Abstract
This article reviews the development and frontier challenges of infrared sensing technology in two major domains: "looking upward" for deep space exploration and "looking downward" for Earth observation. In deep space exploration, represented by the James Webb Space Telescope (JWST), this technology is constantly approaching the physical limits of observation through methods such as ultra-large apertures, very long wavelength band, and ultra-low temperature cooling, aiming to reveal the mysteries of the early universe. The article also elucidates the primary evolution of infrared detection payloads aboard various satellites across different eras, highlighting their representativeness and characteristics. Infrared Earth observation technology has progressed from low spatial resolution and a limited number of bands over a wide swath to improved temporal, spatial, spectral, and radiometric resolution over broad spatial and spectral ranges. New technologies such as large-aperture low-background optics, long-wavelength high-sensitivity detectors, on-chip intelligent sensing, and big data twin systems, as well as the large-scale development of "real-time remote sensing" integrating communication, navigation, and remote sensing, and commercial aerospace, will help build the foundation of "Infrared Digital Earth", enhance human beings′ real-time understanding and accurate prediction of the evolution of anomalous events in the Earth′s multi-spheres and various complex cyclic processes, promoting the popularization of infrared Earth observation technology.
0 Introduction
This study focuses on the infrared band ranging from 1.0 to 1000 µm. Infrared spectroscopy technology analyzes the composition and structure of matter by transmitting unique"molecular fingerprint" information from infrared light. It can detect the reflection-emission characteristics of objects at different wavelengths from a distance and has become a core means for mankind to break through the boundaries of observation and expand the frontiers of knowledge.
In deep space exploration that"looks upwards", due to the continuous expansion of the universe, photons from the early universe are stretched to the infrared band [1], enabling infrared technology to trace the origin of the universe, explore ancient galaxies and various cold celestial bodies. Human beings have established a diverse observation system ranging from planetary probes(such as Juno [2] and New Horizons [3]), interstellar spacecraft(such as the Voyager series [4]), to ground-based large interferometers(such as the Very Large Telescope Interferometer, VLTI [5]) and space telescopes(such as the Herschel Space Observatory, HSO [6] and JWST [7]).
To explore the extremely faint signals of redshift in celestial bodies after the Big Bang, ultra-low temperatures, ultra-large apertures, and high-sensitivity infrared detection have become key cutting-edge technologies for infrared deep space exploration. JWST represents the highest achievement to date, but how to define the physical and technological limits of infrared observation on a light-year scale, and thus clearly reveal the earliest evolutionary processes of the universe, remains a cutting-edge scientific question concerning the boundaries of human cognition.
In the field of Earth observation that"looks downwards", early developments were represented by panchromatic or RGB broadband cameras in the visible light band. Technological advancements primarily focused on improving spatial resolution to meet the demands for detailed observation of Earth's surface morphology. Infrared spectroscopy, by capturing unique information about surface and atmospheric radiation, reveals thermal structures and energy balance states that are imperceptible to visible light. Its detection bands have gradually expanded from the early near-infrared band to shortwave, midwave, and even longwave/very longwave infrared bands. Because the wavelength of infrared light is much longer than that of visible light, its spatial resolution has always been relatively low, and it has mainly been used for wide-swath or multi-spectral payloads, such as those used by NOAA in the United States [8], EOS [9], Landsat [10] series satellites, Europe's Meteosat [11] series satellites, as well as China's Fengyun [12], Haiyang [13] and Environmental Disaster Reduction [14] series satellites, etc.
Due to limitations in technology and the sheer scale required for larger apertures, the development of high spatial resolution infrared observations has been relatively slow, and the number of observations conducted is limited. Infrared detection with lower spatial resolution and ultra-high spectral resolution over a wider swath has attracted more attention and achieved significant technological breakthroughs due to its application value and suitability to the inherent characteristics of infrared light. The Atmospheric Infrared Sounder(AIRS) onboard the US Aqua satellite exemplifies this [15]. The Geostationary Interferometric Infrared Sounder(GIIRS) carried on China's Fengyun-4(FY-4) satellite is a major example.
