Multimodal Deep Learning Framework for Crop Prediction Using Soil Image and Textual Feature with Convolutional Neural Network and Hybrid Convolutional Graphical Neural Network

Abstract Proper classification of soils is essential for the choice of appropriate crops and promoting agricultural productivity. In this work, Convolutional Neural Network (CNN) and Hybrid Convolutional Graphical Neural Network (HCGNN) models are compared for the classification of soil images, which is used to determine the best model to predict crop growth. The soil images were collected from Kaggle and were pre-processed into a dataset of 1156 images across 5 soil classes: Alluvial, Black, Clayey, Red, and Sandy before training a model. The CNN model takes off the visual features of the soil image, while the HCGNN transforms the convolutional feature abstraction with graph-based learning, and integrates the complex spatial relationships between the image features. From the experimental results, it is found that the HCGNN has a advanced accuracy of 97.50% on the test set compared with CNN with 87.41%. Based on the soil types predicted by the selected best performing model, appropriate crops are recommended which offers a reliable decision-support classification for precision of agriculture. The outcomes display that the graph-based deep learning has better classification accuracy and fewer misclassifications compared to the conventional CNN-based methods. This planned framework delivers an efficient and practical solution for soil analysis and crop recommendations that can be implemented in automated systems and it could be used in smart farming and sustainable agricultural decision-making.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-17
DOI
https://doi.org/10.21203/rs.3.rs-11371903/v1
Primary Topic
Smart Agriculture and AI
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Multimodal Deep Learning Framework for Crop Prediction Using Soil Image and Textual Feature with Convolutional Neural Network and Hybrid Convolutional Graphical Neural Network

Bijay Kumar Paikaray, Trilochan Rout, Mamata Garanayak, Chinmaya Kumar Naik et al.
2 links
Research Square
Smart Agriculture and AI
preprint

Multimodal Deep Learning Framework for Crop Prediction Using Soil Image and Textual Feature with Convolutional Neural Network and Hybrid Convolutional Graphical Neural Network

Bijay Kumar Paikaray, Trilochan Rout, Mamata Garanayak, Chinmaya Kumar Naik, Ramesh Patra
preprint en

Abstract

Abstract Proper classification of soils is essential for the choice of appropriate crops and promoting agricultural productivity. In this work, Convolutional Neural Network (CNN) and Hybrid Convolutional Graphical Neural Network (HCGNN) models are compared for the classification of soil images, which is used to determine the best model to predict crop growth. The soil images were collected from Kaggle and were pre-processed into a dataset of 1156 images across 5 soil classes: Alluvial, Black, Clayey, Red, and Sandy before training a model. The CNN model takes off the visual features of the soil image, while the HCGNN transforms the convolutional feature abstraction with graph-based learning, and integrates the complex spatial relationships between the image features. From the experimental results, it is found that the HCGNN has a advanced accuracy of 97.50% on the test set compared with CNN with 87.41%. Based on the soil types predicted by the selected best performing model, appropriate crops are recommended which offers a reliable decision-support classification for precision of agriculture. The outcomes display that the graph-based deep learning has better classification accuracy and fewer misclassifications compared to the conventional CNN-based methods. This planned framework delivers an efficient and practical solution for soil analysis and crop recommendations that can be implemented in automated systems and it could be used in smart farming and sustainable agricultural decision-making.

Research Square
Siksha O Anusandhan University (IN), SRM University, Andhra Pradesh (IN), Kalinga Institute of Social Sciences (IN)
Zero hunger, Life on land
Smart Agriculture and AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon