Search

Objective: Deep convolutional neural network (DCNN) is widely used in automatic road extraction from high-resolution remote sensing images (HRSIs). However, the existing methods make it difficult to model the context relationship of pixels in the predicted results. To solve this problem, some studies have used fully connected conditional random field (FullCRF) to perform secondary optimization of semantic segmentation results combined with context information, but the discontinuity problem of road structure cannot be effectively improved. Methods: In order to improve the integrity of road structure, this paper proposes a short range conditional random filed (SRCRF) model combined with DCNN. SRCRF mainly includes a unary potential function based on road pre-segmentation, a binary potential function based on spectral spatial features, and a K-neighborhood mean field inference algorithm. First, the prior knowledge of road pre-segmentation results is obtained by using the powerful feature extraction capability of DCNN as the unary potential function of SRCRF. Then, the dependence of the binary potential function defined by the linear combination of Gaussian kernel functions on the surrounding nodes is modeled. The binary potential function enables the classification results to have local consistency,that is, adjacent pixels with similar spectral features have the same label. Finally, the K-neighborhood mean field inference algorithm based on the mean-field approximation inference algorithm optimizes the inference range to make full use of the spatial context information and spectral feature context information of the road, and then calculates the optimal label corresponding to each pixel based on the space and spectral feature to optimize the road accurately. The convolution method is adopted to control the inference range of SRCRF within the radius of K in order to improve the proportion of feature vectors. Results: The experimental results show that SRCRF not only alleviates the transition smoothness of FullCRF, but also alleviates the structural discontinuity and incompleteness in the road acquisition results of HRSIs. In Zimbawe-Roads dataset and Cheng-Roads dataset, F1-scores of SRCRF are increased by about 4.01% and 3.73% respectively compared with DCNN, and about 3.25% and 2.28% respectively compared with FullCRF. Conclusions: The proposed SRCRF combines the advantages of DCNN, optimizes the fully connected structure of traditional conditional random fields into the K-neighborhood structure, reduces the inference scope, and improves the proportion of feature vectors. Compared with FullCRF, SRCRF can make better use of image color features and spatial features to accurately optimize the road extraction results of deep learning output. The performance of SRCRF is improved compared with DCNN and FullCRF, and the time is shortened by one order of magnitude compared with FullCRF. In the future, we will further investigate the potential for learning Gaussian features and investigate more complex architecture of conditional random field to better capture global context information. Additionally, we are particularly interested in exploring the application potential of SRCRF in other fields, such as building extraction, vehicle extraction, lake extraction, etc. [ABSTRACT FROM AUTHOR]

Objectives: In the field of human-related geographic information systems (GIS), the spatiotemporal analysis of character information has received increasingly more attention. It is important in that it helps GIS users to generate various thematic maps and achieve the visualization of human geographic con⁃ tent. For adaptation to the development direction of GIS intellectualization, it is of great significance to combine GIS requirements with natural language processing (NLP) methods and build a character information model. Methods: Firstly, we expound the research status of character information models in GIS and NLP and put forward the concept of life-track, which is mainly composed of a series of character event mentions. Secondly, considering the feasibility of the existing information extraction methods, a conceptual character life-track information model is determined. This model focuses on event information to highlight character spatiotemporal elements and also includes character attribute and relationship information. Finally, a complete information extraction process is designed for the model with online character encyclopedia pages as the data source. This paper focuses on two sub-tasks in the process: One is to use time features and Open How Net semantic calculations to identify event mentions, and the other is to use linguistics features and the conditional random field (CRF) model to extract location information. Results: Experiment results show that the method of event mention identification has an accuracy of 91.8%. Although the average F1 value of location information extraction is only 78% under the condition of a limited labeling corpus, some valuable experimental conclusions have been obtained by analyzing the weight of the transmit matrix of the CRF model: (1) The location phrase and its adjacent words have obvious characteristic effects. (2) The dependency syntactic parsing and the relative position of the word in the sentence have little influence on the extraction. (3) The target of location information extraction is the place where the event occurred, but in a few cases, some location phrases are irrelevant to the location of the event. This is the main reason for the low accuracy. Conclusions: Combining GIS with NLP, intelligent GIS development will be promising. The character life-track information model provides an example of the large-scale use of ubiquitous internet information. Improving methods applied in the extraction process and applying those methods to more online text types are the focus of our team's subsequent research. [ABSTRACT FROM AUTHOR]