Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Mar 27;14(1):7313.
doi: 10.1038/s41598-024-57408-0.

A deep inverse convolutional neural network-based semantic classification method for land cover remote sensing images

Ming Wang  1 , Anqi She  2 , Hao Chang  3 , Feifei Cheng  4 , Heming Yang  5
Affiliations

A deep inverse convolutional neural network-based semantic classification method for land cover remote sensing images

Ming Wang et al. Sci Rep. .

Abstract

The imbalance of land cover categories is a common problem. Some categories appear less frequently in the image, while others may occupy the vast majority of the proportion. This imbalance can lead the classifier to tend to predict categories with higher frequency of occurrence, while the recognition effect on minority categories is poor. In view of the difficulty of land cover remote sensing image multi-target semantic classification, a semantic classification method of land cover remote sensing image based on depth deconvolution neural network is proposed. In this method, the land cover remote sensing image semantic segmentation algorithm based on depth deconvolution neural network is used to segment the land cover remote sensing image with multi-target semantic segmentation; Four semantic features of color, texture, shape and size in land cover remote sensing image are extracted by using the semantic feature extraction method of remote sensing image based on improved sequential clustering algorithm; The classification and recognition method of remote sensing image semantic features based on random forest algorithm is adopted to classify and identify four semantic feature types of land cover remote sensing image, and realize the semantic classification of land cover remote sensing image. The experimental results show that after this method classifies the multi-target semantic types of land cover remote sensing images, the average values of Dice similarity coefficient and Hausdorff distance are 0.9877 and 0.9911 respectively, which can accurately classify the multi-target semantic types of land cover remote sensing images.

Keywords: Deep inverse convolutional neural network; Feature extraction; Land cover; Remote sensing images; Semantic classification; Semantic segmentation.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Schematic diagram of supervised learning stage.
Figure 2
Figure 2
Land cover remote sensing image segmentation algorithm based on deep deconvolution neural network.
Figure 3
Figure 3
Upsampling structure.
Figure 4
Figure 4
Division of object positions in images.
Figure 5
Figure 5
Training loss value of deep deconvolution neural network.
Figure 6
Figure 6
Training accuracy value of deep deconvolution neural network.
Figure 7
Figure 7
Original land remote sensing images.
Figure 8
Figure 8
The semantic segmentation effect of land remote sensing images using the method described in this paper.
Figure 9
Figure 9
Dice similarity coefficient calculation results.
Figure 10
Figure 10
Hausdorff distance calculation result.
Figure 11
Figure 11
Test results of image segmentation performance using different methods.
See this image and copyright information in PMC

References

    1. Qinghui L, Michael K, Robert J, Salberg AB. Multi-modal land cover mapping of remote sensing images using pyramid attention and gated fusion networks. Int. J. Remote Sens. 2022;43(9/10):386–412.
    1. Demirkan DC, Koz A, Duzguna HS. Hierarchical classification of Sentinel 2-a images for land use and land cover mapping and its use for the CORINE system. J. Appl. Remote Sens. 2020;14(2):026524. doi: 10.1117/1.JRS.14.026524. - DOI
    1. Coronado A, Moctezuma D. Feature evaluation for land use and land cover classification based on statistical, textural, and shape features over Landsat and Sentinel imagery. J. Appl. Remote Sens. 2020;14(4):048503. doi: 10.1117/1.JRS.14.048503. - DOI
    1. Mohammad R, Stuart R, Phinn CM, Roelfsema A, Abdul A. Modeling forest cover dynamics in Bangladesh using multilayer perceptron neural network with Markov chain. J. Appl. Remote Sens. 2022;16(3):034502.
    1. Jorgen AA, Luigi T, Luppino SNA, Jane UJ. Toward targeted change detection with heterogeneous remote sensing images for forest mortality mapping. Can. J. Remote Sens. 2022;48(6):826–848. doi: 10.1080/07038992.2022.2135497. - DOI