However, infrared hyperspectral imaging surface observation payloads that combine high spatial resolution and high spectral resolution have developed most slowly because they need to take into account both the large range and high resolution in the spatial-spectral dimensions. In 2018, the U.S. National Academy of Sciences called spaceborne hyperspectral imaging"the only method for space-based identification of ground features and quantification of biochemical properties" in the"Decade Earth Observation Plan" [16]. In May 2018, China's Gaofen-5 hyperspectral satellite was successfully launched, and the world’s first wide-spectrum, wide-swath hyperspectral camera—Advanced Hyperspectral Imager(AHSI)—entered orbit and began operation [17]. Subsequently, the Italian PRISMA(2019) [18], the German EnMap(2022) [19] and the American Tanager-1(2024) [20] satellites carrying similar cameras were launched and entered orbit, and hyperspectral imaging technology entered a new stage of development. To date, the working band of spaceborne hyperspectral cameras for Earth observation reported internationally only extends to 2.5 μm.
The intelligent development trend of Earth observation in the future is becoming increasingly clear. The intelligent development of space-ground collaboration from constellations to ground-based public terminal services has become an inevitable trend [21]. In order to achieve real-time and refined perception of the Earth system, breakthroughs are still needed in key areas such as multidimensional extreme detection, intelligent terminals, on-chip systems and full-domain digital twin platforms. The rise of commercial aerospace has injected strong momentum into the development of space-ground collaborative applications. Since the State Council issued the"2016 China's Space Activities" white paper, China's commercial aerospace has flourished, and emerging forces represented by commercial remote sensing constellations have grown rapidly [22-23]. The"2021 China's Space Activities" white paper further established support policies. The core of commercial remote sensing lies in popularization, large-scale and ubiquitous application. This will greatly promote the development of infrared spectroscopy technology for Earth observation.
This article addresses the cutting-edge development of infrared sensing technologies, proposing key scientific issues and technological challenges that need to be overcome in the current and future, focusing on the limits of deep space exploration"looking up" and the new paradigm of digital earth"looking down", and also looks forward to cutting-edge research directions that will promote their large-scale popularization and application.
1 Development and frontier Issues of infrared astronomical observation
China's first national space science plan, the"National Medium-and Long-Term Development Plan for Space Science(2024-2050)", prioritizes the exploration of the origin and evolution of the universe, as well as missions such as high-precision infrared observation in space. What are the limits of infrared astronomical observation? To see farther and clearer, continuous advancements are needed to overcome the limitations of Earth's atmosphere and the instruments themselves. This is mainly manifested in the expansion of observation platforms from ground to space, the continuous increase in the aperture of observation instruments, and the optimization of detection wavelengths, dark current detection, and cooling technologies, pushing detection sensitivity towards its limits.
The development of infrared astronomy has been a continuous breakthrough from the ground to space. Since the 1960s, ground-based infrared observation has achieved many important results. Ground-based telescopes collect and focus starlight to detectors. The increased aperture of the telescopes not only improves the light-gathering ability, but also helps to obtain images with higher spatial resolution within the diffraction limit. Building larger telescopes faces multiple challenges in engineering, including physical laws, atmospheric conditions, technical bottlenecks and resource limitations. Even on the ground, the construction of larger telescopes still faces major technical difficulties [24]. In addition, although ground-based telescopes use high-altitude stations to avoid some atmospheric interference, the absorption and radiation of the Earth's atmosphere itself limits observation. In the early days, in order to overcome the limitation of residual atmospheric absorption, the National Aeronautics and Space Administration(NASA) developed the Kuiper Airborne Observatory(1974-1995) [25] and the Sofia Stratospheric Infrared Observatory(2010-2022) [26].
