Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold
Semantic segmentation is an important method to implement fine-grained semantically understand for high-resolution remote sensing images by dividing images into pixel groupings which can then be labelled and classified. In the field of computer vision (CV), the methods based on fully convolutional n...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2019-04-01
|
| Series: | Connection Science |
| Subjects: | |
| Online Access: | http://dx.doi.org/10.1080/09540091.2018.1510902 |
| _version_ | 1797684083072958464 |
|---|---|
| author | Zhihuan Wu Yongming Gao Lei Li Junshi Xue Yuntao Li |
| author_facet | Zhihuan Wu Yongming Gao Lei Li Junshi Xue Yuntao Li |
| author_sort | Zhihuan Wu |
| collection | DOAJ |
| description | Semantic segmentation is an important method to implement fine-grained semantically understand for high-resolution remote sensing images by dividing images into pixel groupings which can then be labelled and classified. In the field of computer vision (CV), the methods based on fully convolutional network (FCN) are the hotspot and have achieved state-of-the-art results. Compared with popular datasets in CV such as PASCAL and COCO, class imbalance is a problem for multiclass semantic segmentation in remote sensing datasets. In this paper, an FCN-based model is proposed to implement pixel-wise classifications for remote sensing image in an end-to-end way, and an adaptive threshold algorithm is proposed to adjust the threshold of Jaccard index in each class. Experiments on DSTL dataset show that the proposed method produces accurate classifications in an end-to-end way. Results show that the adaptive threshold algorithm can increase the score of average Jaccard index from 0.614 to 0.636 and achieve better segmentation results. |
| first_indexed | 2024-03-12T00:24:27Z |
| format | Article |
| id | doaj.art-92c09d1f9c9f45c38520948b8e46c4e7 |
| institution | Directory Open Access Journal |
| issn | 0954-0091 1360-0494 |
| language | English |
| last_indexed | 2024-03-12T00:24:27Z |
| publishDate | 2019-04-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Connection Science |
| spelling | doaj.art-92c09d1f9c9f45c38520948b8e46c4e72023-09-15T10:47:58ZengTaylor & Francis GroupConnection Science0954-00911360-04942019-04-0131216918410.1080/09540091.2018.15109021510902Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive thresholdZhihuan Wu0Yongming Gao1Lei Li2Junshi Xue3Yuntao Li4Space Engineering UniversitySpace Engineering UniversitySpace Engineering UniversitySpace Engineering UniversitySpace Engineering UniversitySemantic segmentation is an important method to implement fine-grained semantically understand for high-resolution remote sensing images by dividing images into pixel groupings which can then be labelled and classified. In the field of computer vision (CV), the methods based on fully convolutional network (FCN) are the hotspot and have achieved state-of-the-art results. Compared with popular datasets in CV such as PASCAL and COCO, class imbalance is a problem for multiclass semantic segmentation in remote sensing datasets. In this paper, an FCN-based model is proposed to implement pixel-wise classifications for remote sensing image in an end-to-end way, and an adaptive threshold algorithm is proposed to adjust the threshold of Jaccard index in each class. Experiments on DSTL dataset show that the proposed method produces accurate classifications in an end-to-end way. Results show that the adaptive threshold algorithm can increase the score of average Jaccard index from 0.614 to 0.636 and achieve better segmentation results.http://dx.doi.org/10.1080/09540091.2018.1510902semantic segmentationremote sensing imagesfully convolutional networkclass imbalanceadaptive threshold |
| spellingShingle | Zhihuan Wu Yongming Gao Lei Li Junshi Xue Yuntao Li Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold Connection Science semantic segmentation remote sensing images fully convolutional network class imbalance adaptive threshold |
| title | Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold |
| title_full | Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold |
| title_fullStr | Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold |
| title_full_unstemmed | Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold |
| title_short | Semantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold |
| title_sort | semantic segmentation of high resolution remote sensing images using fully convolutional network with adaptive threshold |
| topic | semantic segmentation remote sensing images fully convolutional network class imbalance adaptive threshold |
| url | http://dx.doi.org/10.1080/09540091.2018.1510902 |
| work_keys_str_mv | AT zhihuanwu semanticsegmentationofhighresolutionremotesensingimagesusingfullyconvolutionalnetworkwithadaptivethreshold AT yongminggao semanticsegmentationofhighresolutionremotesensingimagesusingfullyconvolutionalnetworkwithadaptivethreshold AT leili semanticsegmentationofhighresolutionremotesensingimagesusingfullyconvolutionalnetworkwithadaptivethreshold AT junshixue semanticsegmentationofhighresolutionremotesensingimagesusingfullyconvolutionalnetworkwithadaptivethreshold AT yuntaoli semanticsegmentationofhighresolutionremotesensingimagesusingfullyconvolutionalnetworkwithadaptivethreshold |