field of agricultural research and grain quality assessment, suggesting that it could be a valuable tool for rice breeders and geneticists looking to improve crop yields and grain quality.

Conclusion

The study concludes that the novel use of acoustic sensing technology combined with a deep learning analysis model offers a powerful and non-destructive method for accurately predicting the percentage of filled grains in rice. The acoustic absorption spectrum, as analyzed by the DeepSpectra model, provides a reliable indicator of grain quality, which is essential for rice breeding and genetic research. The implications of these findings are significant, as they suggest a shift towards more automated, precise, and efficient methods of grain analysis. This could lead to improvements in rice production strategies and better food security globally. Additionally, the study underscores the potential for expanding this technology to other grains and agricultural products, which could broadly revolutionize quality assessment in agriculture. The researchers highlight the need for further studies to refine the technology and fully realize its potential across different environments and grain types.

A|Lab - File Exchange - MATLAB Central https://www.mathworks.com/matlabcentral/fileexchange/93180-a-lab.
Allegrini, F. & Olivieri, A. C. (2016). Anal Chem 88, 7807–7812.
Boubel, A., Bousshine, S., Garoum, M., Abdelmajid, B. & Laaroussi, N. (2016).
Burlando, B. & Cornara, L. (2014). Trends Food Sci Technol 40, 82–98.
Caballol, D. & Raposo, Á. P. (2020). Constr Build Mater 256, 119297.
Ceyhan, M., Kartal, Y., Özkan, K. & Seke, E. (2023). Multimed Tools Appl 1–23.
Duan, L., Yang, W., Bi, K., Chen, S., Luo, Q. & Liu, Q. (2011). Comput Electron Agric 75, 196–203.
Gasso-Tortajada, V., Ward, A. J., Mansur, H., Brøchner, T., Sørensen, C. G. & Green, O. (2010). Sensors (Basel) 10, 10027–10039.
Geimer, P. R., Ulrich, ; T J, Luke, ;, Beardslee, B., Hayne, M. L., Remillieux, M. C., Saleh, T. A., Freibert, F. J., Ulrich, T. J. & Beardslee, L. B. (2022). J Acoust Soc Am 151, 3633–3640.
Gheibi, A. & Hedayat, A. (2018). Ultrasonics 87, 112–125.
Griffiths, S., Nennig, B. & Job, S. (2017). J Acoust Soc Am 141, 254–264.
Hastie, T., Tibshirani, R. & Friedman, J. (2009). https://doi.org/10.1007/978-0-387-84858-7.
Huang, S. ;, Lu, Z. ;, Shi, Y. ;, Dong, J. ;, Hu, L. ;, Yang, W. ;, Huang, C., Huang, S., Lu, Z., Shi, Y., Dong, J., Hu, L., Yang, W. & Huang, C. (2023). Sensors 2023, Vol. 23, Page 6331 23, 6331.
Ioffe, S. & Szegedy, C. (2015). 32nd International Conference on Machine Learning, ICML 2015 1, 448–456.
ISO 10534-1:1996 - Acoustics — Determination of sound absorption coefficient and impedance in impedance tubes — Part 1: Method using standing wave ratio https://www.iso.org/standard/18603.html.
ISO 10534-2:2023 - Acoustics — Determination of acoustic properties in impedance tubes — Part 2: Two-microphone technique for normal sound absorption coefficient and normal surface impedance https://www.iso.org/standard/81294.html.
Kamboj, U., Guha, P. & Mishra, S. (2021). Mater Today Proc 48, 576–582.
Kingma, D. P. & Ba, J. L. (2014). 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings.
Koulivand, K. & Moniri, | Kamel.
Kumar, A., Taparia, M., Madapu, A., Rajalakshmi, P., Marathi, B. & Desai, U. B. (2020). J Cereal Sci 95, 103037.
Lecun, Y., Bengio, Y. & Hinton, G. (2015). Nature 2015 521:7553 521, 436–444.
Lee, C. M. & Wang, Y. S. (2006). J Sound Vib 298, 350–365.
Lee, C. M. & Xu, Y. (2009). J Sound Vib 326, 290–301.
Li, Z., Chen, S., Wu, Z. & Yang, L. (2020). Https://Doi.Org/10.1177/1077546320926905 27, 332–342.
Lin, P., Chen, Y. M., He, Y. & Hu, G. W. (2014). Comput Electron Agric 109, 124–133.
Liu, T., Wu, W., Chen, W., Sun, C., Chen, C., Wang, R., Zhu, X. & Guo, W. (2016). Biosyst Eng 150, 79–88.
Liu, X., An, H., Cai, W. & Shao, X. (2024a). TrAC Trends in Analytical Chemistry 172, 117612.
Liu, X., An, H., Cai, W. & Shao, X. (2024b). TrAC Trends in Analytical Chemistry 172, 117612.
Maas, A. L., Hannun, A. Y. & Ng, A. Y. (2013).
Manickavasagan, A., Sathya, G., Jayas, D. S. & White, N. D. G. (2008). J Cereal Sci 47, 518–527.
Mosa, A. I., Putra, A., Ramlan, R., Prasetiyo, I. & Esraa, A. A. (2019). Applied Acoustics 146, 409–419.
Muthayya, S., Sugimoto, J. D., Montgomery, S. & Maberly, G. F. (2014). Ann N Y Acad Sci 1324, 7–14.
Nguyên, T. T. T., Doanh, T., Bot, A. Le & Dalmas, D. (2022). Scientific Reports 2022 12:1 12, 1–9.
Ni, W., Nørgaard, L. & Mørup, M. (2014). Anal Chim Acta 813, 1–14.
Ostrovsky, L., Lebedev, A., Matveyev, A., Potapov, A., Sutin, A., Soustova, I. & Johnson, P. (2001). J Acoust Soc Am 110, 1770–1777.
Passos, D. & Mishra, P. (2022). Chemometrics and Intelligent Laboratory Systems 223, 104520.
Rinnan, Å., Berg, F. van den & Engelsen, S. B. (2009). TrAC Trends in Analytical Chemistry 28, 1201–1222.
Roozen, N. B. & Piana, E. A. (2022). Applied Acoustics 198, 109002.
Stender, M., Adams, C., Wedler, M., Grebel, A. & Hoffmann, N. (2021). J Acoust Soc Am 149, 1932–1945.
Sun, M., Zhang, D., Liu, L. & Wang, Z. (2017). Food Chem 218, 413–421.
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V. & Rabinovich, A. (2015). 1–9.
Tsuruha, T., Yamada, Y., Otani, M. & Takano, Y. (2020). J Acoust Soc Am 147, 3418–3428.
Yang, J., Peng, S., Visperas, R. M., Sanico, A. L., Zhu, Q. & Gu, S. (2000). Plant Growth Regul 30, 261–270.
Zhang, J., Dai, L. & Cheng, F. (2021a). Journal of Food Measurement and Characterization 15, 484–494.
Zhang, J., Dai, L. & Cheng, F. (2021b). Food Anal Methods 14, 389.
Zhang, X., Lin, T., Xu, J., Luo, X. & Ying, Y. (2019). Anal Chim Acta 1058, 48–57.
 
Iranian Journal of Biosystem Engineering
Volume 54, Issue 4
Winter 2024
Pages 87-102

  • Receive Date 10 May 2024
  • Revise Date 17 July 2024
  • Accept Date 29 July 2024