| Full text | |
| Author(s): |
Wagner, Fabien H.
[1, 2]
;
Dalagnol, Ricardo
[1]
;
Tarabalka, Yuliya
[3, 4]
;
Segantine, Tassiana Y. F.
[2]
;
Thome, Rogerio
[2]
;
Hirye, Mayumi C. M.
[5]
Total Authors: 6
|
| Affiliation: | [1] Natl Inst Space Res INPE, Remote Sensing Div, BR-12227010 Sao Jose Dos Campos, SP - Brazil
[2] Fdn Sci Technol & Space Applicat FUNCATE, GeoProc Div, BR-12210131 Sao Jose Dos Campos, SP - Brazil
[3] INRIA Sophia Antipolis, F-06902 Sophia Antipolis - France
[4] Luxcarta Technol, Parc Activite Argile, Lot 119b, F-06370 Mouans Sartoux - France
[5] Univ Sao Paulo, Fac Architecture & Urbanism, Quapa Lab, BR-05508900 Sao Paulo, SP - Brazil
Total Affiliations: 5
|
| Document type: | Journal article |
| Source: | REMOTE SENSING; v. 12, n. 10 MAY 2020. |
| Web of Science Citations: | 1 |
| Abstract | |
Currently, there exists a growing demand for individual building mapping in regions of rapid urban growth in less-developed countries. Most existing methods can segment buildings but cannot discriminate adjacent buildings. Here, we present a new convolutional neural network architecture (CNN) called U-net-id that performs building instance segmentation. The proposed network is trained with WorldView-3 satellite RGB images (0.3 m) and three different labeled masks. The first is the building mask; the second is the border mask, which is the border of the building segment with 4 pixels added outside and 3 pixels inside; and the third is the inner segment mask, which is the segment of the building diminished by 2 pixels. The architecture consists of three parallel paths, one for each mask, all starting with a U-net model. To accurately capture the overlap between the masks, all activation layers of the U-nets are copied and concatenated on each path and sent to two additional convolutional layers before the output activation layers. The method was tested with a dataset of 7563 manually delineated individual buildings of the city of Joanopolis-SP, Brazil. On this dataset, the semantic segmentation showed an overall accuracy of 97.67% and an F1-Score of 0.937 and the building individual instance segmentation showed good performance with a mean intersection over union (IoU) of 0.582 (median IoU = 0.694). (AU) | |
| FAPESP's process: | 16/17652-9 - Functional diversity of intact and regenerating Amazon, Atlantic Forest and Cerrado systems using hyperspectral imagery |
| Grantee: | Fabien Hubert Wagner |
| Support Opportunities: | Scholarships in Brazil - Young Researchers |
| FAPESP's process: | 15/50484-0 - Functional diversity of intact and regenerating Amazon, Atlantic Forest, and Cerrado systems using hyperspectral imagery |
| Grantee: | Fabien Hubert Wagner |
| Support Opportunities: | Research Grants - Young Investigators Grants |
| FAPESP's process: | 15/22987-7 - Assessment of climate change impacts on the biomass and carbon dynamics in the Amazon |
| Grantee: | Ricardo Dal'Agnol da Silva |
| Support Opportunities: | Scholarships in Brazil - Doctorate |