Machine learning techniques for real-time tillage quality assessment

B  PRIYADHARSHINI; S  THAMBIDURAI; P K  PADMANATHAN; KAMARAJ  P; PATIL SANTOSH  GANAPATI; R  THIYAGARAJAN

doi:10.56093/ijas.v95i7.161563

Authors

B PRIYADHARSHINI Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India
S THAMBIDURAI Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India
P K PADMANATHAN Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India
KAMARAJ P Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India
PATIL SANTOSH GANAPATI Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India
R THIYAGARAJAN Tamil Nadu Agriculture University, Coimbatore, Tamil Nadu 641 003, India

https://doi.org/10.56093/ijas.v95i7.161563

Keywords:

Image analysis, Machine learning, Random forest, Real-time assessment, Tillage quality

Abstract

Tillage quality significantly impacts soil structure, nutrient availability, and crop growth. The present study was carried out in 2024 in the delta region of Tamil Nadu at Puvalur village, Trichy District to evaluate tillage quality based on high-resolution soil images through a machine-learning approach using a Random Forest (RF) model. Key soil properties such as circularity, aspect ratio, and solidity were analyzed, with the model achieving an R² score of 0.86 and a Mean Squared Error (MSE) of 0.0023. Data from 100 soil images were processed using ImageJ software version 1.54 g. The Random Forest model was chosen for its robustness, interpretability, scalability, and efficiency in handling complex, real-time data. Also, the RF model provided accurate, real-time insights, reducing labor and optimizing resource use, promoting sustainability in agriculture across diverse soil conditions.

Downloads

Download data is not yet available.

References

Blanco‐Canqui H and S Ruis. 2018. No-tillage and soil physical environment. Geoderma 326: 164–200. doi: 10.1016/J. GEODERMA.2018.03.011

Chandra P, Khippal A, Prajapat K, Barman A, Geeta S, Rai A, Ahlawat O P, Verma R, Kumari K and Singh G. 2023. Influence of tillage and residue management practices on productivity, sustainability, and soil biological properties of rice-barley cropping systems in indo-gangetic plain of India. Frontiers in Microbiology 14. doi: 10.3389/fmicb.2023.1130397

Chen R, Christine D, Su-Wen H and Caraka R. 2020. Selecting critical features for data classification based on machine learning methods. Journal of Big Data 7: 1–26. doi: 10.1186/ s40537-020-00327-4

Cooper R, Hama-Aziz Z Q, Hiscock K, Lovett A, Vrain E, Dugdale S, Sünnenberg G, Dockerty T, Hovesen P and Noble L. 2020. Conservation tillage and soil health: Lessons from a 5-year UK farm trial (2013–2018). Soil and Tillage Research 202: 104648. doi: 10.1016/j.still.2020.104648

Cueff S, Coquet Y, Aubertot J, Bel L, Pot V and Alletto L. 2021. Estimation of soil water retention in conservation agriculture using published and new pedotransfer functions. Soil and Tillage Research 209: 104967. doi: 10.1016/J.STILL.2021.104967

Elavarasan D and Vincent P M. 2021. A reinforced random forest model for enhanced crop yield prediction by integrating agrarian parameters. Journal of Ambient Intelligence and Humanized Computing 1–14. doi: 10.1007/s12652-020-02752-y

Igor B, Pereira P, Kisic I, Krunoslav S and Sraka M. 2018. Tillage management impacts on soil compaction, erosion and crop yield in Stagnosols (Croatia). Catena 160: 376–84. doi: 10.1016/J. CATENA.2017.10.009

Kraut-Cohen J, Avihai Z, S H Liora, A Eli, R Rabinovich, S Green and D Minz. 2019. Effects of tillage practices on soil microbiome and agricultural parameters. The Science of the Total Environment 135791. doi: 10.1016/j.scitotenv.2019.135791

Kushal B J, Nikhil J S P, Nikil S Raaju, Kushal Gowda G V, Asha Rani K P and Gowrishankar S. 2022. Real time crop prediction based on soil analysis using internet of things and machine learning. (In) 2022 International Conference on Edge Computing and Applications, Institute of Electronics and Electrical Engineers, pp. 1249–54.

