Detection of Diseases in Malvaceae Family plants using Enhanced Deep Learning Algorithm with Color Level Descriptor

Abstract

The precise and prompt identification of plant diseases constitutes a crucial element in maintaining robust crop production, particularly with regard to ornamental and economically valuable species within the Malvaceae family. This study introduces an advanced deep learning-based methodology for the identification of diseases in Malvaceae leaf images by incorporating a tailored Convolutional Neural Network (CNN) alongside Color Level Descriptor (CLD) feature extraction. The CLD technique enhances the input dataset by capturing spatial color attributes, thereby significantly augmenting the model's capability to differentiate between healthy and diseased leaf patterns. The system underwent training and validation on a meticulously curated dataset containing images of diverse species from the Malvaceae family, exhibiting enhanced accuracy and resilience relative to traditional CNN models. Experimental findings indicate that the integration of CLD facilitates more precise feature representation and superior classification efficacy. This innovative approach holds substantial promise for practical implementation in agricultural diagnostics, fostering early detection and effective management of plant diseases affecting the Malvaceae family.