Generating high-accuracy urban distribution map for short-term change monitoring based on convolutional neural network by utilizing SAR imagery

Shota Iino; Riho Ito; Kento Doi; Tomoyuki Imaizumi; Shuhei Hikosaka

doi:10.1117/12.2277901

5 October 2017 Generating high-accuracy urban distribution map for short-term change monitoring based on convolutional neural network by utilizing SAR imagery

Shota Iino, Riho Ito, Kento Doi, Tomoyuki Imaizumi, Shuhei Hikosaka

Author Affiliations +

Proceedings Volume 10428, Earth Resources and Environmental Remote Sensing/GIS Applications VIII; 1042803 (2017) https://doi.org/10.1117/12.2277901
Event: SPIE Remote Sensing, 2017, Warsaw, Poland

Abstract

In the developing countries, urban areas are expanding rapidly. With the rapid developments, a short term monitoring of urban changes is important. A constant observation and creation of urban distribution map of high accuracy and without noise pollution are the key issues for the short term monitoring. SAR satellites are highly suitable for day or night and regardless of atmospheric weather condition observations for this type of study. The current study highlights the methodology of generating high-accuracy urban distribution maps derived from the SAR satellite imagery based on Convolutional Neural Network (CNN), which showed the outstanding results for image classification. Several improvements on SAR polarization combinations and dataset construction were performed for increasing the accuracy. As an additional data, Digital Surface Model (DSM), which are useful to classify land cover, were added to improve the accuracy. From the obtained result, high-accuracy urban distribution map satisfying the quality for short-term monitoring was generated. For the evaluation, urban changes were extracted by taking the difference of urban distribution maps. The change analysis with time series of imageries revealed the locations of urban change areas for short-term. Comparisons with optical satellites were performed for validating the results. Finally, analysis of the urban changes combining X-band, L-band and C-band SAR satellites was attempted to increase the opportunity of acquiring satellite imageries. Further analysis will be conducted as future work of the present study

Citation Download Citation

Shota Iino, Riho Ito, Kento Doi, Tomoyuki Imaizumi, and Shuhei Hikosaka "Generating high-accuracy urban distribution map for short-term change monitoring based on convolutional neural network by utilizing SAR imagery ", Proc. SPIE 10428, Earth Resources and Environmental Remote Sensing/GIS Applications VIII, 1042803 (5 October 2017); https://doi.org/10.1117/12.2277901

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