SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Photonics West 2019 | Call for Papers

2018 SPIE Optics + Photonics | Register Today



Print PageEmail PageView PDF

Sensing & Measurement

Efficient measurement of large light sources for optical simulation and design

A new method of generating illumination distributions for modeling lighting systems reduces space requirements and improves accuracy.
7 July 2009, SPIE Newsroom. DOI: 10.1117/2.1200907.1696

Like any task in optical design, large lighting projects such as architectural, automotive, street, security, or entertainment systems still need to be characterized for output color and luminance. However, these distributions are difficult to measure because of both the size of the source and the large space required for the measurements. For such lighting systems, the illumination pattern is so large that measurement logistics in controlled laboratory environments become difficult and expensive.

This problem is solved by looking at it from a completely different perspective. Instead of measuring the illumination distribution and having to control the light source, the extended measurement environment, the target surface, and the measurement system, we just measure the light-source parameters directly. This allows measurement from within the near field. The data gathered is then extrapolated to generate the illumination distribution at any distance from the light source. This technique has been demonstrated to work successfully for sources such as down lights, lighting tubes, and automobile headlamps.

The fundamental concept is to combine a series of spatial measurements into a luminance and color model that can then be extrapolated to any distance. The general measurement setup consists of a high-dynamic-range imaging colorimeter for capturing luminance and color data, a two-axis goniometric system to allow tilting and rotating of the light source in the colorimeter's field of view, and a software system that controls an automated measurement sequence (see Figure 1).1 Our tests of the method were implemented using a Radiant Imaging ProMetric® imaging colorimeter and the ProMetric near-field goniometric measurement system (PM-NFMS™ ) and software.

Figure 1. Setup for measuring light-source luminance and color distribution. The setup includes a high-dynamic-range imaging colorimeter for capturing luminance and color data and a two-axis goniometric system to allow tilting and rotating of the light source.

Each individual colorimeter measurement captures both the luminance and color profile of the source from a particular angle. The goniometric system moves the light source through a series of inclination angles relative to the fixed position of the imaging colorimeter, tilting and rotating the source until a complete network of measurements is taken. The multiple images capture the spatial distribution of the light output, which is modeled as an extended source.

Figure 2. Images generated from automobile headlamp-measurements using the near-field method. The illumination distributions have been extrapolated to theoretical planes at the colorimeter location and at a distance of 3.4m from the source. Traditionally, automobile headlamps are measured from about 100ft.

Typically the source is tilted through a ±90° range and then rotated by ±90° from left to right for each tilt angle. The measurement steps can be as small as 0.1°, although for practical applications very good results are obtained with steps as large as 2.5–5°. One can also also have different step sizes for different parts of the measurement. Testing time typically ranges from 0.5 to several hours, depending on the source and measurement resolution desired.

Once the complete set of measurements is obtained, they are combined to form the near-field performance model, which fully describes the light source's luminance and color output. The data is recorded as output light intensities. The latter are expressed as a function of the location on the source and the direction the rays travel in. This intensity function can also include color coordinates if color measurements are made. The intensity distribution can then be used for optical modeling or to generate a far-field distribution.

For optical modeling, the intensity function is used to generate a ray set for importing into optical modeling software. The ray set is generated by randomly selecting the light-ray travel direction and an originating pixel on the light source. This selection is done either with Monte Carlo methods or by weighting the direction and origin based on the measured luminance distribution. This second approach, ‘importance sampling,’ usually results in improved accuracy when generating the same number of rays. To generate the ray, the measured intensity function is used to extrapolate an intensity value with the specific origin and direction, and the origin is the point along that ray where it intersects a user-defined optical surface. Usually, hundreds of thousands to tens of millions of rays are generated to produce a ray set.

To generate a far-field distribution, this same methodology is followed except that the intensity distribution can be simplified by setting the location coordinates to zero for all measured data. This has the effect of treating the light source as a point source and so yields the far-field luminous-intensity distribution.

Traditionally the output of automobile headlamps is measured from about 100ft (30m). This requires a specialized (and expensive) testing facility. With the near-field measurement method the source can be measured in the space of just a few meters (see Figure 2). Comparisons of illumination distributions extrapolated using our near-field method with those obtained by direct measurement have shown an excellent match.

The imaging-colorimetry approach is broadly applicable to general lighting systems such as lighting tubes (see Figure 3). The method allows measurement in a compact laboratory space and is automated, making it easy to use. The measurements provide full near-field data for optical design and simulation, and color distribution by angle, and capture detailed images of the lighting system in multiple configurations useful for development and design assessment. We are currently working to adapt the system for a broader variety of light sources that have complex mounting requirements and to define the optimal measurement step sizes for these sources. We have shown that the measurements obtained with this system have excellent correlation with other standardized illumination-distribution measurement methods, in particular those for automobile headlamps. We hope to establish our approach as equivalent for the purposes of light-source certification.

Figure 3. Architectural lighting system mounted for measurement.
This work was presented in the conference Novel Optical Systems Design & Optimization at the SPIE Optics + Photonics symposium in August 2009 in San Diego.

Hubert Kostal, Douglas Kreysar, Ronald Rykowski
Radiant Imaging Inc.
Redmond, WA

Hubert Kostal has been vice president of sales and marketing since 2007. He received a PhD in mathematics from the University of Texas at Austin, has published numerous papers, and holds several patents.

Douglas Kreysar is chief operating officer. He received a BS in physics from Vanderbilt University in 1991, and an MS in applied physics from the University of Michigan, Ann Arbor, in 1993. He has authored multiple technical papers and holds three US patents.

Ronald Rykowski, president and co-founder of Radiant Imaging, has over 25 years of experience in software development, illumination, optical design, and photometry and colorimetry. He received a BS in physics and chemistry with a minor in computer science from the University of California, Irvine, in 1980.