Road Network Extraction Using Atrous Spatial Pyramid Pooling
Aravapalli Sri Chaitanya1, Suvarna Vani Koneru2, Praveen Kumar Kollu3
1Aravapalli Sri Chaitanya, Computer Science and Engineering, V R Siddhartha Engineering College, Vijayawada, India.
2Suvarna Vani Koneru, Computer Science and Engineering, V R Siddhartha Engineering College, Vijayawada, India.
3Praveen Kumar Kollu, s Computer Science and Engineering, V R Siddhartha Engineering College, Vijayawada, India.
Manuscript received on 30 June 2019 | Revised Manuscript received on 05 July 2019 | Manuscript published on 30 July 2019 | PP: 31-33 | Volume-8 Issue-9, July 2019 | Retrieval Number: H7459068819/19©BEIESP | DOI: 10.35940/ijitee.H74590.78919
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Road extraction from satellite images has several Applications such as geographic information system (GIS). Having an accurate and up-to-date road network database will facilitate transportation, disaster management and GPS navigation. Most active field of research for automatic extraction of road network involves semantic segmentation using convolutional neural network (CNN). Although they can produce accurate results, typically the models give up performance for accuracy and vice-versa. In this paper, we are proposing architecture for semantic segmentation of road networks using Atrous Spatial Pyramid Pooling (ASPP). The network contains residual blocks for extracting low level features. Atrous convolutions with different dilation rates are taken and spatial pyramid pooling is performed on these features for extracting the spatial information. The low level features from residual blocks are added to the multi scale context information to produce the final segmentation image. Our proposed model significantly reduces the number of parameters that are required to train the model. The proposed model was trained on the Massachusetts roads dataset and the results have shown that our model produces superior results than that of popular state-of-the art models.
Keywords: Convolutional neural network, Geographical information systems, Road network, Semantic segmentation.
Scope of the Article: Advanced Computer Networking