Christopher D. Elvidge · Sensors 2010 · Laboratory spectroscopy and sensor simulation study · n=43 lamps

Spectral Identification of Lighting Type and Character

Cited 312 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Laboratory spectroscopy and sensor simulation study (non-clinical design analogy).

OpenAlex W2080916144 · doi:10.3390/s100403961 · record verified 2026-08-27

What was done

High-resolution emission spectra (350 to 2,500 nm) were measured for 43 lamps spanning nine major worldwide lamp types. These narrow-band spectra were used to simulate sensor radiances in eight broad spectral bands: three human eye photoreceptor bands (photopic, scotopic, and "meltopic") and five visible and near-infrared (NIR) bands modeled on the Landsat Thematic Mapper (TM). The study evaluated how accurately these band sets identify lighting types and estimate four lighting character and efficiency metrics: Luminous Efficacy of Radiation (LER), Correlated Color Temperature (CCT), Color Rendering Index (CRI), and Luminous Efficacy (LE).

What was found

Continuous high-resolution spectra performed best for lamp classification and indexing. Among the broad-band configurations, the four-band set modeled on Landsat TM (blue, green, red, and NIR) achieved low errors for lighting type identification and yielded reasonable estimates for LER and CCT. The photopic band was useful for estimating LER, but the three photoreceptor bands performed poorly for lighting type classification. None of the broad-band sets could provide reasonable estimates of LE or CRI. The abstract does not provide specific numerical error rates, correlations, or classification percentages.

Why it matters

It demonstrates the feasibility of identifying lighting types and estimating key efficiency indices from space using multispectral sensors with four broad visible-to-NIR bands (0.4 to 1.0 µm), avoiding the high cost of orbital hyperspectral systems.

Limits

The study is based on laboratory-measured lamp spectra and simulated band passes rather than actual on-orbit or aerial imagery. Real-world confounding factors such as atmospheric absorption/scattering, ground surface reflectance, lamp fixture shielding, and mixed lighting types per pixel were not assessed. No exact numerical metrics or error rates are reported in the abstract.

Cited by