Pano2Geo: An efficient and robust building height estimation model using street-view panoramas

ISPRS Journal of Photogrammetry and Remote Sensing, Год журнала: 2024, Номер 215, С. 177 - 191

Опубликована: Июль 11, 2024

Building height serves as a crucial parameter in characterizing urban vertical structure, which has profound impact on sustainable development. The emergence of street-view data offers the opportunity to observe 3D scenarios from human perspective, benefiting estimation building height. In this paper, we propose an efficient and robust model, call Pano2Geo by precisely projecting panorama (SVP) coordinates geospatial coordinates. Firstly, SVP refinement stratagem is designed, incorporating NENO rules for observation quality assessment four aspects: number buildings, extent nodes, orthogonal observations, followed application art gallery theorem further refine SVPs. Secondly, model constructed, provides pixel-level projection transformation locating features buildings SVP. Finally, valid feature points are extracted based slope mutation test, projected using so obtain proposed was evaluated city Wuhan China, results indicate that can accurately estimate height, with average error 1.85 m. Furthermore, compared three state-of-the-art methods, shows superior performance, only 10.2 % have absolute errors exceeding 2 m, Map-image-based (27.2 %), Corner-based (16.8 Single-view-based (13.9 %) methods. method achieves optimal less than 50 existing SVPs, leading highly estimation, particularly areas high density. Moreover, exhibits robustness maintaining within m even shape complexity occlusion degree increase Our source dataset code available at https://github.com/Giser317/Pano2Geo.git.

Язык: Английский

---

Mapping urban functional zones with remote sensing and geospatial big data: a systematic review

GIScience & Remote Sensing, Год журнала: 2024, Номер 61(1)

Опубликована: Сен. 22, 2024

Язык: Английский

---

An urban building use identification framework based on integrated remote sensing and social sensing data with spatial constraints

Geo-spatial Information Science, Год журнала: 2024, Номер unknown, С. 1 - 25

Опубликована: Авг. 12, 2024

Building use identification is crucial in urban planning and management. Current methods often rely on a single data source neglect spatial proximity. In this paper, we propose stepwise building framework that integrates remote sensing social with constraints based the association of buildings Point Interest (POI), Area (AOI) data. First, study are preprocessed using geometric correction POI AOI reclassification. Then, identify quantitative-density index POIs as well relationships between AOIs buildings. Next, generate Traffic Analysis Zones (TAZs) from road network utilize similarity physical features to within constraints. Finally, kernel density estimation used determine semantic buildings, utilized remaining The specificity our proposed lies not only combination multiple at building-level but also introduction Shenyang selected an example. identifies commercial, residential, industrial, public service scenic spots. accuracy assessment indicates performs well, Overall Accuracy (OA) 87.1% kappa coefficient (kappa) 73.4%. results comparison experiments show consideration integration sources help improve identification. provides new tool for better use, generated suitable in-depth analyses such heat islands.

Язык: Английский

---

Hierarchical partition of urban land-use units by unsupervised graph learning from high-resolution satellite images

International Journal of Digital Earth, Год журнала: 2024, Номер 17(1)

Опубликована: Ноя. 27, 2024

Urban land use information can be effectively extracted from high-resolution satellite images for many urban applications. A significant challenge remains the accurate partition of fine-grained land-use units these images. This paper presents a novel method deriving based on unsupervised graph learning techniques using and open street boundaries. Our constructs to represent spatial relations between cover objects as nodes within block. These are characterized by composition structure features their surrounding neighborhood. We then apply into subgraphs, which communities spatially bounded boundaries correspond units. Next, neural network is used extract deep structural classification. Experiments were conducted cities Fuzhou Quanzhou, China. Results showed that our surpassed traditional grid block techniques, improving classification accuracy 24% 9%, respectively. Furthermore, it achieved results comparable those reference units, with an overall 0.87 versus 0.89.

Язык: Английский

---

Ct-Hiffnet: A Contour-Texture Hierarchical Feature Fusion Network for Cropland Field Parcel Extraction from High-Resolution Remote Sensing Images

Опубликована: Янв. 1, 2024

Язык: Английский

---