A New Remote Sensing Image Retrieval Method Based on CNN and YOLO
Abstract
Retrieving remote sensing images plays a key role in RS fields, which activates researchers to design a highly effective extraction method of image high-level features. However, despite the advanced features of the image can be extracted by deep-learning method, features fail to cover the overall information due to its rich and complex background. Extract key regions from RS images, recede background interference and retrieval accuracy needs to be solved. Combined YOLOv5 target recognition algorithm with deep-learning method, this paper proposes a novel retrieval method based on target critical region detection. Firstly, the key retrieval regions have been identified. YOLOv5 target recognition algorithm has been used to identify the key regions of the image and served as the retrieval regions. Secondly, the retrieval characteristics are determined. Combining with the CNN model ResNet50, the retrieval features are extracted from the retrieval regions acquired in the previous step, in addition, PCA method has been used to reduce the dimension of the retrieval features. Finally, using weighted distance based on class probability to measure the similarity between a query and retrieve images. Experimental results show that the proposed method can extract better image retrieval features and improve the retrieval performance of RS image.
Keywords
RS image retrieval, CNN, PCA, Weighted distance, Key region
Citation Format:
Junwei Xin, Famao Ye, Yuanping Xia, Yan Luo, Xiaoyong Chen, "A New Remote Sensing Image Retrieval Method Based on CNN and YOLO," Journal of Internet Technology, vol. 24, no. 2 , pp. 233-242, Mar. 2023.
Junwei Xin, Famao Ye, Yuanping Xia, Yan Luo, Xiaoyong Chen, "A New Remote Sensing Image Retrieval Method Based on CNN and YOLO," Journal of Internet Technology, vol. 24, no. 2 , pp. 233-242, Mar. 2023.
Full Text:
PDFRefbacks
- There are currently no refbacks.
Published by Executive Committee, Taiwan Academic Network, Ministry of Education, Taipei, Taiwan, R.O.C
JIT Editorial Office, Office of Library and Information Services, National Dong Hwa University
No. 1, Sec. 2, Da Hsueh Rd., Shoufeng, Hualien 974301, Taiwan, R.O.C.
Tel: +886-3-931-7314 E-mail: jit.editorial@gmail.com