Study on building extraction with the DC-CTransNet model

Changdong Ji; Zihan Li

doi:10.1117/12.3059051

2 January 2025 Study on building extraction with the DC-CTransNet model

Changdong Ji, Zihan Li

Author Affiliations +

Proceedings Volume 13514, International Conference on Remote Sensing and Digital Earth (RSDE 2024); 135140E (2025) https://doi.org/10.1117/12.3059051
Event: 2024 International Conference on Remote Sensing and Digital Earth, 2024, Chengdu, China

Abstract

To address the issues of missing, misclassified, and distorted building features in the U-Net's single-channel double convolution and skip connection approach, this paper proposes a DC-CTransNet building extraction model. This model expands the receptive field by introducing a DC-Block module with dual-channel triple convolutions and feature concatenation. Additionally, it replaces the traditional skip connection with a CTrans component that includes a Channel Cross-Fusion Transformer (CCT) module and a Channel Cross-Fusion Attention (CCA) block to enhance building detection performance. Experimental results show that DC-CTransNet achieves OA, MIoU, Precision, and FWIoU scores of 98.5%, 97.5%, 95%, and 97.1%, and 94.3%, 96.9%, 96.1%, and 95.4% on the Aerial Image Dataset and Aerial Imagery Labeling datasets, respectively. The building extraction accuracy improved by 1.8%, 2.1%, 4.1%, 3.2%, and 4.1%, 2.8%, 4.2%, 2.0%, respectively. This research provides valuable insights for similar models in building extraction.

(2025) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Changdong Ji and Zihan Li "Study on building extraction with the DC-CTransNet model", Proc. SPIE 13514, International Conference on Remote Sensing and Digital Earth (RSDE 2024), 135140E (2 January 2025); https://doi.org/10.1117/12.3059051

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
10 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Feature extraction

Convolution

Performance modeling

Data modeling

Airborne remote sensing

Education and training

Simulation of CCA and DLA aggregates

Show All Keywords

Keywords/Phrases

Search In:

Publication Years