Abstract. The China-Europe Railway Express (CER-Express) has developed rapidly since its opening, but little attention has been paid to the long-term ecological environment changes along its route, especially in scale of city. This paper studies the long-term changes of cities along the route before and after the opening of CER-Express. Based on the Google Earth Engine platform, we apply the RSEI model for the cities along the route before and after the opening of the CER-Express from 2010 to 2020. Taking Warsaw, Poland as an example, quantitative analysis is carried out regarding its ecological environment changes and the reasons for these changes are discussed. Experimental results show that the mean values of RSEI between 0.6 and 0.8 before the opening, but declined significantly after the opening, and changed steadily in the later stage. This preliminary research can provide theoretical basis for cities along the CER-Express, protect to the sustainable development of CER-Express and response to the call of Goal 15 of the Sustainable Development Goals (SDGs).

Abstract. 3D building model is an important part of 3D GIS. However, with the rapid development of data acquisition technology, the data volume of the 3D building model has increased dramatically. Levels of detail (LOD) technology determines the resource allocation for object rendering based on the position (Screen Size) and importance of the nodes of the model in the display environment, reducing the model’s volume and thus obtaining efficient rendering computations. To ensure the smooth rendering and efficient loading of 3D building models, it is essential to simplify the 3D building models to generate LOD. By taking the texture discontinuity and topological complexity into account, this paper proposes a quadratic error metric mesh simplification algorithm based on "local-vertex" texture features for 3D building models simplification. Using the texture features of the model texture map and the vertex curvature at local vertices, we increase the edge collapse cost of the model in the rich texture areas. After each simplification operation, we use centre of gravity coordinate method to optimize the texture coordinates to preserve the model’s detailed features and topological relationships. A series of experimental comparisons with other state-of-the-art methods verify the effectiveness of the proposed method for simplifying 3D building models. The method in this paper helps to maintain the detailed features and topological relationships of 3D building models while reducing the volume of model and better generating LOD for application in 3D GIS.

Abstract. Urban Historical Buildings (UHBs) preserve the past and memory of a city. Collecting and managing UHB knowledge efficiently is of vital importance to protect local culture heritages and meet the demands of sustainable development. However, UHB knowledge is commonly hidden in the texts from different sources, making it hard to access and manage by general methods efficiently. Meanwhile, domain agnostic knowledge graphs are challenging to express the rich semantic and multi-hop relationships among UHB knowledge. To address the above problems, we proposed a general framework for the extraction and management of the key knowledge from free texts about urban historical buildings. In this study, we attempted to build an UHB knowledge graph from free texts obtained from Internet. Firstly, we designed the UHB domain ontology to regulate the knowledge of urban historical building. Next, the structured knowledge was extracted from web text with Natural Language Processing (NLP) technologies. Finally, we built the UHB knowledge base and developed a Beijing UHB knowledge graph. Furthermore, we demonstrated knowledge retrieval and visualization based on UHB knowledge graph. Our experiment shows that mining UHB knowledge from free text is feasible and important for urban historical building conservation efforts.

Abstract. Accurate and effective extraction of water body information is an important prerequisite for hydrological studies of the Yellow River. However, there is a scattered and frequently swing water flow in the middle and lower reaches of the Yellow River. Traditional water body extraction methods mainly rely on handcrafted statistical features, which cannot fully extract river body in real-world conditions. To deal with these problems and achieve more accurate results, an AU-Net network is proposed to expand the receptive field of the convolutional kernel and incorporate the detailed information of multi-scale features, which improves the ability to extract the middle and lower reaches of the Yellow River from remote sensing images. The experimental results illustrate that compared to the other methods, the AU-NET model has higher recognition accuracy (MPA = 0.97 and MIoU = 0.99) on the water body dataset in the middle and lower reaches of the Yellow River. And the network has high robustness and good fitting, which can better extract the middle and lower reaches of the Yellow River.