A Data-Driven Approach for Mitigating Dark Current Noise and Bad Pixels in Complementary Metal Oxide Semiconductor Cameras for Space-based Telescopes

TL;DR

This work tackles dark current noise and bad pixels in CMOS sensors for space telescopes by proposing a data-driven framework that combines pixel clustering with temperature-dependent function fitting. Pixels are clustered with a Gaussian Mixture Model to capture shared dark-current–temperature trends, and cluster-specific Arrhenius-inspired models map frame sequence (temperature proxy) to dark current, enabling both dark-current estimation and bad-pixel detection. The method is validated on Yangwang-1 data (near-UV and optical), showing reduced dark current, cleaned pixel masks, and improved target-detection performance when paired with SExtractor and Faster RCNN detectors. By relying on ground-based tests and on-orbit observations, the approach offers a practical path to enhance the science return of small satellites without active cooling, with potential applicability to future space telescopes and broader noise sources such as sky background in subsequent work.

Abstract

In recent years, there has been a gradual increase in the performance of Complementary Metal Oxide Semiconductor (CMOS) cameras. These cameras have gained popularity as a viable alternative to charge-coupled device (CCD) cameras in a wide range of applications. One particular application is the CMOS camera installed in small space telescopes. However, the limited power and spatial resources available on satellites present challenges in maintaining ideal observation conditions, including temperature and radiation environment. Consequently, images captured by CMOS cameras are susceptible to issues such as dark current noise and defective pixels. In this paper, we introduce a data-driven framework for mitigating dark current noise and bad pixels for CMOS cameras. Our approach involves two key steps: pixel clustering and function fitting. During pixel clustering step, we identify and group pixels exhibiting similar dark current noise properties. Subsequently, in the function fitting step, we formulate functions that capture the relationship between dark current and temperature, as dictated by the Arrhenius law. Our framework leverages ground-based test data to establish distinct temperature-dark current relations for pixels within different clusters. The cluster results could then be utilized to estimate the dark current noise level and detect bad pixels from real observational data. To assess the effectiveness of our approach, we have conducted tests using real observation data obtained from the Yangwang-1 satellite, equipped with a near-ultraviolet telescope and an optical telescope. The results show a considerable improvement in the detection efficiency of space-based telescopes.

Figures (10)

• Figure 1: The left panel shows the throughput of the optical telescope and the right panel shows that of the near-ultraviolet telescope.
• Figure 2: The same sky area captured by the near-ultraviolet telescope (the left panel) and the optical telescope (the right panel), when they are pointing to $RA=148.88822$ and $Dec=69.06529$. Since these two telescopes have different spatial resolutions, these two images have different field of views ($0.11^{\circ} \times 0.11^{\circ}$ for optical images and $0.17^{\circ} \times 0.17^{\circ}$ for near-ultraviolet images).
• Figure 3: The schematic of the data-driven dark current noise model of the CMOS camera encompasses two essential steps: the pixel clustering step and the function fitting step. Initially, we collect test data following the observation strategy. Subsequently, we cluster the pixels based on their temperature (frame sequence number)-dark current relations. Using the clustering results, we fit functions to each cluster, representing the temperature (frame sequence number)-dark current relationship specific to the CMOS camera. In real observations, we start by acquiring the CMOS temperature (frame sequence number) using the gray scale values of non-celestial object pixels. This information enables us to determine the dark current for all pixels in the CMOS camera.
• Figure 4: The left panel displays the BIC values corresponding to various clusters. The right panel shows the distribution of pixels across different clusters.
• Figure 5: The left panel shows the relation between logarithm of dark current level and 1/ Frame Sequence Number. The right panel shows the relation between the dark current level and Frame Sequence Number.