Abstract
Building and road extraction from remote sensing images is of great significance to urban planning. At present, most of building and road extraction models adopt deep learning semantic segmentation method. However, the existing semantic segmentation methods did not pay enough attention to the feature information between hidden layers, which led to the neglect of the category of context pixels in pixel classification, resulting in these two problems of large-scale misjudgment of buildings and disconnection of road extraction. In order to solve these problem, this paper proposes a Non-Local Feature Search Network (NFSNet) that can improve the segmentation accuracy of remote sensing images of buildings and roads, and to help achieve accurate urban planning. By strengthening the exploration of hidden layer feature information, it can effectively reduce the large area misclassification of buildings and road disconnection in the process of segmentation. Firstly, a Self-Attention Feature Transfer (SAFT) module is proposed, which searches the importance of hidden layer on channel dimension, it can obtain the correlation between channels. Secondly, the Global Feature Refinement (GFR) module is introduced to integrate the features extracted from the backbone network and SAFT module, it enhances the semantic information of the feature map and obtains more detailed segmentation output. The comparative experiments demonstrate that the proposed method outperforms state-of-the-art methods, and the model complexity is the lowest.
Highlights
As the material carrier of human survival and development, land resources have the characteristics of fixed location, non-renewable, unbalanced distribution of resources and so on [1]
In order to verify the effectiveness of the proposed Non-Local Feature Search Network (NFSNet), experiments were carried out on the open dataset Aerial Image Segmentation Dataset (AISD) [36] and International Society for Photogrammetry and Remote Sensing (ISPRS) 2D Semantic Labeling Contest (ISPRS) [37]
The model proposed in this paper was compared with the current excellent semantic segmentation models Full Convolutional Network (FCN)-8S [12], U-Net [13], DeeplabV3+ [15] and Pyramid Scene Parsing Network (PSPNet) [14]
Summary
As the material carrier of human survival and development, land resources have the characteristics of fixed location, non-renewable, unbalanced distribution of resources and so on [1]. Many scholars had proposed effective feature engineering remote sensing image segmentation methods. Fan et al [4] proposed a remote sensing image segmentation method based on prior information. This method used single point iterative weighted fuzzy c-means clustering algorithm to solve the impact of data distribution and random initialization of clustering center on clustering quality. The above feature engineering segmentation methods could effectively segment remote sensing images They have some problems, such as poor noise resistance, slow segmentation speed and artificial parameter design, and could not competent for the tasks of automatic segmentation of large quantities of data
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have