Author(s): Pankaj Deoskar and Ajay Kumar Sachan

Abstract: Plant leaf diseases in essential crops such as wheat and rice pose a significant challenge to global food security, demanding timely and accurate detection methods. This study introduces a hybrid deep learning approach that integrates multiple Convolutional Neural Network (CNN) architectures for the early detection of leaf diseases in wheat and rice plants. The process begins with the collection of images followed by image preprocessing steps including resizing and grayscale conversion to ensure uniformity. Texture and intensity-based features are extracted using the Grey Level Co-Occurrence Matrix (GLCM) along with statistical measures like mean and standard deviation. To enhance detection accuracy and robustness, the system combines the strengths of multiple CNN models CNN-2D, VGG16, ResNet50, InceptionV3, and MobileNet in a hybrid and integrated framework. This ensemble-like integration enables the model to learn both low-level and high-level discriminative features more effectively. The models are evaluated using key metrics such as accuracy, precision, recall, F1-score, and error rate, with comparative graphs illustrating the performance across two distinct datasets.

DOI: 10.33545/27075923.2025.v6.i1a.89

Pages: 35-50 | Views: 418 | Downloads: 167

Download Full Article: Click Here