Deep Learning-Based Yield Prediction Using Data Collected from Soil Sensors in Real Time

IPEM Journal Computer Applications > Deep Learning-Based Yield Prediction Using Data Collected from Soil Sensors in Real Time

Deep Learning-Based Yield Prediction Using Data Collected from Soil Sensors in Real Time

Mr. Aakash Pundhir, Dr. Waseem Ahmad

Vol 10 , Issue 1 , December 2025 | Pages: 52-64

DOI:

Published Online: December, 2025

Download Article

Author Details

1. Mr. Aakash Pundhir (Research Scholar), Vishveshwarya Group of Institutions, Dadri, Ghaziabad, Uttar Pradesh, India
2. Dr. Waseem Ahmad(HoD, Department of Computer Science & Engineering), Vishveshwarya Group of Institutions, Dadri, Ghaziabad, Uttar Pradesh, India

Accurate yield prediction is an important basis for modern precision agriculture, ensuring optimized resource utilization, timely decision-making, and sustainable farming. While newer research on artificial intelligence has improved yield forecasting by machine learning and deep learning techniques, many of the existing approaches rely heavily on remote sensing imagery and completely lack continuous soil-level information. For these lacunae, this paper proposes a novel deep-learning-based yield prediction framework driven by real-time data from soil sensors. The proposed system would collect multi-parameter data, such as soil moisture, pH, temperature, electrical conductivity, and nutrient composition (NPK), through field-deployed IoT sensors. This paper designs an LSTM–CNN hybrid architecture to model the temporal dependencies and nonlinearities in sensor data, while proposing dimensionality reduction to improve feature extraction and denoise the dataset. Experimental assessment using field-simulation datasets establishes that the proposed model arrives at a prediction accuracy of 92.8%, outperforming traditional regression-based models and unassisted deep learning architectures. The results show that deep learning integrates with real-time soil monitoring and significantly improves yield estimation reliability, supports data-driven agricultural decision making, and provides a scalable framework for future smart farming systems.

Keywords

Deep learning, cropyiel dprediction, soilsensors, IoT agriculture, LSTM, CNN, hybrid model, real-time monitoring, precision farming, smart agriculture, time-series analysis, sensor fusion, sustainable agriculture.

  • van Klompenburg, T., Kassahun, A., &Catal, C. (2020). Crop yieldpredictionusingmachinelearning:Asystematicreview.ComputersandElectronics in Agriculture. Available at: ScienceDirect Link.
  • Hashemi-Beni,L.,etal.(2020).Deeplearningforremotesensingimageanalysis in agriculture. ISPRS Archives. Available at: ISPRS Link.
  • Kamilaris, A., &Prenafeta-Boldu´, F. X. (2020). Deep learning in agri-culture: A survey. Computers and Electronics in Agriculture. Availableat: ScienceDirect Link.
  • Chlingaryan, A., et al. (2020). Machine learning approaches for cropyield prediction. Remote Sensing. Available at: MDPI Link.
  • Sharma, P., et al. (2021). Deep CNN models for crop disease diagnosis.IEEE Access. Available at: IEEE Link.
  • Koirala,A.,etal.(2021).Deeplearningforfruitdetectionandcounting.Sensors. Available at: MDPI Link.
  • Khaki, S., & Wang, L. (2021). Crop yield prediction using deep neuralnetworks. Frontiers in Plant Science. Available at: Frontiers Link.
  • Rustowicz,S.,etal.(2021).Semanticsegmentationforagriculturalfieldanalysis. arXiv. Available at: arXiv Link.
  • Islam,M.M.,etal.(2023).Deeplearning-basedcropdiseasepredictionwithwebinterface.ComputersandElectronicsinAgriculture.Availableat: ScienceDirect Link.
  • Attri, I., et al. (2023). Review of deep learning models in smartagriculture. Ecological Informatics. Available at: ScienceDirect Link.
  • Patidar, S., et al. (2023). Hybrid CNN–LSTM for crop yield prediction.Neural Computing and Applications. Available at: Springer Link.
  • Zhang, R., et al. (2024). Bibliometric analysis of deep learning in cropmonitoring. Remote Sensing. Available at: PubMed Link.
  • Wang, Y., et al. (2024). Advances in deep learning-based crop yieldprediction. Agronomy. Available at: MDPI Link.
  • Lee, H., et al. (2024). Deep learning-based crop disease detection.Applied Sciences. Available at: MDPI Link.
  • Subramaniam,L.K.,etal.(2024).Cropyieldpredictionusingdimension-ality reduction + DL. Intelligent Systems with Applications. Availableat: ScienceDirect Link.
  • Dahiya, N., et al. (2024). U-Net v5 change detection in agriculture.Heliyon. Available at: ScienceDirect Link.
  • Rahman, A., et al. (2022). Transformer-based crop yield forecasting.IEEE Access. Available at: IEEE Link.
  • Khan,F.,etal.(2022).IoTanddeeplearningforreal-timecropmonitoring. Sensors. Available at: MDPI Link.
  • Ramesh, S., et al. (2023). Fusion of satellite and soil-sensor data usingDL.GIScience&RemoteSensing.Availableat:Taylor&FrancisLink.
  • Yu,X.,etal.(2023).Federatedlearningforpredictiveagriculture.InformationProcessinginAgriculture.Availableat:ScienceDirectLink.
  • Mehdipour, V., et al. (2024). DL-based phenotyping and biomass esti-mation. Plants (MDPI). Available at: MDPI Link.
  • Zhou, Y., et al. (2020). DL-based soil moisture estimation using IoTsensors. Sensors. Available at: MDPI Link.
  • Ghosal,S.,etal.(2021).PlantstressdetectionusingCNNs.Frontiersin Plant Science. Available at: Frontiers Link.
  • Kachamba, D.J., et al. (2024). DL for vegetation index estimation from UAV imagery. Remote Sensing. Available at: MDPI Link.
  • Wen, W., et al. (2025). Next-generation AI for smart farming and sensor networks. Artificial Intelligence in Agriculture. Available at: ScienceDirect Link.
  • Kaggle Dataset — Crop Yield Prediction (2024). Crop Yield Prediction Dataset including soil parameters, weather data, and crop output. Available at: Kaggle Link.

A. Pundhir and W. Ahmad, “Deep learning-based yield prediction using data collected from soil sensors in real time,” IPEM Journal of Computer Application & Research, vol. 10, pp. 52–64, Dec. 2025.