Average Exterior Visibility Prediction using Polynomial Linear Regression

Visibility inside space is a big design decision driver, but also exterior view quality is an important criterion. Especially in the NEOM Mountain Resort project and it’s museum tower, the “Tower of Discovery” had to be situated with the maximum quality of exterior views. Maximising the quality of exterior views, from the perspective of the tower visitors and inhabitants, also mean visibility from the outside across the NEOM Mountain Resort. Computing the exterior visibility for each facade panel on a large terrain mesh can be expensive, so I computed the visibility only on a few locations of a given grid, across the terrain, to then interpolate in-between, using kernel density estimation. The heatmap of average visibility computed is used as a parameter to predict the average visibility for any given location on the terrain, alongside the vertices x, y and z values.

Focus Area: Museum and Observatory Renderings

Tower of Discovery: Museum lower Mountain Entrance "Grotto"
Tower of Discovery: Museum Interior View Gallery
Tower of Discovery: Museum

Exterior View Analysis: Concept

Position Sample: From one facade panel to all observable points
This is a sample ray-cast, from the tower "Position Option A", of one facade panel to all possible observable points on the terrain, including rays which hit an obstruction in form of building geometry or terrain.
Facade Panel Observation Point Close-up
An offsetted centre point of each facade panel is used as the ray starting point.
Observation Rays With Clear View
Ray-casting showing only obstruction free rays.
Heatmap Of Visibility Analysis: Tower of Discovery Position A
Position A average visibility: 2.6%
Heatmap Of Visibility Analysis: Tower of Discovery Position B
Position B average visibility: 2.9%

Average Exterior View Prediction

To predict the average visibility for any given tower positon, four parameters are used: Mesh-vertices value in x, y and z direction as well as the average visibility heatmap of the initially analised
Parameter: X-Value of Mesh Vertices
Parameter: Y-Value of Mesh Vertices
Parameter: Z-Value of Mesh Vertices
Ground Truth: Average Exterior Visibility
A wide grid of points on the mesh are used to create the ground truth average exterior visibility map and serves as a parameter for the prediction model.
Prediction using Polynomial Regression: Average Exterior Visibility Height-map
This is the result of polynomial regression using the parameters described above and is visualized as a heat and height-map.