Chip HardwareWavelength range, number of spectral channels, spatial resolution, package type, data interface, etc.

Software AlgorithmsSpectral acquisition algorithm, reconstruction algorithm, auto white balance, image quality optimization, etc.

Built on proprietary computational spectral technology, Seetrum's full-chain multispectral imaging solution spans four layers: core sensors, module development, algorithm models and calibration systems. Designed for smartphones, action cameras and other smart imaging devices, it combines high performance, low power consumption and broad compatibility to deliver outstanding color performance.

I. Multispectral Chips

Seetrum's consumer-grade multispectral imaging chips cover the core wavelength range of 380–960 nm, with up to 12 physical spectral channels. Programmable functions include frame-rate adjustment, image flipping, windowing and multiple binning modes. The chips support standard output formats such as 10-bit RGB RAW.

Each chip integrates OTP memory, static defect-pixel correction and automatic black level calibration. Dual phase-locked loops (PLLs) and multi-camera synchronization interfaces ensure excellent imaging consistency and synchronized operation across multiple devices. Built on an advanced stacked BSI (backside-illuminated) pixel architecture, the chip sustains 60 fps at resolutions from 0.5 MP to 1.3 MP, with power consumption as low as 15 mW — balancing high performance and low power to meet the slim-profile and long-battery-life requirements of smartphones, action cameras and similar devices.

II. Spectral Algorithms & Applications

Seetrum has built an end-to-end algorithm chain spanning spectral data acquisition, spectral reconstruction, auto white balance and image quality optimization.

Our proprietary global multispectral auto white balance (AWB) algorithm accurately estimates key parameters — including the light source spectrum, illuminant type and correlated color temperature (CCT) — and supports both color constancy and chromatic adaptation output modes. A one-click reconstruction and rendering tool enables intuitive, efficient evaluation of imaging results.

Validated across comprehensive scenario testing covering indoor, outdoor, multi-illuminant and solid-color background environments, the solution significantly reduces the inherent color deviation of raw sensor output, delivering results closely aligned with reference standards. It also addresses the long-standing industry challenge of color inconsistency across multiple cameras.

III. Spectral Calibration System

To address long-standing industry challenges in the manufacturing processes and consistency of multispectral sensors, Seetrum has developed a large calibration model that enhances spectral consistency. It overcomes technical bottlenecks in high-dimensional spectral data representation and enables high-performance implementation on mainstream production processes.

Our proprietary calibration algorithms generalize strongly across devices. In production testing across 1,092 modules, they improved spectral and CCT calibration performance by more than 83%, and by up to 90% — raising both yield and production throughput.

In addition, we have established an end-to-end calibration system that combines local and cloud platforms and spans every stage of module production. This standardized, intelligent calibration flow helps ensure consistent performance from device to device.