The real revolutionary progress comes from space infrared telescopes. Due to their ability to conduct far-infrared observations that are difficult to achieve on the ground and their advantages of being unaffected by the atmosphere, they are the inevitable choice for obtaining high-precision, high-quality astronomical spectral data [27]. The aperture of space infrared telescopes is gradually increasing, and their detection capabilities are also constantly improving. Tables 1 and 2 list the main indicator data of typical large-aperture telescopes at home and abroad. At present, space infrared observation, represented by the JWST launched in 2021, is opening a new era. It is deployed at the Sun-Earth Lagrange point L2, far from the Earth's heat source, with a 6.5 m-diameter spliced mirror and the technology of deep cooling to below 7 K(approximately-266 °C), it has achieved unprecedented sensitivity and resolution in the near-infrared and infrared bands of 0.6-28.3 µm for the first time. Currently, the limit of human detection that JSWT can reach is 13.8 billion light-years. Scientists at the Bohr Institute at the University of Copenhagen in Denmark used the near-infrared and mid-wave infrared cameras on JSWT to discover that the earliest reionized(an important transition in the early universe) galaxies may have been formed 330 million years after the Big Bang. The relevant results were published in Nature [28]. However, astronomers speculate that the observable diameter of the universe is about 93 billion light-years [ 29-30].
Table
1
Key Indicators of Typical Large-Aperture Telescopes in China
Table
2
Key Indicators of Typical Large-Aperture Telescopes Abroad
Human exploration of the limits of cosmic observation has revealed several core challenges in infrared observation(see Figure 1). First, ultra-large aperture: to capture extremely faint signals from nearly a hundred billion years ago, future development of deployable/on-orbit mosaicking technologies at the 10 m or even 100 m level is needed. The Shanghai Institute of Technical Physics, Chinese Academy of Sciences, has already begun research and development on 6 m aperture aspherical mosaicking primary mirror technology. Second, extremely low temperature environment: to detect celestial radiation against the cosmic background(approximately 3 K), the key components of the optomechanical system and the detector itself need to be cooled to extremely low temperatures(e.g., below 4 K). Third, the redshift-driven shift to far-infrared detection bands: the demand for very long-wave infrared(VLWI) detection is rising sharply, requiring the deployment of VLWI payload satellites to conduct light-year-scale deep space exploration using VLWI imaging. Fourth, intelligent processing: it requires the use of massive computing power and ultra-large intelligent model algorithms to extract extremely distant and extremely weak near-edge celestial signals from complex cosmic signals and noise.
Fig.
1
Major challenges in the limits of infrared astronomical detection.
In the future, the scope of astronomical research will expand to a more ambitious observational system architecture. Cutting-edge concepts include an international lunar research station, utilizing its exceptionally unobstructed atmospheric environment to deploy an array of ultra-large aperture telescopes; or extending the observation platform to more distant heliocentric orbits or Martian bases to obtain longer interferometric baselines, thereby achieving ultra-high resolution imaging similar to"synthetic aperture". A distributed"astronomical observation network" stretching from near-Earth orbit, lunar space, to interstellar space is moving from conception to planning.
2 Development and frontier issues of infrared earth observation
Because of the long wavelength of infrared light, achieving spatial resolution comparable to visible light requires larger aperture optical equipment with strong background radiation suppression capabilities, resulting in a large and complex payload. Furthermore, the Earth itself is a massive source of infrared radiation, making target signals susceptible to high background noise and hindering the cooling of the front-end optical system, placing stringent demands on the sensitivity and anti-interference capabilities of the detector. These factors have led to a relatively slower development of infrared remote sensing technology compared to visible light remote sensing.
The origin of infrared remote sensing can be traced back to the infrared horizon sensor carried on the American"Tiros II" satellite in 1960, which achieved the first infrared experiment in space. In 1970, the conical scanning infrared horizon sensor probe, developed by Kuang Dingbo and others from the Shanghai Institute of Technical Physics, Chinese Academy of Sciences, was launched into space aboard China's first artificial Earth satellite,"Dongfanghong-1", marking the official start of China's first generation of infrared space remote sensing. Early infrared technology was used only for simple observations; for example, infrared horizon sensors were used to provide information on satellite attitude deviations relative to the Earth.
