Leaf Disease Classification and Crop Damage Estimation using Advanced Deep Learning Models

Current soil classification systems provide accuracy outperforming 90% displaying the efficiency of deep learning. These systems that provide advice maximize crop yield and reduce Implementation of resources in evaluations. This paper introduces hybrid deep learning model for soil image classification and automatic crop recommendation to improve agricultural productivity. By leveraging Convolutional Neural Networks (CNNs), particularly MobileNetV2 and ResNet50, the proposed hybrid model extracts complex visual features from soil images to classify them into predefined types. This classification is used to recommend optimal crops established regarding soil features, such as surface quality, pH levels, and water retention capacity. The system addresses the limitations of traditional soil classification methods, which are prolonged and prone to fault. Empirical outcomes using the Soil_Data_V3 dataset illustrate that the hybrid model exceeds separate models in accuracy, abstraction, and reducing validation loss. This study contributes to precision agriculture by systematizing soil classification and crop recommendation or optimizing crop yield and enhancing maintainable farming practices.