07-10, 16:30–17:00 (Europe/Amsterdam), Method (1.5)
We published a paper to help Geospatial data scientists when dealing with large noisy contours like electoral maps, water bodies or digital contours. We developed EditBoundary.jl, a Julia native semi-manual technique to reduce noise and points while respecting fidelity with 2 new algorithms.
We are excited to apply this method to bring structured quad meshes back into the mainstream - they deserve better!
In some problems of cartography and hydrology the contours of water bodies and coastlines are represented by polygonal curves with a high level of detail and thus require a large number of points. The visualization and processing of these contours may have an enormous computing demand on Geographic Information Systems due to such data scales; therefore shape preserving contours with fewer points are needed to deal with these issues. Our overarching research project, UNAMALLA, relies on EditBoundary.jl to carry out this task.
We developed algorithms in Julia for such polygonal approximation. In our experiments, they handle 90% point elimination adequately, and at interactive speeds, even for highly detailed and noisy polygons.
This work will become part of Meshes.jl and the GeoStats.jl ecosystem for the benefit of all, and is 100% written in Julia.
Julia. Rust. Wannabe time bandit.
Mexican mathematician interested in Julia and Geometry Processing