In the decades that followed, infrared remote sensing technology gradually developed, making significant progress, especially in meteorology and oceanography. Its payloads were mostly wide-swath or multispectral imagers with low spatial resolution and a limited number of channels(see Table 3). For example, the European Meteosat series of geostationary meteorological satellites, developed over nearly 50 years since 1977, evolved from the first-generation Meteosat Visible and Infrared Imager(MVIRI) with only one infrared channel, to the second-generation Spinning Enhanced Visible and Infrared Imager(SEVIRI) with five infrared channels, to the third-generation Flexible Combined Imager(FCI) with eight infrared channels. The Advanced Very High Resolution Radiometer(AVHRR) carried by the US NOAA series of meteorological satellites has undergone three generations of updates since its launch in 1979: the number of infrared channels increased from three to five, and the spatial resolution increased from 4 km increased to 1.1 km.
The Visible and Infrared Spin Scan Radiometer(VISSR) carried on the FY-2A satellite launched by China in 1997 had two infrared channels(with a spatial resolution of 5 km). In the second generation, the number of infrared channels increased to four. The Advanced Geostationary Radiation Imager(AGRI) on the FY-4A satellite launched in 2021 has 11 infrared channels, with the highest spatial resolution reaching 2 km. In terms of ocean observation, China's Haiyang-1 series satellites also employ wide-swath payloads suitable for ocean exploration. For instance, the new generation of the Ocean Color and Temperature Scanner has a swath width of over 3,000 km and five infrared channels [31]. The Moderate Resolution Imaging Spectroradiometer(MODIS) carried on the US Terra satellite(launched in 1999) and Aqua satellite(launched in 2002) has 32 infrared channels, with a spatial resolution of 0.5 km in the short-wave infrared band and 1 km in the mid-wave infrared and thermal infrared bands.
With the continuous improvement of indicators of spatial resolution and other indicators, infrared Earth observation has gradually expanded to terrestrial applications. The Thematic Mapper(TM) instrument on the Landsat-5 satellite launched by the United States in 1984 had only one thermal infrared channel(with a spatial resolution of 120 m and a swath width of 185 km). The Advanced Spaceborne Thermal Emission and Reflection Radiometer(ASTER), carried by the Terra satellite launched in 1999, has five thermal infrared channels(with a spatial resolution of 90 m and a swath width of 60 km). The Multi-Spectral Imager(MSI) on the European Sentinel-2A/B satellites launched in 2015 and 2018 respectively have three short-wave infrared channels, providing multi-spectral images with a maximum spatial resolution of 20 m. The infrared cameras on the Chinese Environment Disaster Reduction-2A/B satellites launched in 2020 have nine(three thermal infrared) bands, a swath width of 720 km, and a spatial resolution of 48 m/96 m. The infrared camera on the Sustainable Development 1 satellite launched in 2021 has three thermal infrared bands(with a spatial resolution of 30 m and a swath width of 300 km). The Wide-swath Thermal Infrared Imager(WTI) on the GF-5 01A satellite launched in 2022 has four thermal infrared bands, with a spatial resolution of 100 m and a swath width of 1500 km.
Table
3
Key indicators of early and current representative earth observation infrared payloads
Note: The load parameters in the table only list the infrared-related parts.
Meanwhile, infrared detection has been developing towards higher spectral resolution, but its spatial resolution remains relatively low, primarily used for vertical infrared atmospheric sounding. For example, the first hyperspectral resolution infrared sounder—the Atmospheric Infrared Sounder(AIRS)—was carried on the Aqua satellite launched by the United States in 2002, increasing the number of infrared channels from a few to 2378. China's GIIRS, carried on the FY-4B satellite launched in 2021, is the world's first operational geostationary hyperspectral infrared sounder.
In 2000, the US EO-1 satellite carried the Hyperion infrared hyperspectral imaging Earth observation payload, which combines high spatial resolution and high spectral resolution, covering a wavelength range of 0.4–2.5 nm with a width of only 7.5 μm. The signal-to-noise ratio in the short-wave infrared band is low. Subsequent developments of spaceborne hyperspectral cameras, both domestically and internationally, only extended to approximately 1.0 μm in the near-infrared region. China's Gaofen-5(GF-5) satellite, launched in 2018, carried the world's first spaceborne wide-spectrum, wide-swath hyperspectral imager—AHSI(operating from 0.4 to 2.5 μm with a swath width of 60 km). Subsequently, Europe and the United States successively launched spaceborne hyperspectral imagers with spectral coverage also spanning 0.4–2.5 μm(swath width ≤ 30 km at 30 m spatial resolution). In 2023, China launched the world's first hyperspectral imager with a spectral range covering 0.4–12.5 μm(spatial resolution: 10 m/40 m). Tables 4 and 5 list the key specification data of representative early and current near-infrared/infrared hyperspectral Earth imaging payloads developed domestically and abroad, respectively.
