Journal
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
Volume 60, Issue -, Pages -Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2021.3068280
Keywords
Noise measurement; Training; Task analysis; Image resolution; Satellites; Noise robustness; Image segmentation; Deep learning; high-resolution imagery; noisy label; semantic segmentation
Categories
Funding
- National Key Research and Development Plan of China [2017YFA0604500, 2017YFB0202204, 2017YFA0604401, 2020YFB0204700]
- National Natural Science Foundation of China [U1839206]
Ask authors/readers for more resources
This article proposes an online noise correction approach and a synergistic noise correction loss to address the challenges of high-resolution land cover mapping over large areas. By correcting the noisy labels and optimizing the network parameters, the proposed method improves the accuracy of the land cover map from a lower-resolution product in China.
High-resolution land cover mapping over large areas is a challenging task due to the lack of high-quality labels. A potential solution is to leverage the existing knowledge contained in the freely available lower-resolution land cover products. However, the relatively low resolution and low accuracy of the products lead to numerous inaccurate labels, which harms the performance of the neural network. This article addresses the challenge by jointly optimizing the network parameters and correcting the noisy labels with a novel online noise correction approach and a synergistic noise correction loss. By incorporating the information entropy as a measurement to determine the probable correct labels, the proposed noise correction approach learns to make effective correction of the noisy labels during training and eventually boosts the performance with a training set containing less noisy labels. Experimental results show that the proposed method can effectively correct the noisy labels and reduce their negative impact on network training. By employing the proposed method, we produce a refined high-resolution (3-m) land cover map from a lower-resolution (10-m) product in China and improve the accuracy from 74.96x0025; (10-m) to 81.32x0025; (3-m). Such an approach that can effectively learn from noisy data sets leads to many potential opportunities for using and magnifying existing knowledge and results.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available