Cong zhong Wu¹, Hao Dong¹, Xuan jie Lin¹, Han tong Jiang¹, Li quan Wang¹, Xin zhi Liu¹ and Wei kai Shi², ¹Hefei University of Technology, P.R.China, ²Macau University of Science and Technology, P.R.China

It is difficult to segment small objects and the edge of the object because of larger-scale variation, larger intra-class variance of background and foreground-background imbalance in the remote sensing imagery. In convolutional neural networks, high frequency signals may degenerate into completely different ones after downsampling. We define this phenomenon as aliasing. Meanwhile, although dilated convolution can expand the receptive field of feature map, a much more complex background can cause serious alarms. To alleviate the above problems, we propose an attention-based mechanism adaptive filtered segmentation network. Experimental results on the Deepglobe Road Extraction dataset and Inria Aerial Image Labeling dataset showed that our method can effectively improve the segmentation accuracy. The F1 value on the two data sets reached 82.67% and 85.71% respectively.

Convolutional Neural Network, Remote Sensing Imagery Segmentation, Adaptive Filter, Attention Mechanism, Feature Fusion.