Table
4
Key indicators of representative early and current near-infrared/hyperspectral earth imaging payloads in China
Table
5
Key indicators of representative early and current near-infrared/hyperspectral earth imaging payloads from abroad
With the rapid development of infrared technology and the increasing impact of human activities on Earth's ecology, environment, climate, and resources, the need to understand the interactions of multiple spheres(ice-snow-rock-soil-vegetation-water-air) and the complex evolution of processes such as the air-temperature-carbon-water cycle has become more urgent. Infrared Earth observation technology is undergoing a profound paradigm shift from traditional imaging to multidimensional, precise perception. The core of this transformation lies in constructing a completely new technological development path, with the following key scientific questions and cutting-edge directions: exploring novel detection mechanisms that couple multidimensional information such as spatiotemporal, spectral, and radiation intensity to collaboratively achieve observation goals with wide coverage, high resolution, and high quantitative accuracy; breaking through the limits of physical parameter detection capabilities for key elements in multiple global spheres; solving the complex system challenges brought about by high dynamic range and multi-element collaborative observation; and ultimately developing an infrared virtual Earth system based on intelligent learning from multi-source data. Within this framework, the constructed"multidimensional, multi-sphere infrared digital Earth" will become a major infrastructure for realizing fully digital Earth perception, providing crucial support for accurately simulating and predicting the evolutionary patterns and cyclical processes of the Earth system, and promoting humanity's" ubiquitous and constant" fully digital perception of the Earth. Achieving multi-dimensional infrared limit detection at the minute, sub-meter, nanometer, and mK scales, as well as constructing a digital infrared base for the Earth, has become the main development trend of infrared Earth observation(see Figure 2).
Fig.
2
Development paradigm of infrared earth observation systems.
The development of infrared technology in multi-dimensional, spatiotemporal, spectral, and radiometric detection will generate massive amounts of data. However, it also faces significant challenges, including weak intelligent processing capabilities, slow system response, and long application service chains, leading to time-consuming and energy-intensive infrared information systems. This contradicts the goals of real-time and mass-market applications. Therefore, it is necessary to construct embedded intelligent terminals based on satellite platforms that enable real-time, definable, and reconfigurable data access. This allows for the transition from complex sensing to real-time service, where end-users can remotely define observation parameters(such as band, resolution, and target type) from the ground and receive real-time data downloads. Among these challenges, programmability of spaceborne infrared spectrometers, lightweight deployment of neural network models, and the operation of high-energy-efficiency edge artificial intelligence(AI) chips on satellites remain unresolved and present significant research difficulties.
In recent years, high-end third-generation infrared and photoelectric detectors, characterized by ultra-high resolution, ultra-high sensitivity and ultra-fast response, have made significant breakthroughs both domestically and internationally. Their detection capabilities in multiple dimensions such as space, energy, time and spectrum have been continuously improved and are gradually approaching the theoretical limit [32]. However, the achievement of high performance has long depended on extremely low operating environments. Infrared detectors and even optomechanical systems need to operate at low temperatures of tens of Kelvin to suppress detector dark current and optomechanical background radiation. This has led to a large and expensive cooling system, which in turn has resulted in the high size, cost and power consumption of the entire device, which has seriously restricted its popularization and application. Therefore, how to improve the operating temperature of detectors and optomechanical systems through the innovation of new materials and structural systems has become an important way to promote the field towards low cost, miniaturization and large-scale manufacturing.
