Empowering Forestry Conservation with AI: Severity Classification of Sequoia Canker Using Siamese CNNs

The rise of plant diseases recognized to be a great menace to global biodiversity and forestry must call for modern methods of diagnostic and classification tools. Within the list of these diseases, the most notable one is the Sequoia cypress canker, which is a pathogenic infection caused by the organism Seiridium company and its impacts on the health and lifespan of Sequoia cypress trees are very important for both ecological balance and the industries connected with this tree as a raw material. Conventional approaches that involve disease detection and classification of severity levels can be both tedious and inaccurate due to human errors, i.e. subjective decisions. This research will make use of one of the innovative applications of the Siamese Convolutional Neural Networks (Siamese CNNs) which is the classification system of varying the levels of severity in the Sequoia Canker a disease in the cypress trees, that uses the network in recognizing fine differences in the manifestations of the disease. The dataset preparation stems from heterogenic sources, the design of a Siamese CNN which adapts to the assignment, and an extensive training and evaluation process as well. The model proved to be a better predictor of the disease severity, discriminating the subject into five levels of disease which have an accuracy of 92.9%, outperforming both the traditional CNN models and manual classification methods in metrics such as accuracy, precision, recall, and f1 score. The findings reveal that Siamese CNN is an outstanding tool to transform the way we classify plant disease severities, providing us with scalable and effective work that does not get influenced by the experts' personal bias. Firstly, the conducted research is an addition to the expanding collection of literature at the junction between artificial intelligence and environmental science, though it can be learned that appropriate tools can solve key issues of plant pathology and forestry management. The successful application of CNN for disease severity distinction suggests further use of these AI tools for addressing environmental problems. Thus, one of the possible pathways to sustainable management and the conservation of natural resources is the possibility of introducing Artificial intelligence (AI)-driven solutions.