Predict Unity Help - 1.7

Sensor

The Sensor section of the UVR Camera Settings component defines the actual sensor behaviour :

The Compute Polarization field of the sensor defines whether the polarization should be computed or not. You should always favour a polarized sensor as it ensures the polarization is taken into account in the render. You should only use a non-polarized sensor if you have memory issues,
The Type defines the sensor behaviour,

Integration spectrum range

If the expert mode is enabled, you can specify the integration spectrum range of all sensors.

By default, sensors of type 'Human', 'Still Camera', 'Luminance Meter' and 'Custom' automatically use the maximum spectrum range of the given observer, spectral response or channels. By selecting a spectrum range of type "Custom", you can reduce this integration range to a custom value.

Preset values are also available for usual integration ranges. Presets involving values in the infra red or ultra violet require the "Infra Red" and/or "Ultra Violet" options to be enabled in the Matter preferences.

Flat Field

In real life, the intensity of the image captured by an optical system is not constant for every pixel of the sensor. Usually, the pixels at the center receive more energy than the outer ones. This results in an image that is brighter at the center which is called the flat field.

If the Expert Mode is enabled, you can specify the Flat Field with an RGB texture.

Figure A : Example of a Flat Field texture

Figure B : rendering of a scene without Flat Field

Figure C : rendering of the same scene as in Figure B, using the Flat Field defined in Figure A

Blooming

When using an actual camera and a pixel gets saturated, its tends to fill the adjacents pixels. This behaviour is called Blooming and results in a luminous blur around saturated pixels.

If the Expert Mode is enabled, you can specify the blooming by :

A texture that defines how a pixel fills its neighbouring pixels : the pixel is represented at the center of the texture, surrounded by its neighbours,
The Intensity defines the impact of the blooming on the final output,
The Threshold defines the saturation level for the pixel : only pixels above this value will bloom.

Figure A : Example of a Blooming texture

Figure B : rendering of a scene without Blooming

Figure C : rendering of the same scene as in Figure B, using the Blooming defined in Figure A

Sensor Type

Predict Engine supports 5 types of sensor :

Human

In order to visualize a scene as the human eye would perceive it, the CIE (Commission Internationale de l'Eclairage) has defined several standard "observers" that represent the color perception of an average person. This sensor converts the raw spectral radiance to the CIE XYZ color space using these observers.

The observers are characterized by 3 curves commonly named the Color Matching Function (CMF). We use this CMF to convert the spectral radiance to the XYZ color space. After an XYZ to RGB conversion, the resulting image corresponds to what an average person would see.

We offer a choice between some commonly used CMF :

CIE 1931 2° standard observer,
CIE 1964 10° standard observer,
CIE 2015 2° standard observer,
CIE 2015 10° standard observer.

Figure A : comparison of the perceived colors of a color checker using a CIE 1931 observer (left), a CIE 1964 observer (middle), and a CIE 2015 observer (right).

Still Camera

This sensor behaves like the CCD or CMOS sensor of a classic camera. The sensor computes a raw image using the camera’s spectral response curves. The raw image is then converted to a LDR image in the post pipeline function. Be aware that the RGB values do not correspond to any physical quantity as it depends on the post pipeline.

This is the sensor you should use to model a measured sensor.

The sensor requires an RGB spectral response file. You can use a default preset or an actual measurement of a sensor. The measurement should be stored in an *.xml file that contains the integration curve of each type of pixel in the sensor. The file should be constructed as follow :

...

</observer>

with the wavelength in nanometers.

Luminancemeter

This sensor converts the raw spectral radiance to the CIE XYZ color space and displays the Y component : the luminance in cd/m².

In order to visualize a scene as the human eye would perceive it, the CIE (Commission Internationale de l'Eclairage) has defined several standard "observers" that represent the color perception of an average person. These observers are characterized by 3 curves commonly named the Color Matching Function (CMF). We use this CMF to convert the spectral radiance to the XYZ color space.

We offer a choice between some commonly used CMF :

CIE 1931 2° standard observer,
CIE 1964 10° standard observer,
CIE 2015 2° standard observer,
CIE 2015 10° standard observer.

Spectroradiometer

This sensor computes the radiance of the light integrated over each spectral band in W/m²/sr. The sensor is defined by its spectrum range (its range of detection) and the number of spectral bands in this range.

Central wavelength

If the expert mode is enabled, you can specify how the spectroradiometer sensor channels are computed.

The "Use central wavelength" toggle defines whether the channels are centered on their definition wavelength (the channel at 380nm goes from 360nm to 400nm) or they start at their definition wavelength (the channel at 380nm goes from 380nm to 420nm).

The "Use central wavelength" toggle is set to true by default.

Polarimeter

This sensor computes the different components of the light’s polarization state, the image contains the four components (S0, S1, S2 and S3) of the Stokes vector. See the polarization section for more details on polarization.

Custom sensor

When the expert mode is enabled, a new custom sensor type is available.

The custom sensor enables you to choose precisely which channels are computed by Predict Engine. A channel is defined by a spectrum -- its integration curve -- that determine how much light is detected by the sensor at each wavelength. For instance, a photographic sensor has three channels (one for each pixel type : R, G or B) defined by the pixel integration curve.

Some preset channels are available to create automatically photographic, photometric or radiometric sensor channels :

the photographic preset creates three RGB channels defined by the given preset spectrums,
the photometric preset creates three XYZ channels defined by the given preset spectrums
the radiometric preset creates n (n = number of bands) constant channels defined regularly on the given spectrum range.

The channels can also be defined manually with a name, a custom spectrum and a unit. The unit is only indicative and will not have any impact on the simulation, it is only used for reference in the output saved file.

A custom channel should never have the same name as a preset channel ("R","G","B" for the photographic channels, "X","Y","Z" for the photometric channels).

Page updated

Report abuse