Skip to content

Hybrid Quantum Deep Learning and Machine Learning Framework for Soil Image Analysis with XGBoost-based Crop and Random Forest Fertilizer Recommendation

Oct 2026 · Recent Patents on Engineering · 0 citations

Abstract

Farm productivity largely depends on soil conditions, choosing the right crops, and using fertilizers in a balanced way. This work presents a multi-stage, AI-driven decision support system that integrates quantum computing with classical machine learning, leveraging quantum circuits for increased feature representation. To address the challenges of accurate soil analysis and crop selection, an intelligent agricultural decision support system is developed to classify soil categories from images and provide suitable crop and fertilizer recommendations. Multiple machine learning models were integrated into a single web application, enabling real-time analysis and decision support for farmers. The HQCNN (Hybrid Quantum Convolutional Neural Network) model classifies soil types from input images in this system. It processes soil images using CNN layers for feature extraction and a quantum layer for advanced feature transformation. By combining CNN with quantum learning, it accurately predicts soil type. The XGB Classifier model acts as the crop recommendation engine in this system. It takes soil nutrients, climate parameters, and encoded soil type as input and predicts the most suitable crop. Learning patterns from historical data using gradient boosting provides accurate multi-class crop classification based on environmental conditions. Random Forest (RF) Classifier serves as the fertilizer recommendation engine in this system. It takes inputs such as moisture, recommend crop type, NPK values, encoded soil type, temperature, and humidity to predict the most suitable fertilizer. The HQCNN model achieved 97.62% training accuracy, 98.83% validation accuracy, and 98.25% test accuracy. Additionally, the model demonstrated excellent classification capability with a micro-average ROC (AUC = 1.00) and macro-average ROC (AUC = 1.00) score, indicating strong generalization performance. High accuracy validates that HQCNN performs efficiently in soil classification. Combining it with XGBClassifier and RF enhances prediction capability. The system has more reliable crop and fertilizer recommendations. It provides clear insight into soil nutrient status. Web application integration improves usability for real-world farming use. The proposed system demonstrated high accuracy and practical applicability, offering a valuable tool for farmers to optimize crop selection and improve productivity. The work is potentially useful for ML-based soil classification and crop recommendation.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.