Road Extraction by Multiscale Deformable Transformer From Remote Sensing Images
- 31 July 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 20 (1545598X) , 1-5
Abstract
Rapid progress has been made in the research of high-resolution remote sensing road extraction tasks in the past years, but due to the diversity of road types and the complexity of road context, extracting the perfect road network is still fraught with difficulties and challenges. Many Convolutional Neural Networks (CNNs) based on encoder-decoder structures have demonstrated their effectiveness. Transformer’s self-attention mechanism shows more powerful performance than CNNs in modeling global feature dependencies. In this paper, we propose a Multi-scale Deformable Transformer Network (MDTNet) based on encoder-decoder structure to extract road networks from remote sensing images. The core of MDTNet is our proposed Multi-scale Deformable Self-Attention (MDSA) mechanism. MDSA can capture more comprehensive features than conventional self-attention. In addition, roads are not present in certain blocks of areas like other objects, but are interwoven throughout the image in such a long, linear fashion that information about certain road segments may be overlooked. To minimize residual errors in road segmentations, our MDSA incorporates a deformable design on feature maps, which effectively enhances the salience of road features relative to their surroundings. Extensive experiments on several public remote sensing road datasets show that our MDTNet achieves higher segmentation [F1 score and Intersection over Union (IoU)] and connectivity [Average Path Length Similarity (APLS)] accuracy, which verifies the effectiveness of our approach.Keywords
Funding Information
- NSFC Key Project of International (Regional) Cooperation and Exchanges (61860206004)
- NSFC Key Project of Joint Fund for Enterprise Innovation and Development (U20B2068)
- National Natural Science Foundation of China (61976004)
This publication has 17 references indexed in Scilit:
- Cascaded Residual Attention Enhanced Road Extraction from Remote Sensing ImagesISPRS International Journal of Geo-Information, 2021
- Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions2021
- DBRANet: Road Extraction by Dual-Branch Encoder and Regional Attention DecoderIEEE Geoscience and Remote Sensing Letters, 2021
- DeepGlobe 2018: A Challenge to Parse the Earth through Satellite Images2018
- D-LinkNet: LinkNet with Pretrained Encoder and Dilated Convolution for High Resolution Satellite Imagery Road Extraction2018
- Pyramid Scene Parsing Network2017
- Automatic Road Detection and Centerline Extraction via Cascaded End-to-End Convolutional Neural NetworkIEEE Transactions on Geoscience and Remote Sensing, 2017
- SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image SegmentationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- Analysis on causes of flash flood in Jeddah city (Kingdom of Saudi Arabia) of 2009 and 2011 using multi-sensor remote sensing data and GISGeomatics, Natural Hazards and Risk, 2015
- GIS-based sustainable city compactness assessment using integration of MCDM, Bayes theorem and RADAR technologyGeocarto International, 2014