Keywords:

Summary

This demo shows how to use the "spherical collector" sensor plugin to emulate a goniometer used to measure bi-directional reflectance distribution (BRDF) functions of surfaces.

polar area 5000

Related Materials

The following demos, manuals and tutorials can provide additional information about the topics at the focus of this demo:

Details

In this scenario, a surface is populated with a distribution of spheres that are illuminated from a single direction and the sensor plugin captures the light scattering into the hemisphere above the complex surface.

Important Files

Scene Description

The candidate scene for this demo is a flat plate that has a collection of 1,300+ uniformly sized and uniformly attributed spheres on it.

The Simulations

There are three simulation scenarios:

overhead.jsim

An overhead imaging scenario of the scene using the OrthoImage plugin.

polar_point.jsim

An overhead hemispherical collection of a point in the center of the scene using the SphericalCollector plugin.

polar_area.jsim

An overhead hemispherical collection of an area in the center of the scene using the SphericalCollector plugin.

Setup

This section includes any step-by-step instructions for running and visualizing the various simulations in this demo.

The Overhead Simulation

To run the overhead simulation, perform the following steps:

  1. Run the DIRSIG overhead.jsim file

  2. Load the resulting overhead.img radiance file in the image viewer.

The Spherical Simulations

There are two spherical collector simulations. One that samples a point (zero area) at the center of the hemisphere and a second that samples an area at the center of the scene.

The Point Collection

  1. Run the DIRSIG polar_point.jsim file

  2. Load the resulting polar_point.img radiance file in the image viewer.

The Area Collection

  1. Run the DIRSIG polar_area.jsim file

  2. Load the resulting polar_area.img radiance file in the image viewer.

Results

Overhead Simulation

The image below is the output of the overhead collection simulation (overhead.jsim) and shows the collection of spheres that compose the scene. The fixed illumination geometry results in illumination of the scene from the direction corresponding to the bottom in this image, with shadows falling toward the top. The extent of this scene is 30 x 30 meters.

overhead
Figure 1. The overhead image of the sphere surface scene.

Outbound Polar Simulation

The image below is from the spherical collector simulation (polar.jsim) using the default radiometric convergence settings. In this case, the collector is looking in at a point in the middle of the scene.

Key parameters in the polar_point.jsim file.
    "mode" : "outbound",
    "coverage" : "hemisphere",
    "projection" : "polar",
    "radius" : 1000,
    "xsamples" : 180,
    "ysamples" : 180

As a result, the collector images the distribution of spheres.

polar point
Figure 2. The outbound spherical collection for a point centered in the scene.

The goal of this demo is to show how the spherical collector can be used as virtual goniometer to measure bi-directional reflectance distribution (BRDF) functions of a complex surface. In this case, that complex surface is this distribution of spheres. To model the effective BRDF of this scene for a 30 meter resolution sensor, we can use the spherical collector and tell it to sample a 30 x 30 meter area (rather than stare at a point) by adding the xwidth and ywidth parameters to the input:

Key parameters in the polar_area.jsim file.
    "mode" : "outbound",
    "coverage" : "hemisphere",
    "projection" : "polar",
    "radius" : 1000,
    "xwidth" : 30.0,
    "ywidth" : 30.0,
    "xsamples" : 180,
    "ysamples" : 180

The image below is output from the spherical collector using the default convergence settings when sampling the 30 x 30 meter scene. The data provides a polar visualization of the radiance reflected by the scene into the hemisphere above the scene. The distribution shows that the radiance reflected by the scene is higher into the 180 degree azimuth direction, which corresponds to the -Y direction in the scene. This is the backscatter direction of the spheres.

polar area default
Figure 3. The outbound spherical collection for a region centered in the scene (default convergence).

The noise in the data is due to the complexity of the scene and variation in illumination and shadows across the 30 x 30 meter area. To drive down the noise, more samples per "pixel" (angular sample in the hemisphere) need to be used. The data below was produced using 5,000 paths/pixel by overriding the default convergence parameters via the command-line --convergence=5000,5000,0 option:

polar area 5000
Figure 4. The outbound spherical collection for a region centered in the scene (5,000 samples per angle).