Date of Publication :2nd September 2024

Abstract:Mangrove species identification is a critical task for the preservation and management of coastal ecosystems. Accurate identification of mangrove species can aid in biodiversity conservation, climate change studies, and coastal zone management. This paper presents a comparative study of three deep learning models—Convolutional Neural Network (CNN), MobileNetV2, and EfficientNet—for the classification of mangrove species. The models were trained and evaluated on a dataset comprising images of various mangrove species. Key performance metrics, including precision, recall, and accuracy, were used to assess and compare the models. The CNN achieved an accuracy of 89.63%, MobileNetV2 achieved 99.80%, and EfficientNet achieved 95.50%. MobileNetV2 performed better than CNN and EfficientNet models in terms of accuracy and robustness, especially in precision and recall metrics for identifying mangrove species. EfficientNet model outperformed CNN but did not match MobileNetV2's performance. MobileNetV2 is identified as the most effective model for this task, showing potential for environmental monitoring and conservation efforts.

Reference :