Spectral radiometric calibration: a key to the quantitative transformation of remote sensing data

With the continuous deployment of Earth observation platforms such as meteorological satellites, land satellites, and ocean satellites, remote sensing data is experiencing explosive growth. However, the raw data exists in the form of digital voltage values. If it cannot be accurately converted into physically meaningful spectral radiation information, it will be difficult to support the reliability of Earth observation and analysis, let alone meet the quantitative needs of fields such as climate change monitoring and disaster early warning.Spectral radiometric calibration techniquesThis is precisely the key to resolving this core contradiction.

From a technical perspective, the core of spectral radiometric calibration technology lies in establishing a precise correspondence between the actual radiance of ground objects and the instrument's output voltage signal across different spectral bands. This process not only achieves instrument calibration but also establishes a reliability benchmark for remote sensing data—ensuring consistency in radiance data of the same ground object acquired at different observation times and on different platforms. This provides support for remote sensing observation to transition from laboratory verification to global dynamic monitoring and promotes the transformation of remote sensing applications from qualitative description to quantitative analysis.

The calibration technology system for spectral imagers includes multiple implementation paths (as shown in Figure 1). Among them, the calibration process before satellite launch is particularly critical, and it is mainly divided into two core modules: laboratory calibration and field calibration.

Figure 1. Classification of Spectral Imager Calibration Techniques

Laboratory calibration is a core calibration step before the satellite leaves the factory. Its core objective is to eliminate inherent instrument biases (such as detector response non-uniformity), establish an initial accuracy benchmark for on-orbit observations, and avoid rework of massive amounts of data due to errors later on. This step specifically covers three key aspects: First, spectral calibration, using a monochromator as the standard light source, involves adjusting the output of monochromatic light at a specific wavelength in a stepwise manner to construct the instrument's spectral response function (SRF), accurately determining the center wavelength, bandwidth, and spectral line curvature of the detector pixels, and reducing ground feature identification bias caused by spectral drift; Second, relative radiometric calibration, which calculates the relative radiometric coefficient to achieve pixel-level response equalization based on the response differences between pixels in the detector array, eliminating image striping effects and edge attenuation, and ensuring image brightness uniformity (as shown in Figure 2); Third, absolute radiometric calibration, which uses an integrating sphere to simulate the diffuse light environment in space (as shown in Figure 3), and converts the digital count (DN value) output by the instrument into the physical quantity of radiative flux through the transfer equation. During this process, dark current offset, lens field of view attenuation, and gain factor need to be corrected to achieve end-to-end traceability from digital signal to physical quantity.

Figure 2. Comparison of a certain model of Xi'an Institute of Optics and Precision Mechanics CAS Aerospace Science and Technology Group Co.,LTD before and after relative calibration.

Figure 3. Schematic diagram of absolute calibration scheme for a certain model of Zhongke Xiguang.

Field calibration is the on-site verification step that connects the ideal laboratory environment with the actual on-orbit scenario. In the real outdoor atmospheric environment, the radiation and spectral mapping model established in the laboratory is verified and fine-tuned by simulating the satellite's on-orbit observation conditions using a ground-based standard source. Its technical value is reflected in three aspects: First, it incorporates correction mechanisms for environmental factors such as atmospheric scattering and absorption, further optimizing measurement accuracy and ensuring stable on-orbit observation capabilities immediately after satellite launch; second, it verifies the integrated response characteristics of the instrument and satellite platform, avoiding the impact of platform vibration and installation errors on data quality, and supporting the consistent fusion of multi-band data; third, it exposes the performance fluctuations and potential problems of the instrument in real-world environments in advance, reducing error risks in the initial launch phase and providing pre-operational assurance for the remote sensing system.

Laboratory calibration and field calibration together form a complete accuracy assurance chain of "benchmark establishment - field verification": the former eliminates inherent biases of the instrument and lays a solid foundation for high-precision observation, while the latter compensates for the differences between the ideal environment and the actual space scenario and gives the accuracy benchmark applicability. This collaborative mechanism provides a reliable basis for the conversion of remote sensing data from "digital signals" to "quantitative radiometric information," realizing the quantitative conversion of remote sensing data.