Driven by new physical mechanisms, China's infrared detection technology is evolving toward multi-dimensional perception, on-chip integration, bionic detection, and integrated sensing and computing. Among them, the deep integration of digital twin technology and infrared spectral detection technology is becoming a cutting-edge trend. By constructing virtual mappings of physical entities, a new paradigm of real-time monitoring, simulation prediction, and optimized control has been realized for global networked infrared remote sensing [33]. Existing research has successfully constructed high-precision dynamic scenes such as Poyang Lake wetland by integrating multi-source data and game engines, verifying its potential in improving scene fidelity and interactive experience [34]. From digital twins to digital prototypes, and then to digital application scenarios, infrared detection technology will develop from"system demonstration–design–research and development–application" to a new paradigm of"complete real-time seamless connection and integration of reality and virtuality".
The rapid development of commercial spaceflight has injected new momentum into the construction of a space-air-ground collaborative observation system. For example, low-Earth orbit(LEO) communication constellations, represented by Starlink, not only enhance global data transmission capabilities, but their satellite platforms can also be used as new carriers for remote sensing payloads. Furthermore, companies such as Planet and Capella Space have significantly improved the efficiency and spatiotemporal coverage of infrared and other band remote sensing data acquisition through high-revisit-frequency, high-resolution microsatellite constellations. Innovation in business models has also driven the cost reduction, popularization, and industrialization of data services, accelerating the expansion of remote sensing technology from specialized fields to public services. China's currently widely used BeiDou navigation satellites, and the developing"StarNet" and"Qianfan" constellations, which will reach tens of thousands of LEO communication satellites, combined with the unstoppable development trend of commercial remote sensing satellites, will facilitate the formation of integrated communication, navigation, and remote sensing, inter-satellite interconnection, and space-ground communication for real-time remote sensing. This will significantly improve the monitoring capabilities for sudden anomalies on the Earth's surface, enabling the public application of infrared remote sensing technology and ushering in a new chapter in the integrated development of infrared technology.
3 Conclusion
Infrared technology plays an irreplaceable role in humankind's exploration of the origins of the universe" looking up" and the multi-layered evolution of the Earth system"looking down". It is rapidly developing towards breakthroughs in detection limits and the intelligentization and ubiquity of the sensing system. At the same time, it faces systemic challenges from physical laws, technological bottlenecks, and application requirements.
In the field of space exploration, which is"looking upwards", the evolution of infrared technology is reflected in the shift of observation platforms from ground-based to space-based, the continuous increase in telescope aperture, the advancement of cooling technology towards extremely low temperatures, and the expansion of detection wavelengths to the longer very long-wave infrared band. The frontier for the future points to the construction of a distributed astronomical observation network. Through the development of ultra-large-aperture on-orbit segmented mirrors, very-long-wavelength infrared and cryogenic detection, as well as interferometric arrays deployed at cislunar and even interstellar base stations, humanity will be able to better understand the earliest history of the universe and continuously expand the spatiotemporal boundaries of human knowledge.
In the field of Earth observation, which focuses on"looking downwards", infrared technology has progressed from early low spatial resolution and a limited number of bands to simultaneously improving spectral and spatial resolution across a wider band range and swath width. Furthermore, it has enhanced the temporal resolution of infrared observations through geostationary orbits or constellations. Achieving multi-dimensional infrared limit detection at the minute, sub-meter, nanometer, and mK levels, along with constructing a digital infrared base for the Earth, has become the main development trend of infrared Earth observation. The rise of digital twins, artificial intelligence, big data processing, and commercial aerospace is jointly driving infrared Earth observation towards real-time, widespread, and mass-market service applications that integrate space-air-ground coordination and communication/navigation/remote sensing. This provides crucial support for accurately simulating and predicting the evolution and cyclical processes of the Earth system, and for promoting a fully digital, ubiquitous, and constant perception of the Earth.
This article primarily focuses on the forefront of space infrared technology, aiming to provide a framework and reference for the development of infrared technology. Due to space limitations, the connotation and extension of infrared technology development extend far beyond this, and there is an urgent need for more scholars and experts in the field of infrared research to summarize and share their findings, jointly promoting the continuous progress of infrared technology.