Low-light cameras: Employ specialized imaging devices such as image intensifiers, EMCCD, ICCD/ICMOS, which amplify weak photons by multiple orders of magnitude through physical-level photoelectric conversion. They can operate at illuminance below 10⁻⁴ Lux and directly output enhanced visible-light images.

Ultra-low-light cameras: Based on high-sensitivity CCD/CMOS sensors, imaging is optimized via large-aperture lenses, low-noise circuits, and software algorithms. Their operating illuminance typically ranges from 10⁻² to 10⁻³ Lux (moonlight level to starlight level), with the limit rarely exceeding 10⁻⁴ Lux.

Minimum Illuminance Levels for Low-Light Cameras

0.1 Lux: Standard Level

0.01 ~ 0.1 Lux: Moonlight Level

0.001 ~ 0.01 Lux: Starlight LevelBelow

 0.001 Lux: Ultra Starlight Level

0.0005 Lux: Black Light Level

Three Pillars of Low-Light Imaging

1. ISP Image Processor: The Core of Image Quality Optimization

Functions: Performs post-processing on signals output by front-end image sensors, including linear correction, noise reduction, dead pixel

removal, interpolation

white balance, and automatic exposure control. Reliance on ISP technology enables satisfactory restoration of on-site details under various

optical conditions, and ISP technology

largely determines the imaging quality of cameras. Parameter tuning of the ISP involves processing such as 3D noise reduction, color

noise suppression, edge outlining, wide dynamic range, gain control, noise filtering, and edge enhancement. 2. Hardware Collaborative Optimization: Enhancing Physical Light Sensitivity Optical System: Large-aperture lens: The light intake of an F1.0 aperture is four times that of an F2.0 aperture (a larger aperture increases light intake),

adopting anti-reflection coatings and ultra-low dispersion glass; Infrared enhancement: Improves the quantum efficiency (QE) of sensors in the near-infrared band (700~1100nm) and expands the light-sensing

range.

Sensor Innovation: Large-size high-sensitivity sensors: Large-format and stacked sensors effectively enhance light-sensing capability; Infrared enhancement: The use of sensors with near-infrared enhancement technology captures brighter images; Dual-sensor fusion: Visible-light sensors capture colors while near-infrared sensors collect brightness, and algorithms synthesize

full-color images (e.g., Exview HAD technology). 3. AI-ISP: Driven by Intelligent Algorithms Deep Learning Empowerment: Spectral reconstruction: Recovers multi-spectral information from RGB images based on CNN networks, improving color authenticity. Dynamic noise reduction: Adaptively distinguishes moving targets from background noise, increasing the signal-to-noise ratio by

15dB compared with traditional algorithms at 0.001Lux. Modular Integration: Optimizes modules such as HDR and RLTM to solve the problem of dynamic range compression in low-illuminance

environments.

Conclusion Low-illuminance cameras are evolving from "being able to see" to "seeing clearly and accurately". With the development of AI-ISP

and sensor technologies, future core breakthroughs will focus on extreme environmental adaptability and the embedding of

intelligent analysis at the front end, driving low-illuminance imaging into an era of full-time, full-scene and full-color performance.