Lightweight U-Net for cloud detection of visible and thermal infrared remote sensing images
- PDF / 2,272,077 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 25 Downloads / 269 Views
Lightweight U‑Net for cloud detection of visible and thermal infrared remote sensing images Jiaqiang Zhang1,3,4 · Xiaoyan Li1,2,3,4 · Liyuan Li1,3,4 · Pengcheng Sun3,5 · Xiaofeng Su1,3 · Tingliang Hu1,3 · Fansheng Chen1,2,3 Received: 17 May 2020 / Accepted: 31 July 2020 © The Author(s) 2020
Abstract Accurate and rapid cloud detection is exceedingly significant for improving the downlink efficiency of on-orbit data, especially for the microsatellites with limited power and computational ability. However, the inference speed and large model limit the potential of onorbit implementation of deep-learning-based cloud detection method. In view of the above problems, this paper proposes a lightweight network based on depthwise separable convolutions to reduce the size of model and computational cost of pixel-wise cloud detection methods. The network achieves lightweight end-to-end cloud detection through extracting feature maps from the images to generate the mask with the obtained maps. For the visible and thermal infrared bands of the Landsat 8 cloud cover assessment validation dataset, the experimental results show that the pixel accuracy of the proposed method for cloud detection is higher than 90%, the inference speed is about 5 times faster than that of U-Net, and the model parameters and floating-point operations are reduced to 12.4% and 12.8% of U-Net, respectively. Keywords Fully convolutional network · Depthwise separable convolution · Cloud detection · Semantic segmentation · Lightweight network
1 Introduction According to the results of the International Satellite Cloud Climatology Project (ISCCP), clouds cover two-thirds of the land surface on earth (Rossow and Schiffer 1991), and high cloud coverage will reduce the accuracy and application of remote sensing data, resulting
Jiaqiang Zhang and Xiaoyan Li have contributed equally to this work. * Xiaoyan Li [email protected] * Xiaofeng Su [email protected] * Fansheng Chen [email protected] Extended author information available on the last page of the article
13
Vol.:(0123456789)
397
Page 2 of 14
J. Zhang et al.
in changes in the texture and spectral information of remote sensing images and affecting the radiation correction, geometric calibration and distortion correction (Li et al. 2019). Since the time window for satellite data downlink transmission is only 5 to 10 min, cloud detection preprocessing of the data before transmission can improve the quality and efficiency of the remote sensing images (Williams et al. 2002). With the continuous development of remote sensing technology, miniaturization of satellites has limited the computing power and power consumption of electronic systems. Therefore, a lightweight cloud detection model with low computing power requirements is necessary. Spectral analysis method is widely used in cloud detection. Radiance and reflectance of cloud in visible, short-wave infrared, thermal infrared and other wave bands are used for cloud detection through threshold method. Irish et al. (2006) proposed a cl
Data Loading...
