基于纹理特征的深度学习云和云阴影检测
Deep learning based on texture features for cloud and cloud shadow detection
张昊 1焦瑞莉 1乔聪聪 2霍娟 2宗雪梅2
作者信息
- 1. 北京信息科技大学信息与通信工程学院,北京 100101
- 2. 中国科学院大气物理研究所中层大气和全球环境探测重点实验室,北京 100029
- 折叠
摘要
针对云和云阴影检测过程中存在边界不准确以及易与地表混淆等问题,构建一种融合纹理特征模块的卷积神经网络模型对Landsat 8遥感图像进行云和云阴影检测.引入基于统计特性的纹理特征模块进行纹理特征的提取和学习,在训练过程采用焦点损失函数削弱样本不均衡带来的影响.实验结果表明,该模型细化了云和云阴影的边界等纹理细节,减少了云和云阴影的误检和漏检现象,提高了云和云阴影的检测精度.
Abstract
Aiming at the problems of inaccurate boundary and easiness to be confused with the surface in the process of cloud and cloud shadow detection,a convolutional neural network model incorporating texture feature module was constructed to detect clouds and cloud shadows in Landsat 8 remote sensing images.A texture feature module based on statistical properties was intro-duced for texture feature extraction and learning,and focal loss function was adopted in the training process to weaken the inf-luence of sample imbalance.Experimental results show that the proposed model refines the texture details such as boundaries of clouds and cloud shadows,reduces the false detection and missed detection of clouds and cloud shadows,and improves the detec-tion accuracy of clouds and cloud shadows.
关键词
云检测/云阴影检测/统计特性/纹理特征/卷积神经网络/遥感图像/焦点损失函数Key words
cloud detection/cloud shadow detection/statistical properties/texture feature/convolutional neural network/re-mote sensing images/focal loss function引用本文复制引用
基金项目
国家自然科学基金重点项目(42030107)
国家自然科学基金面上项目(42175150)
出版年
2024
NETL
NSTL
维普
万方数据