A Hybrid CNN-LSTM Model for Temporal-Spatial Analysis of Pine Wilt Disease Severity

Pine Wilt Disease poses a great threat to coniferous forests, accurate and early detection is required for effective disease control. This study provides a powerful model with CNNs and LSTMs that can effectively categorize PWD into five distinct levels of severity. Its overall accuracy is 97.58% The plant species share similar features in their structural design and appearance, enabling the scientists to accurately identify them from each other based on these distinct characteristics of all three statues at a greater-than-average level percentage-wise. Its ability to distinguish different stages of disease severity is impressive, and the key performance parameters that it has shown – precision, recall but also F1-score are all beyond acceptable values. Comparative analysis against existing methodologies reveals superiority as benchmarks are well above current state-of-the-art systems. This breakthrough has the potential to revolutionize forest disease monitoring and intervention strategies. However, due to the restricted variability of datasets and environmental factors, there is a need for further research so that their results can also apply more generally. This research thus signifies a milestone toward precise definitions of the level of severity in PWD, and such innovations have had revolutionary impacts on forest health management across the globe.