Abstract
Abstract. With the development of deep learning, remote sensing image scene classification technology has been greatly improved. However, current deep networks used for scene classification usually introduce ingenious extra modules to fit the characteristics of remote sensing images. It causes a high labor cost and brings more parameters, which makes the network more complicated and poses new intractable problems. In this paper, we rethink this popular “add module” pattern and propose a more lightweight model, called ProbDenseNet (PDN). PDN is obtained via a random search strategy in Neural Architecture Search (NAS) which is an automated network design manner. In our method, all topological connections are assigned importance degrees which subject to a uniform distribution. And we set a regulator to adjust the sparsity of the network. By this way, the design procedure is more automated and the network structure becomes more lightweight. Experimental results on AID benchmark demonstrate that the proposed PDN model can achieve competitive performance even with much fewer parameters. And we also find that excessive connections do not always improve the network’s performance while they can drag down the network’s behavior as well. Furthermore, we conduct experiments on Vaihingen dataset with classical Fully Convolutional Network (FCN) framework. Quantitative and qualitative results both indicate that the features learned by PDN can also transfer in semantic segmentation task.
Highlights
Remote sensing image scene classification is a fundamental work in Computer Vision and in Earth Vision (Cheng, Han, 2016, Xia et al, 2018)
We describe the steps of experiments and analyze corresponding results to confirm the effectiveness of PDN model
4.1.1 AID Dataset We confirm the validity of the proposed PDN model on AID which is widely used in remote sensing image scene classification task
Summary
Remote sensing image scene classification is a fundamental work in Computer Vision and in Earth Vision (Cheng, Han, 2016, Xia et al, 2018). The purpose of remote sensing image scene classification is to efficiently and automatically identify the semantic category label through some algorithms It has a significant impact on Land Use and Land Cover (LULC) determination (Zhang et al, 2013, Zhu et al, 2016), vegetation mapping (Li, Shao, 2013, Mishra, Crews, 2014), urban planning and so on. While it offers a foundation for semantic segmentation (Kampffmeyer et al, 2016), object detection (Wang et al, 2019a, Fu et al, 2020, Feng et al, 2019), Fine Grained Visual Classification (FGVC) (Fu et al, 2019) and other extension tasks. They have a strong interpretability in neurological theory and these architec-
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
