Optimized Tomato Rot Disease Severity Profiling: A Hybrid CNN-Random Forest Algorithm for Five-Tier Categorization

Tomato agriculture is often plagued by many illnesses that adversely affect the health of plants and reduce crop output. This project aims to create and validate an advanced tomato disease severity scale that accurately differentiates between five distinct stages of sickness in tomatoes. In our study, we employed a sophisticated cross-breeding method that involved the integration of Convolutional Neural Networks (CNN) and Random forests (RF) to achieve a high level of accuracy. Our model achieved an amazing overall accuracy of 96.81% at that point. The model's exceptional accuracy demonstrates its proficiency in accurately distinguishing between the five severity ratings. Through this methodology, we have established distinct connections between mild indications and symptoms, which progressively escalate in intensity, ultimately culminating in the emergence of potentially fatal ailments. While our model outperforms existing operational modes, we acknowledge certain inherent limitations due to minor variations in disease pathology and image quality, which may hinder the identification of subtle differences in severity. We must optimize the model's sensitivity to intermediate severity levels and get extensive datasets that encompass various degrees of severity. This work will play a crucial role in evaluating the severity of diseases in tomato cultivation, with groundbreaking implications for precision agriculture. The objective is to enhance crop management strategies to minimize crop loss and enhance agricultural resilience against a wide range of disease intensities, therefore ushering in a new era of disease mitigation.