Acta Photonica Sinica, Volume. 50, Issue 2, 166(2021)
Optimization Algorithm for Polarization Remote Sensing Cloud Detection Based on Machine Learning
Figures & Tables(9)
Fig. 5. Fitting relationship distribution of verification set and test set
Table 1. Feature data set
View tableView in ArticleTable 1. Feature data set
Data dimension Characteristic data name 1 443 nm reflectivity in 2, 5, 8 and 12 directions 2 670 nm reflectivity in 2, 5, 8 and 12 directions 3 490 nm polarized reflectance in 2, 5, 8 and 12 directions 4 670 nm polarized reflectance in 2, 5, 8 and 12 directions 5 865 nm polarized reflectance in 2, 5, 8 and 12 directions 6 R670/R865 reflectivity combination in 2, 5, 8 and 12 directions 7 R763/R765 reflectivity combination in 2, 5, 8 and 12 directions 8 CALIPSO lidar cloud detection label data
Table 2. Quantitative comparison of each pixel algorithm
View tableView in ArticleTable 2. Quantitative comparison of each pixel algorithm
Methods adopted Method type Cloud pixel Clear pixel Undetermined pixel POLDER3 official Threshold method 1147 112 112 MODIS official Threshold method 1429(312) 947(-200) 0 BPNN Machine learning method 1279(162) 1097(-50) 0 PSO-BPNN Machine learning method 1388(271) 988(-159) 0
Tools
Get Citation
Copy Citation Text
Jiejun WANG, Shaohui LIU, Shu LI, Song YE, Xinqiang WANG, Fangyuan WANG. Optimization Algorithm for Polarization Remote Sensing Cloud Detection Based on Machine Learning[J]. Acta Photonica Sinica, 2021, 50(2): 166
Paper Information
Category: Remote Sensing and Sensors
Received: --
Accepted: --
Published Online: Aug. 26, 2021
The Author Email: LI Shu (lishu_owen@sina.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20