Chinese rapeseed mapping coupled with flower index and one-class classification
Topics: Remote Sensing
, Agricultural Geography
, China
Keywords: remote sensing, rapeseed mapping, DDYI, one-class classification
Session Type: Virtual Paper Abstract
Day: Saturday
Session Start / End Time: 2/26/2022 11:20 AM (Eastern Time (US & Canada)) - 2/26/2022 12:40 PM (Eastern Time (US & Canada))
Room: Virtual 15
Authors:
Yunze Zang, Beijing Normal University
,
,
,
,
,
,
,
,
,
Abstract
Rapeseed mapping is important for food security and government regulation. Several previous studies have been made to map rapeseed. However, these studies are not suitable for Chinese rapeseed mapping due to cloud cover during key phenology stages, variation of phenology, and intraclass variability. In this study, a rapeseed mapping method coupled with flower index and one-class classification was proposed to overcome these problems. Firstly, a novel index, called Differential-Differential Yellow Index (DDYI) was proposed firstly to select rapeseed samples. Second, phenology-based synthesis features were extracted, and sample enhancements were employed to improve the generalization of the model. Finally, a binary random forest classifier using a positive unlabeled learning strategy (PUL-RF) was applied to classify rapeseed. With the proposed method, China rapeseed was mapped at 20m scale for 2019-2021 unprecedently. Validation results demonstrate that our method achieves an overall accuracy of 95.08% and an F1-Score of 0.937. Compared with census rapeseed area data, the area extracted by the method was well correlated with the census area. In contrast, the performance of other methods was poorer than our method. These results indicate that the method is effective and robust for Chinese rapeseed mapping. More significantly, we filled the gap of rapeseed maps in China.
Chinese rapeseed mapping coupled with flower index and one-class classification
Category
Virtual Paper Abstract
Description
This abstract is part of a session. Click here to view the session.
| Slides