TY - GEN
T1 - Hierarchical reinforcement learning for saliency detection of low-resolution airports
AU - Zhao, Danpei
AU - Ma, Yuanyuan
AU - Wang, Jiajia
AU - Jiang, Zhiguo
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2016/11/1
Y1 - 2016/11/1
N2 - The traditional airport detection methods usually utilize geometric characteristics, which are limited by large amount of data and low-resolution of the remote sensing images. In this paper, we present a novel hierarchical reinforcement learning (HRL) saliency model for quickly airports detecting in large cover area. In contrast with conventional saliency models which usually are effective for high-resolution nature images, our method learns hierarchically high-level features via multi-scale superpixels segmentation and Least Absolute Shrinkage and Selection Operator (LASSO). More importantly, we introduce back-propagation theory for hierarchical learning to adaptively control and generate saliency map. Therefore our unsupervised saliency model is more simple and effective for low-resolution airport detection. Compared with 18 state-of-the-art saliency models, experimental results demonstrate the excellent performance of our method on the remote sensing image datasets. It is more robust and accurate for long-range airports detection.
AB - The traditional airport detection methods usually utilize geometric characteristics, which are limited by large amount of data and low-resolution of the remote sensing images. In this paper, we present a novel hierarchical reinforcement learning (HRL) saliency model for quickly airports detecting in large cover area. In contrast with conventional saliency models which usually are effective for high-resolution nature images, our method learns hierarchically high-level features via multi-scale superpixels segmentation and Least Absolute Shrinkage and Selection Operator (LASSO). More importantly, we introduce back-propagation theory for hierarchical learning to adaptively control and generate saliency map. Therefore our unsupervised saliency model is more simple and effective for low-resolution airport detection. Compared with 18 state-of-the-art saliency models, experimental results demonstrate the excellent performance of our method on the remote sensing image datasets. It is more robust and accurate for long-range airports detection.
KW - airport detection
KW - high-level features
KW - low-resolution
KW - reinforcement learning saliency model
KW - remote sensing image
UR - https://www.scopus.com/pages/publications/85007415326
U2 - 10.1109/IGARSS.2016.7729414
DO - 10.1109/IGARSS.2016.7729414
M3 - 会议稿件
AN - SCOPUS:85007415326
T3 - International Geoscience and Remote Sensing Symposium (IGARSS)
SP - 1622
EP - 1625
BT - 2016 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2016 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 36th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2016
Y2 - 10 July 2016 through 15 July 2016
ER -