Desenvolvimento de software aplicativo para diagnóstico de doenças, pragas e distúrbios fisiológicos em tomateiro (Solanum lycopersicum L.)

Abstract

Tomato (Solanum lycopersicum L.) holds significant economic importance in Brazil; however, a wide range of diseases, pests, and physiological disorders can affect its production. Therefore, this study aims to develop a software application for the diagnosis of diseases, pests, and physiological disorders in tomato crops, based on symptoms and signs that are easily identifiable by producers. A database was created in MySQL, containing signs, symptoms, and management recommendations. The interface and diagnostic logic were developed in Java, allowing the producer to select the observed symptoms. The application operates interactively, enabling the producer to gradually select the symptoms observed in their plants. Based on this information, the system analyzes and provides an accurate diagnosis, along with recommendations for appropriate management. The efficiency and functionality of the application were evaluated through usability and integrity tests. Since Test 1, the software showed satisfactory usability performance, with 100% approval regarding its intuitive interface and response speed. The main challenge was related to agronomic information. In Test 2, diagnostic accuracy increased from 5% to 95%, and agreement with management suggestions rose from 5% to 80%, with a significant decrease in errors. The application received 100% recommendation. Despite not achieving 100% accuracy, the errors were occasional, with some problems being difficult to diagnose due to symptoms similar to other issues. The evolution of the application increased its accuracy, speed, and usability, making it more reliable and recommended. Thus, it can be concluded that the application has great potential to assist in the identification and management of phytosanitary problems, with possibilities for expansion to other crops and integration with machine learning to further enhance diagnostic accuracy.