Oil Tank Extraction in High-Resolution Remote Sensing Images Based on Deep Learning

The general methods of circular target extraction include Hough transform, circle fitting method, template circle detection method, etc. However, due to the abundance of information in high resolution remote sensing images, the result of the extraction is disturbed by the background, resulting in po...

Full description

Saved in:
Bibliographic Details
Published in2018 26th International Conference on Geoinformatics pp. 1 - 6
Main Authors Xia, Xian, Liang, Hong, RongFeng, Yang, Kun, Yang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2018
Subjects
Accuracy
Caffe
Data mining
Deep learning
Feature extraction
Fuel storage
Oil tank extraction
Oils
Remote sensing
Selective search
Target extraction
Target recognition
Training
Transforms
Online AccessGet full text

Cover

More Information
Summary:The general methods of circular target extraction include Hough transform, circle fitting method, template circle detection method, etc. However, due to the abundance of information in high resolution remote sensing images, the result of the extraction is disturbed by the background, resulting in poor results. In order to solve this problem, this paper proposes an oil tank extraction method in high-resolution remote sensing image based on deep learning. Our experiment uses the RSOD-Dataset shared by Wuhan University. Firstly, it uses the Selective Search algorithm for target recognition, then trains the CaffeNet network model under the deep learning Caffe framework as a feature extraction classifier, and finally marks the oil tank in the image. Experiments show that the method proposed in this paper can effectively carry out oil tank extraction. The proposed method is robust in different complex backgrounds which has high detection rate and low false alarm rate.
ISSN:2161-0258
DOI:10.1109/GEOINFORMATICS.2018.8557161