Light-Weight Cloud Detection Network for Optical Remote Sensing Images with Attention-Based DeeplabV3+ Architecture

Abstract

Clouds in optical remote sensing images cause spectral information change or loss, that affects image analysis and application. Therefore, cloud detection is of great significance. However, there are some shortcomings in current methods, such as the insufficient extendibility due to using the information of multiple bands, the intense extendibility due to relying on some manually determined thresholds, and the limited accuracy, especially for thin clouds or complex scenes caused by low-level manual features. Combining the above shortcomings and the requirements for efficiency in practical applications, we propose a light-weight deep learning cloud detection network based on DeeplabV3+ architecture and channel attention module (CD-AttDLV3+), only using the most common red–green–blue and near-infrared bands. In the CD-AttDLV3+ architecture, an optimized backbone network-MobileNetV2 is used to reduce the number of parameters and calculations. Atrous spatial pyramid pooling effectively reduces the information loss caused by multiple down-samplings while extracting multi-scale features. CD-AttDLV3+ concatenates more low-level features than DeeplabV3+ to improve the cloud boundary quality. The channel attention module is introduced to strengthen the learning of important channels and improve the training efficiency. Moreover, the loss function is improved to alleviate the imbalance of samples. For the Landsat-8 Biome set, CD-AttDLV3+ achieves the highest accuracy in comparison with other methods, including Fmask, SVM, and SegNet, especially for distinguishing clouds from bright surfaces and detecting light-transmitting thin clouds. It can also perform well on other Landsat-8 and Sentinel-2 images. Experimental results indicate that CD-AttDLV3+ is robust, with a high accuracy and extendibility.

Semantic segmentation of UAV remote sensing images based on edge feature fusing and multi-level upsampling integrated with Deeplabv3+

“…In addition, the balance between the accuracy and efficiency of detection models in largescale remote sensing image segmentation tasks is also a research point of interest. Yao et al [36] combined the channel attention mechanism with a lightweight deep convolutional neural networks (DCNN) to achieve efficient cloud detection on remote sensing images. For the convenience of readers, we summarize the above methods in Table 1.…”

“…As shown in Fig 1, the semantic segmentation module consists of a backbone feature extraction network and an ASPP module. In order to reduce the model computation and memory footprint so that image features can be mined more efficiently and quickly [36] 2, where t is the multiplication factor (i.e., expansion factor) of the input channels, c denotes the number of output channels, n represents the number of repetitions of the module, while s is the step size.…”

A brief analysis of remote sensing image cloud extraction method based on deep learning

“…CD AttDLV3+connects more lowlevel features than DeeplabV3+to improve cloud boundary quality. Introduce channel attention module to strengthen the learning of important channels and improve training efficiency [30] .…”

LCDNet: Light-Weighted Cloud Detection Network for High-Resolution Remote Sensing Images

“…For example, Zhang et al [16] introduced channel attention and spatial attention modules in cloud and snow segmentation in order to enhance the information of key areas, so that key information is highlighted in the channel and spatial dimensions, thereby improving the detection accuracy of thin clouds. Another example is Yao et al [44] designed a lightweight cloud detection network CD-AttDLV3+ based on DeeplabV3+. This network introduces a channel attention module similar to the SE architecture to strengthen the learning of important channels and improve training efficiency.…”