Representing Vector Geographic Information As a Tensor for Deep Learning Based Map Generalisation

Courtial, Azelle; Touya, Guillaume; Zhang, Xiang

doi:10.5194/agile-giss-3-32-2022

Cited by 12 publications

(2 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, when it emerges geographic context, we can set it correctly on Earth and then analyze cell values which can be thought of as color wavelength bands [17]. Raster datasets can be used to store a continuous representation of data, such as weather events and climate or surface elevation forecasts [18]. In this context, the grid representation is more closely related to the representation of physical properties or phenomena found in the world [19].…”

Section: Introductionmentioning

confidence: 99%

The Comparison of Vector and Raster Data for The Calculation of Landscape Environment Using a Geographic Information System Approach

Giofandi¹,

Munibah²,

Kraugusteeliana

et al. 2023

ITJRD

View full text Add to dashboard Cite

GIS-based measurements can combine vector and raster data to produce thematic data obtained from remote sensing data. The data used in this study uses data on land use in the city of Pekanbaru. After the data is obtained, the pixel calculation process is carried out using three methods: the cell center method, the maximum area, and the maximum combined area. This data describes information with multiple raster data resolutions and then interprets the level of distortion in the data. The research findings found that in the process of raster data from 8 different resolution levels for the 5x5 meter category, it is able to provide results that are closest to the area of vector data, where PL 1 code produces 404229 pixels, PL 2 code ranges from 225717 pixels, PL 3 code ranges from 160323 pixels, code PL 4 ranges from 92268 pixels, PL 5 code reaches 73384 pixels, PL 6 code reaches 57237 pixels, and PL 7 code reaches 48315 pixels. Meanwhile, of the 3 methods that were compared to determine distortion with vector data, the cell center approach was the closest to raster data by calculation through eight levels of raster resolution compared to the other two methods. In choosing the right pixel resolution for further use in mathematical modeling, it is necessary to pay attention to the level of resolution by generalizing the resolution of satellite imagery data so that the data can have the same resolution. The weakness of the three methods lies in increasing the resolution the greater it will make the data coarser. This research is expected to be used as a consideration in future research to add a more precise process and be able to produce less storage capacity.

show abstract

Section: Introductionmentioning

confidence: 99%

The Comparison of Vector and Raster Data for The Calculation of Landscape Environment Using a Geographic Information System Approach

Giofandi¹,

Munibah²,

Kraugusteeliana

et al. 2023

ITJRD

View full text Add to dashboard Cite

show abstract

“…orthomosaics, drone or satellite images). A critical goal consists in enabling computer vision models to be trained directly with spatial annotations (Touya et al, 2019, Courtial et al, 2022, as well as delivering model predictions through spatial data formats to automate the production of marine maps from raster data. Moreover, another goal is addressing large raster constraints (whose size exceeds the GPU cache memory) in terms of machine resources for training deep learning models.…”

Section: Introductionmentioning

confidence: 99%

Geoai for Marine Ecosystem Monitoring: A Complete Workflow to Generate Maps From Ai Model Predictions

Talpaert Daudon,

Contini,

Urbina-Barreto

et al. 2023

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Mapping and monitoring marine ecosystems imply several challenges for data collection and processing: water depth, restricted access to locations, instrumentation costs or weather constraints for sampling, among others. Nowadays, Artificial Intelligence (AI) and Geographic Information System (GIS) open source software can be combined in new kinds of workflows, to annotate and predict objects directly on georeferenced raster data (e.g. orthomosaics). Here, we describe and share the code of a generic method to train a deep learning model with spatial annotations and use it to directly generate model predictions as spatial features. This workflow has been tested and validated in three use cases related to marine ecosystem monitoring at different geographic scales: (i) segmentation of corals on orthomosaics made of underwater images to automate coral reef habitats mapping, (ii) detection and classification of fishing vessels on remote sensing satellite imagery to estimate a proxy of fishing effort (iii) segmentation of marine species and habitats on underwater images with a simple geolocation. Models have been successfully trained and the models predictions are displayed with maps in the three use cases.

show abstract

Determining the optimal generalization operators for building footprints using an improved graph neural network model

Niu,

Qian,

Wang

et al. 2024

Geocarto International

View full text Add to dashboard Cite

Representing Vector Geographic Information As a Tensor for Deep Learning Based Map Generalisation

Cited by 12 publications

References 23 publications

The Comparison of Vector and Raster Data for The Calculation of Landscape Environment Using a Geographic Information System Approach

The Comparison of Vector and Raster Data for The Calculation of Landscape Environment Using a Geographic Information System Approach

Geoai for Marine Ecosystem Monitoring: A Complete Workflow to Generate Maps From Ai Model Predictions

Determining the optimal generalization operators for building footprints using an improved graph neural network model

Contact Info

Product

Resources

About