Abstract
Cloud and snow detection is one of the most significant tasks for remote sensing image processing. However, it is a challenging task to distinguish between clouds and snow in high-resolution multispectral images due to their similar spectral distributions. The shortwave infrared band (SWIR, e.g., Sentinel-2A 1.55–1.75 µm band) is widely applied to the detection of snow and clouds. However, high-resolution multispectral images have a lack of SWIR, and such traditional methods are no longer practical. To solve this problem, a novel convolutional neural network (CNN) to classify cloud and snow on an object level is proposed in this paper. Specifically, a novel CNN structure capable of learning cloud and snow multiscale semantic features from high-resolution multispectral imagery is presented. In order to solve the shortcoming of “salt-and-pepper” in pixel level predictions, we extend a simple linear iterative clustering algorithm for segmenting high-resolution multispectral images and generating superpixels. Results demonstrated that the new proposed method can with better precision separate the cloud and snow in the high-resolution image, and results are more accurate and robust compared to the other methods.
Highlights
High-resolution images are widely used in land cover monitoring, target detection and geographic mapping [1]
Compared with traditional convolutional neural network (CNN)-based pixel methods, experimental results reveal that our CNN architecture can achieve more accurate cloud and snow detection
“salt-and-pepper” noise in clouds and snow detection methods based on the pixel-level, A high-resolution multispectral image contains four bands, and if the SLIC algorithm is used to a new method based on the object level is proposed in clouds and in the snow detection field
Summary
High-resolution images are widely used in land cover monitoring, target detection and geographic mapping [1]. Cloud and snow significantly influence the spectral bands of high-resolution optical images [2]. The previous methods-based spectral is no longer practical due to a lack of SWIR for high-resolution images. Distinguishing between cloud and snow in high-resolution satellite images is a challenging task when using only four optical bands (blue, green, red, and infrared). Many methods for cloud and snow detection are based on pixel level predictions, and tend to produce poor detection results with “salt-and-pepper” noise in high-resolution imagery [12]. Compared with traditional CNN-based pixel methods, experimental results reveal that our CNN architecture can achieve more accurate cloud and snow detection. A new CNN structure is proposed to classify cloud and snow on an object level, which is capable of learning cloud and snow multi-scale semantic features from high-resolution multispectral imagery.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
