Progress and development trend of space backlight detection imaging technology

In recent years, rapidly advanced space technology has resulted in a growing number of objects in orbit, including both operational and decommissioned spacecraft, as well as space debris, which pose potential threats to China’s on-orbit spacecraft. To ensure the safe operation of spacecraft in near-Earth and high-orbit environments, a key research focus in the field of space technology is the high-resolution and high-accuracy detection and perception of environmental objects under strong backlight conditions. Such capabilities are critical for providing timely early warnings to on-orbit spacecraft. However, a plethora of studies have demonstrated marked challenges in imaging dark and weak targets under strong light backgrounds. It is essential to analyze the challenges of current backlight imaging technologies, summarize the working principles of existing backlight detection methods, and explore future development trends. In a review article recently published in Space: Science & Technology, a team from Harbin Institute of Technology examined the space backlight imaging scenario and the current state of backlight imaging applications, aiming to provide valuable insights and references for advancing space backlighting detection technology.

First, analyses of a space inverse imaging scene and the characteristics of backlight imaging for space targets are conducted. The schematic diagram of backlighting and forward lighting detection is illustrated in Fig. 1. A phase angle β greater than 90° between the direction of incident sunlight and the detector’s pointing direction is defined as backlighting, while a phase angle α smaller than 90° is defined as forward lighting. Imaging a target under backlight conditions is prone to issues such as detector saturation or damage due to excessive light exposure, degrading the target’s imaging quality. As a result, the detection target may be overwhelmed by the bright background and cannot accurately capture the geometric features or detailed textures of the target. Dynamic range is defined as the ratio between the full well capacity and the noise floor, representing the measurable luminance difference in an image. The essence of backlight imaging technology lies in enhancing the dynamic range of imaging. Thus, a high-dynamic-range (HDR) imaging detector is a prerequisite to realizing target imaging in a backlight background.

Subsequently, based on the solar radiation propagation path, the imaging technology for dim targets under strong light backgrounds is summarized and categorized into stages. Based on the propagation path of solar radiation, current backlight imaging technologies can be categorized into 4 main types, as illustrated in Fig. 2: (1) optical radiation energy attenuation imaging technology, (2) HDR detectors, (3) HDR backlight image synthesis technology, and (4) computational optical imaging technology.

Optical radiation energy attenuation backlight imaging technology utilizes solar filters, aperture reduction, and field-of-view adjustments at the front end of the detection imaging system. These measures reduce the solar radiation energy received by the imaging detector, thereby preventing detector saturation and damage. Commonly used structural combinations for attenuating optical radiation include front filters, multilayer film mirrors, and focal plane filters. Additionally, filters commonly used in the field of optics can be categorized into cutoff filters, band-pass filters, and solar spectrum correction filters. These filters are optical components or independent substrates coated with one or more layers of dielectric films, which alter the transmission characteristics of light waves. By leveraging the properties of these films—such as transmission, reflection, polarization, and phase changes—various filtering effects can be achieved in practical applications. While reducing the imaging field of view and optical aperture can also decrease the light radiation incident on the detector, this method markedly limits the detection field of view and is ineffective for achieving large-field-of-view perception.

When optical information reaches the detector, an HDR detector can be employed to capture richer image details. HDR imaging can be achieved by increasing the sensitivity of the sensor, but this method tends to make the sensor produce a nonlinear response and increase the difficulty of subsequent digital image processing. Other HDR detection imaging technologies, such as multiple exposure fusion of a single sensor, multisensor long- and short-exposure information fusion, intensity modulation of single-frame spatial exposure, dynamic intensity modulation, and special imaging devices, can obtain images with an HDR and realize broadband imaging of targets under strong light background. To date, leading global sensor manufacturers such as Sony, Samsung, and ON Semiconductor have developed HDR sensor products and equipment.

In the digital imaging stage, HDR image synthesis techniques are applied to generate higher-resolution backlight images, which generally combine multiple exposures with the varying brightness range of the same scene to restore the true illumination data. Currently, mainstream HDR image synthesis algorithms are primarily categorized into 2 approaches: radiance-domain-based methods and image-domain-based methods. In the radiance-domain-based method, a mathematical model is used to establish the relationship between the radiance domain and the image domain, constructing a scene radiance map from pixel values to restore the dynamic range of the entire scene as much as possible. The image-domain-based method skips the stage of estimating the radiation map and tone mapping and directly fuses LDR images under different exposure conditions into HDR images.

System-wide computational optical imaging is an emerging imaging technology that integrates optical hardware, image sensors, and algorithmic software. This technology processes light using the detector’s front-end optical imaging system and then recovers digital images, enabling clear imaging of dark and weak targets under strong light backgrounds. Polarization-based inverse imaging and wave-front coding are two main approaches. Polarization detection can capture multidimensional information about the target and eliminate sunlight interference, due to the fact that the reflection and radiation of electromagnetic waves generated by the inherent properties of a target exhibit specific polarization states while sunlight exhibits no polarization characteristics. Currently, polarization imaging technology can be broadly categorized into time-division polarization and simultaneous polarization. Based on the principles of Fourier optics, wave-front coding computational optical imaging technology introduces a specially designed asymmetric phase mask at the aperture of a traditional optical system. This mask performs aperture coding and phase modulation on the incident light, dispersing and suppressing strong light across a wide field of view.

Finally, the paper concludes with a summary and outlook. On the one hand, as the complexity of space exploration tasks increases, the scene of perceptual detection will vary constantly. Future optics methods should adapt to the complex and changeable task environment and adjust the illumination conditions of different scenes. It can be concluded that the backlight imaging will develop toward a combined structure. In this structure, the front end is a special lens system that can encode incident light information. The imaging detector with a large dynamic range receives broadband light information. Finally, the image information of the initial encoding processing is restored on the digital image, and the multiscene multiexposure image is taken for fusion. On the other hand, as an emerging imaging technology, computational optics integrates optical hardware, image sensors, and algorithmic software into a unified system. It overcomes the limitations of traditional imaging technologies in terms of depth (dynamic range and illumination) and breadth (spectrum, 3-dimensionality, and field of view) of information acquisition. The integration of multitype computational optics technologies, system-wide hardware devices, and multilevel algorithmic software provides greater flexibility and more solutions for computational imaging applications in backlight environments and other complex illumination conditions. In summary, emerging computational imaging technologies—based on filters, HDR detectors, HDR image synthesis, and integrated computational optics—can be applied to space detection in strong light backgrounds. These technologies provide critical support for information security and the safe operation of on-orbit spacecraft.