MULTI-LEAF DISEASE DETECTION AND RECOMMENDS PESTICIDE USING DEEP LEARNING

Le résumé fourni par la source

This research describes a proposed system that uses deep learning and image processing methods to identify and categorize plant leaf diseases. The approach that is suggested comprises of two categorization methods and compares them. The first approach comprised of multiple steps leading up to the classification stage and was based on the support vector machine (SVM) algorithm. Because they are the most prevalent plant species worldwide and in Iraq specifically, tomatoes, peppers, and potatoes are the specific plant species that we use in our work. Convolution neural networks (CNNs) were employed in the second approach for classification. Fifteen classes were identified using these two methods: three classes for healthy leaves and twelve classes for illnesses of various plants that were found, such as fungi, bacteria, etc. The comparisons outcome demonstrates that the CNN algorithm is preferred over the SVM algorithm in terms of speed and accuracy, creating a system that is reliable and accurate for the identification and categorization of plant leaf diseases. The effectiveness of the system is evaluated using evaluation measures such as accuracy, precision, recall, and F1-score. Additionally discussed are the moral and environmental aspects of using pesticides. This strategys practical use is illustrated by case studies and real-world examples, which also highlight how it may be used to increase agricultural yields, lower resource consumption, and support sustainable farming methods. Finally, this study highlights how CNNs can revolutionize both pesticide recommendation and the detection of multi-leaf diseases. The integration of advanced technology with agricultural methods in this study promotes sustainable farming practices and global food security. KEYWORDS: Deep Learning, Convolutional Neural Networks (CNNs), Multi-Leaf Disease Detection, Pesticide Recommendation, Agriculture, Plant Disease Diagnosis, Image Analysis, Data Augmentation, Transfer Learning.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.