A Smart Agricultural AI Model For RGB Image-Based Disease Detection On Apple Trees

Abstract

The early and accurate detection of plant diseases is crucial because of its contribution to socioeconomic growth in agricultural productivity and worldwide food security. Traditional methods of plant disease detection often depend upon time-consuming, intensive research surveys and onsite field inspections, which are time-consuming and liable to human error. In the last few decades, the incorporation of imaging technology with automated artificial intelligence (AI) algorithms has appeared as a promising answer, allowing speedy and accurate early identification of plant diseases. In this work, an automated framework is developed to identify and classify diseases in apple plants at the right time to reduce financial loss and human labor. However, advancements in sensor technology, information analytics, and artificial intelligence algorithms continue to enhance smart agriculture. In this work, we have used a multispectral dataset analyzing grayscale and RGB sample images with preprocessing, and classification to discover apple leaf illnesses. Color spatial capabilities have been recognized as crucial for assessing the severity of apple plant species infections. Our findings indicate that blue channel color space supplied better clarity and noise-unfastened outputs, making them more effective for detecting diseased leaves than other color space channels and grayscale images. Two AI-based models, Random Forest and Convolutional Neural Networks (CNNs) were fine-tuned and used for disease detection. The CNN model outperformed Random Forest, achieving an accuracy of 89.05%, precision of 90.71%, remember of 89.05%, and an F1 score of 89.87%. These effects underscore the high functionality of CNNs to hit upon and classify plant diseases with precision while minimizing false positives and negatives. The integration of CNNs into RGB channel color space detection workflows facilitate early diagnosis and timely interventions, improving plant control, safeguarding yield, and promoting agricultural sustainability.