Stray light is an imaging artifact in which non-imaging rays reaches the image sensor plane through optical paths other than the intended imaging path. It is commonly observed in optical imaging systems under high-irradiance conditions, including backlighting and direct solar illumination. Stray light may lead to reduced image contrast, veiling glare, and various flare-related artifacts. In digital camera , stray light degrades image quality by obscuring fine details and introducing non-imaging signals that interfere with subsequent image analysis. Consequently, stray light characterization and evaluation are essential components of imaging system performance assessment.

I. What is stray light?
Stray light refers to non-imaging rays that reaches the image plane within an imaging system.

II. Forms of stray light
Typical examples include:
Figure 3: Under strong sunlight, the buildings appear as if covered by a “veil,” with large areas of reduced image contrast due to veiling glare, accompanied by multiple blurred circular artifacts.
Figure 4: Petal-shaped red flares are visible around the sun, a phenomenon known as petal flare.
Figure 5: A series of colored ghost-like artifacts extend below the sun, representing ghost reflections.
Figure 6: Radial streaks emanate from the sun, forming a starburst effect, also called starburst.

III. Sources of Stray Light
Stray light is light that reaches the image plane through unintended optical paths in an imaging system. It results from internal reflections, scattering, and diffraction rather than the designed imaging process.
At a macroscopic level, the main sources include:
1.Reflections
Single or multiple reflections between lens elements, between lenses and mechanical structures (e.g., lens barrel or aperture stop), and between the optics and the image sensor.
2.Scattering
Caused by surface contamination, scratches, or material inhomogeneities that redirect light away from the intended path.
3.Diffraction
Produced by optical structures such as the aperture stop or sensor microstructures, leading to redistribution of light energy.
Typical manifestations include:

• Figure 3 (Veiling glare): Reduced contrast and blurred circular artifacts caused by internal scattering of non-imaging rays.
  

• Figure 4 (Petal flare): Petal-shaped flares around bright light sources due to diffraction and filter reflections.

• Figure 5 (Ghost reflections): Circular artifacts caused by multiple reflections between lens surfaces.

• Figure 6 (Starburst): Radial streaks caused by diffraction at the aperture stop.

IV. Standard Test Methods for Stray Light
Three major standards are commonly used for evaluating stray light performance.
1.ISO 18844:2017 Method
ISO 18844:2017 defines a stray light measurement method for photographic devices.The method uses a uniformly illuminated or uniformly emitting white background containing multiple circular black regions with very low reflectance. These regions are arranged diagonally to form an array of black spots, as shown in Figure 7.

Stray light is quantified using the stray light index $F$, defined as the luminance ratio between the black regions and the white background in the captured image. A lower $F$ value indicates better stray light suppression.

Figure 7. Diagonal array of black spots.

The stray light index is calculated as follows:
$$F = \frac{S_{black}}{S_{white}} \times 100\%$$

Figure 8. ROI for calculation of stray light index F.

2. QC/T 1128-2019 Testing Method
QC/T 1128-2019 defines a stray light test method for automotive cameras.According to QC/T 1128, the device under test (DUT) is operated in a dark environment while imaging a point light source. Stray light suppression by the DUT is evaluated based on the area ratio of flare and ghost artifacts in the captured images, as defined by the standard.
Flare evaluation
The flare spot area shall not exceed 25% of the display area under test conditions.
Test procedure
a) Without the light source, the luminance of the black test scene shall be below $2\text{cd/m}^2$;
b) The light source luminance is $(5～10)$ $\text{Mcd/m}^2$, with an angular size of $(30～34)$ $\text{arcmin}$ ;
c) The light source remains within the DUT field of view while the incident angle is adjusted to maximize flare.
d)Pixels with intensity greater than $50\%$ of the maximum intensity are extracted, and their area ratio is calculated.

Ghost evaluation
Ghost performance is evaluated by the ghost area ratio within the field of view.
The allowable limits of the ghost area ratio depend on the peak luminance ratio between the ghost image and the primary image:

• When the peak luminance ratio is > $50\%$, the ghost area ratio shall be < $1\%$.
• When the peak luminance ratio is > $30\%$ and ≤ $50\%$, the ghost area ratio shall be ≤ $8\%$.
• When the peak luminance ratio is ≤ $30\%$, no requirement is specified for the ghost area ratio.

Test procedure

• A 1 W, 6000 K point light source is used, with a half-intensity angle of and illuminance of $(110 \pm 10)^\circ$ and illuminance of $(220 \pm 22)\ \mathrm{lx}$.
• The DUT operates in Mode B1 in a dark chamber.
• The distance between the lens and the light source is $(400 \pm 10)\ \mathrm{mm}$.
• The DUT is rotated about the geometric center of the lens in horizontal field-of-view steps.
• Rotation is performed clockwise until the light source exceeds $40\%$ of the horizontal field of view.
• The DUT is then reset and rotated counterclockwise under the same conditions.
• All captured images are analyzed to calculate ghost area ratio and luminance ratio.

3.IEEE 2020–2024 Method
IEEE 2020–2024 defines a stray light (flare) evaluation framework.The device under test (DUT) is evaluated under controlled illumination conditions.
The standard defines two test methods: Flare A and Flare B.

Flare A
Flare A is based on the ISO 18844 black spot method. Additional test points are added along the horizontal and vertical directions to extend field-of-view coverage.
Stray light is quantified using the stray light index $F$, consistent with ISO 18844.

Figure 9.IEEE 2020–2024 Flare A Test Chart

Flare B
Flare B evaluates stray light under multi-angle illumination. Either the point light source or the device under test (DUT) is rotated to introduce strong incident light from horizontal, vertical, and diagonal directions. A sequence of images is captured and analyzed to calculate Average Flare Attenuation and Worst Flare Attenuation for quantitative evaluation of camera stray light performance.

Figure 10.Schematic of flare attenuation measurement setups: (a) fixed device under test (DUT) with rotating light source; (b) fixed light source with rotating DUT.

Average Flare Attenuation
The average flare attenuation is defined as the arithmetic mean of the flare attenuation values across all pixels in the captured image.Its formula is:
$$FlareAverage_{dB \theta，\phi}=avg(Flare_{dB \theta，\phi}(x,y))$$ where:

• $\text{FlareAverage}_{\text{dB},\theta,\phi}$: Average flare attenuation, measured in decibels (dB);
• $\theta$ : Azimuth angle of the light source;
• ${\phi}$ :Field-of-view angle of the light source;
• $x,y$:Horizontal and vertical coordinates of the pixel in the image;
• $\text{Flare}_{\text{dB},\theta,\phi}(x,y)$ :Flare attenuation at pixel $(x,y)$, measured in decibels (dB).

Worst Flare Attenuation
Worst Flare Attenuation is defined as the maximum flare attenuation value across all pixels in the captured image. Its formula is:
$$FlareWorst_{dB \theta，\phi}=\max_{x,y}(Flare_{dB\theta,\phi}(x, y))$$ Where:

• $\text{FlareWorst}_{\text{dB},\theta,\phi}$: Worst-case flare attenuation, measured in decibels (dB);
• $\theta$ :Azimuth angle of the light source;
• ${\phi}$:Field-of-view angle of the light source;
• $x,y$:Horizontal and vertical coordinates of the pixel in the image
• $Flare_{dB }$:Flare attenuation at pixel $(x,y)$, measured in decibels (dB)