Road Extraction from Remote Sensing Images Using Parallel Softplus Networks
- PDF / 1,406,212 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 17 Downloads / 235 Views
SHORT NOTE
Road Extraction from Remote Sensing Images Using Parallel Softplus Networks Zhiqiang Li1 Received: 27 May 2019 / Accepted: 25 September 2020 / Published online: 3 October 2020 Ó Indian Society of Remote Sensing 2020
Abstract Road extraction from remote sensing images plays an important role in traffic management, urban planning, automatic vehicle navigation and emergency management. It is a hot issue that how to extract effectively road information from remote sensing images. Here, a new model, namely parallel softplus network (PSNet), has been proposed, which uses parallel network structure and softplus activation function. Specially, the model uses a new weight initialization for extraction effectiveness. Moreover, compared with the popular models, it extracts more complete and continuous road information on the same road remote sensing images. Meanwhile, it outperforms other extraction models, with a high F1score. Experimental results indicate that it is a promising model, which effectively extracts road information from remote sensing images with a little noise. Keywords Parallel networks Weight initialization Softplus activation function Road extraction Remote sensing images
Introduction Remote sensing is a technology that obtains information about objects or phenomena without direct contact with objects, so as to compare with field observations on the earth (Remote sensing 2019). The application of remote sensing technology is very extensive. On the one hand, it can be applied to information data acquisition of different disciplines (e.g., ecology, meteorology, oceanography, geology, etc.). For example, it can monitor climate change and disasters in meteorology. In addition, it can detect marine hydrology, submarine organisms and submarine topographic structure in oceanography. Especially, geographic information of geoscience (e.g., geological structure and topography) is the focus of research. On the other hand, it can also be widely used in industry, agriculture, commerce, military intelligence, environment and so on. For instance, in agriculture, the water used for crops can be monitored by observing the surface temperature, so as to & Zhiqiang Li [email protected] 1
School of Computer Science and Engineering, Yulin Normal University, Yulin 537000, China
judge the health status of crops [Shivers et al. (2019)]; in the environment, the dynamic information of human living environment (including the changes of the ecological environment around the living environment, the changes of water quality, the shortage of resources and the improvement of environmental deterioration) can be obtained by remote sensing technology (Coskun et al. 2008). Furthermore, an important application of remote sensing technology is in the field of transportation, that is, to obtain traffic remote sensing images. Traffic remote sensing images include a large number of traffic data, e.g., road network information, the number and distribution of vehicles, traffic flow status, road condition information (congesti
Data Loading...
