Integrating Sparrow Search Algorithm with Deep Learning for Tomato Fruit Disease Detection and Classification

Lately, the agricultural industry has considerably increased the applications of deep learning (DL) systems for recognizing and alleviating tomato fruit diseases, which address the major problems in food security and crop management. Leveraging the power of DL approaches, practitioners and researchers have established more efficient solutions capable of accurately detecting and classifying different tomato fruit diseases, enabling timely intervention to safeguard crop health and prevent yield losses. The DL techniques have been trained to detect subtle visual cues indicative of common diseases, including late blight, bacterial spot, and early blight, by harnessing vast datasets comprising annotated images of diseased tomato fruits. This sophisticated detection system offers a high level of performance. It demonstrates adaptability and scalability over different environmental conditions, empowering the farmers with actionable insight to optimize agricultural productivity and deal with disease outbreaks successfully. This study introduces a Sparrow Search Algorithm with Deep Learning for Tomato Fruit Disease Detection and Classification (SSADL-TFDDC) technique. The SSADL-TFDDC technique mainly exploits the optimal DL model to detect diseases in tomato fruits. In the SSADL-TFDDC technique, the bilateral filtering (BF) technique removes noise. The SSADL-TFDDC technique exploits the residual network (ResNet50) model for feature extraction, and the SSA performs the hyperparameter tuning process. At last, a long short-term memory (LSTM) classifier was employed to identify and classify the existence of tomato fruit diseases. A brief set of simulations exemplifies the enhanced solution of the SSADL-TFDDC technique. The experimental validation stated that the SSADL-TFDDC technique accomplishes better detection results.