Research on Fill Land Identification Method Integrating Object-Oriented Analysis and CNN
This study aims to explore the integration of object-oriented analysis with deep learning models,particularly Convolutional Neural Networks(CNN),to enhance the accuracy and efficiency of reclaimed land monitoring.Focusing on the Guangdong-Hong Kong-Ma-cao Greater Bay Area as the research region,the geometric and semantic features of reclaimed land targets are utilized.Based on the precise segmentation of remote sensing images using object-oriented techniques,the model further extracts advanced features of land targets and con-ducts classification through multi-layer convolution and pooling operations of the CNN model.The model's classification accuracy is verified by comparing with annotated training samples,using accuracy and recall rate as evaluation criteria.The results show that this method performs excellently in key evaluation metrics such as accuracy,recall rate,and F1 score.It surpasses traditional methods in terms of identification ac-curacy and robustness,providing a new technical approach for rapid identification in reclaimed land monitoring.
land reclamationobject-oriented analysisconvolutional neural networkGreater Bay Area of Guangdong,Hong Kong,and Macau
万方数据
维普
