Enhancing Disease and Pest Detection in Greenhouse Tomato Cultivation Using Advanced Machine Learning on New Dataset of Images

Increasing food production is a continuous need. In this context, agriculture is a fundamental part of meeting the ever-increasing demand for food. Plant diseases are one of the factors that compromise food production goals, and the characteristics and climate of each production region influence them. Tomatoes are one of the world's most consumed vegetables and are widely affected by various diseases. However, tomato cultivation in greenhouses allows its continuous production. In this context, this research work focuses on the problem of identifying diseases in scenarios of tomato cultivation in greenhouses, where we have specific occurrences of diseases that are affected by regional climatic conditions. Brazil is a major producer of tomatoes, producing more than 3 million tons annually, with 8% of this production being made in the state of Paraná. This study was developed through data collection in collaboration with greenhouse tomato producers from an important region in North Paraná. For this study, we created new datasets with two image sizes: the Tomato Leaf Image Dataset (TLID) with image sizes of 256x256 pixels and 15,256 images, and the Patch Tomato Leaf Image Dataset (PTLID) with patch sizes of 32x32 pixels and 227,218 images. Both datasets comprise seven classes, including four types of diseases, two combinations of diseases on the same leaf, and the healthy leaf. Machine Learning techniques have been widely used to identify plant diseases. This work presents two machine learning methods tested with both datasets. In the proposed models, we combine three convolutional neural networks, a customized CNN, VGG19, and Resnet50, and two voting classification methods using hard and soft decisions. The evaluation performed on the datasets showed that when the patches are used, the results improve significantly, reaching an accuracy of 90.48%. It is also possible to identify the stage of the disease.