基于MS-UNet的Landsat影像云检测

Cloud Detection of Landsat Image Based on MS-UNet

针对在进行RGB彩色遥感图像云检测研究时,云的形态多变导致对薄云、碎云的检测十分困难的问题,提出一种基于多尺度特征提取的U型网络(MS-UNet)。为了在获得更大感受野的同时保留图像更多的语义信息,提出了一种多尺度模块;在第一组卷积中引入FReLU(Funnel Rectified Linear Unit)激活函数,以获得更多的空间信息。经过下采样进一步进行特征提取,在上采样像素恢复中通过跳跃层将丢失的信息补全,将云的深层语义特征与浅层细节特征结合,以更好地实现对云的分割。实验结果表明,所提方法能有效地分割出薄云与碎云,与UNet、MF-CNN、SegNet、DeepLabV3_ResNet50和DeepLabV3_ResNet101网络相比,所提方法的总体精度分别提升了0.075,0.065,0.070,0.013,0.005。

In order to solve the problem that the detection of thin clouds and broken clouds is very difficult due to the changeable cloud shapes in the research of cloud detection in RGB color remote sensing images,a U-shaped network based on multi-scale feature extraction(MS-UNet)is proposed.Firstly,a multi-scale module is proposed in order to obtain a larger receptive field while retaining more semantic information of the image.Secondly,the FReLU(Funnel Rectified Linear Unit)activation function is introduced in the first group of convolutions to obtain more spatial information.Finally,further feature extraction is performed after down-sampling,and in the up-sampling pixel recovery,the missing information is completed by jump layers,and the deep semantic features of the cloud are combined with the shallow detail features to achieve better cloud segmentation.Experimental results show that this method can effectively segment thin clouds and broken clouds.Compared with UNet,MF-CNN,SegNet,DeepLabV3ResNet50,and DeepLabV3ResNet101 networks,the overall accuracy is increased by 0.075,0.065,0.070,0.013,and 0.005,respectively.