Road extraction is a unique and difficult problem in the field of semantic segmentation because roads have attributes such as slenderness, long span, complexity, and topological connectivity, etc. Therefore, we propose a novel road extraction network, abbreviated HsgNet, based on high-order spatial information global perception network using bilinear pooling. HsgNet, taking the efficient LinkNet as its basic architecture, embeds a Middle Block between the Encoder and Decoder. The Middle Block learns to preserve global-context semantic information, long-distance spatial information and relationships, and different feature channels’ information and dependencies. It is different from other road segmentation methods which lose spatial information, such as those using dilated convolution and multiscale feature fusion to record local-context semantic information. The Middle Block consists of three important steps: (1) forming a feature resource pool to gather high-order global spatial information; (2) selecting a feature weight distribution, enabling each pixel position to obtain complementary features according to its own needs; and (3) inversely mapping the intermediate output feature encoding to the size of the input image by expanding the number of channels of the intermediate output feature. We compared multiple road extraction methods on two open datasets, SpaceNet and DeepGlobe. The results show that compared to the efficient road extraction model D-LinkNet, our model has fewer parameters and better performance: we achieved higher mean intersection over union (71.1%), and the model parameters were reduced in number by about 1/4.

Road semantic segmentation is unique and difficult. Road extraction from remote sensing imagery often produce fragmented road segments leading to road network disconnection due to the occlusion of trees, buildings, shadows, cloud, etc. In this paper, we propose a novel fusion network (FuNet) with fusion of remote sensing imagery and location data, which plays an important role of location data in road connectivity reasoning. A universal iteration reinforcement (IteR) module is embedded into FuNet to enhance the ability of network learning. We designed the IteR formula to repeatedly integrate original information and prediction information and designed the reinforcement loss function to control the accuracy of road prediction output. Another contribution of this paper is the use of histogram equalization data pre-processing to enhance image contrast and improve the accuracy by nearly 1%. We take the excellent D-LinkNet as the backbone network, designing experiments based on the open dataset. The experiment result shows that our method improves over the compared advanced road extraction methods, which not only increases the accuracy of road extraction, but also improves the road topological connectivity.

For space targets with the same shapes and composite relationships in terms of micromotion forms (STSSCRMFs), the accuracy of deep learning approaches with single-domain data is seriously degraded. To address this issue, this article proposes a hybrid neural network based on feature-level fusion that utilizes multidomain radar information to recognize STSSCRMFs. The proposed network simultaneously processes radar cross-section (RCS) time series, high-resolution range profile (HRRP) sequences, and time-frequency (TF) spectrograms through three branches, leveraging their complementary characteristics. Additionally, an attention mechanism is incorporated into the feature-level fusion module to selectively enhance the features of important domains, thereby achieving improved space target recognition. Consequently, the temporal-spatial features of multidomain data are adaptively fused, capturing discriminative and complementary representa-tions, which effectively enhances the accuracy of space target recognition. Experimental results demonstrate the superiority and effectiveness of the proposed method.

In this paper, an accurate off-line map-matching (OM2) system is designed for complex trajectory networks. It is difficult to complex trajectories input into the hidden Markov model (HMM) directly. OM2 includes three key modules that are pre-processing, map-matching based on weight adaptation HMM (WA-HMM), and post-processing. The pre-processing module divides complex multi-trajectory into singletrajectory sets based on the self-defined trajectory division model (TDM) of crossroads. Another core module is the WA-HMM based on Boxplot, which is used to balance efficiency and accuracy of off-line mapmatching. The post-processing is used to map the points of the crossroads so as to further improve the mapmatching accuracy. OM2 employs the actual GPS trajectories of the internet company and road map of Sichuan province police GIS (PGIS). Our evaluation results show that the accuracy is about 98%, which is suitable for off-line map-matching and solves the problem of complex trajectory network matching being difficult and time-consuming.