Lal R. 1991. Tillage and agricultural sustainability. Soil and Tillage Research 20: 133–46. doi: 10.1016/0167-1987(91)90036-W

Luz F B d, V R d Silva, F J K Mallmann, C A B Pires, H Debiasi, J C Franchini and M Cherubin. 2019. Monitoring soil quality changes in diversified agricultural cropping systems by the Soil Management Assessment Framework (SMAF) in southern Brazil. Agriculture, Ecosystems and Environment. doi: 10.1016/J.AGEE.2019.05.006

Mallah S, K Bahareh Delsouz, Davatgar N, Scholten T, A C Alireza, E Mostafa, R Kerry, M Amir and R Taghizadeh‐Mehrjardi. 2022. Predicting Soil Textural Classes Using Random Forest Models: Learning from Imbalanced Dataset. Agronomy. doi: 10.3390/agronomy12112613.

Nguyen Q, Ly H, Ho L, Al‐Ansari, N Le, H Tran V, Prakash I and Pham B. 2021. Influence of data splitting on performance of machine learning models in prediction of shear strength of soil. Mathematical Problems in Engineering, pp. 1–15. https://doi.org/10.1155/2021/4832864

Nunes M R, Harold M van Es, R Schindelbeck, R Aaron and M Ryan. 2018. No-till and cropping system diversification improve soil health and crop yield. Geoderma. doi: 10.1016/J. GEODERMA.2018.04.031

Pal M. 2005. Random forest classifier for remote sensing classification. International Journal of Remote Sensing 26: 217–22. doi: 10.1080/01431160412331269698

Paul G, Saha S and Ghosh K. 2020. Assessing the soil quality of Bansloi river basin, eastern India using soil-quality indices (SQIs) and Random Forest machine learning technique. Ecological Indicators 118: 106804. doi: 10.1016/j.ecolind.2020.106804

Putte A, G Gerard, J Diels, K Gillijns and M Demuzere. 2010. Assessing the effect of soil tillage on crop growth: A meta- regression analysis on European crop yields under conservation agriculture. European Journal of Agronomy 33: 231–41. doi: 10.1016/J.EJA.2010.05.008

Rodriguez-Galiano, V M Sanchez-Castillo, M Chica-Olmo and M Chica-Rivas. 2015. Machine learning predictive models for mineral prospectivity: An evaluation of neural networks, random forest, regression trees and support vector machines. Ore Geology Reviews 71: 804–18. doi: 10.1016/J. OREGEOREV.2015.01.001

Shahid Nawaz, K L Dapeng and M Maimaitijiang. 2022. A geographically weighted random forest approach to predict corn yield in the US corn belt. Remote Sensing 14: 2843. doi: 10.3390/rs14122843

Taghizadeh‐Mehrjardi R., M Mahdianpari, F Mohammadimanesh, T Behrens, N Toomanian, T Scholten and K Schmidt. 2020. Multi-task convolutional neural networks outperformed random forest for mapping soil particle size fractions in central Iran. Geoderma 376: 114552. doi: 10.1016/j.geoderma.2020.114552

Tesfaye A, X Wenbo and F Jinlong. 2022. Comparison of random forest and support vector machine classifiers for regional land cover mapping using coarse resolution FY-3C Images. Remote Sensing 14: 574. doi: 10.3390/rs14030574

Were K, D Bui D T, O B Dick, and Singh B R. 2015. A comparative assessment of support vector regression, artificial neural networks, and random forests for predicting and mapping soil organic carbon stocks across an Afromontane landscape. Ecological Indicators 52: 394–403. doi: 10.1016/J. ECOLIND.2014.12.028

Yuan L, L Zhou, S Cui, S Jagadamma and Z Qingping. 2019. Residue retention and minimum tillage improve physical environment of the soil in croplands: A global meta-analysis. Soil and Tillage Research. doi: 10.1016/J.STILL.2019.06.009

Zizheng D, H Mingjing, Z Wuping, W Guofang, H Xuefang, L Gaimei and L Nana. 2023. Effects of five years conservation tillage for hedging against drought, stabilizing maize yield, and improving soil environment in the drylands of northern China. PLoS One 18. doi: 10.1371/journal.pone.0282359