Open Access
Add to BibSonomy
Abstract
As a new type of active Earth observation technology, airborne hyperspectral lidar combines the advantages of traditional lidar 3D information acquisition and passive hyperspectral imaging technology, and it can achieve integrated imaging detection with a high spatial and hyperspectral resolution. Thus, it has become an important future direction of Earth surface remote sensing technology. This article introduces the design and development of an airborne hyperspectral imaging lidar system. The hyperspectral lidar adopts a focal plane splitting method, combined with an array of 168 optical fibers, to couple wide-spectral-range laser echo signals one by one to the corresponding single tube detector, achieving efficient splitting and precise coupling of supercontinuum laser pulse echo signals. This article proposes a fast synchronous calibration method that is suitable for hyperspectral imaging lidar systems. Results show that the spectral range of the hyperspectral lidar system is 400–900 nm, and the spectral resolution of single-fiber detection is greater than 3 nm. Notably, this article focuses on analyzing the abnormal detection channels based on the calibration results. With the test results of adjacent channels combined, the reason for the abnormal spectral bandwidth of channel 17 is analyzed as an example. This research points out the direction for verifying the design parameters of the hyperspectral lidar prototype and lays an important foundation for airborne flight test of the hyperspectral lidar.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Airborne lidar technology, as an important air-to-ground observation technology [1–3], is playing an increasingly important role in many applications such as forest remote sensing [4], environmental monitoring [5], urban planning [6], and ocean exploration [7,8] through active laser remote sensing detection, combined with technologies such as global positioning system and inertial measuring unit. As the core means of Earth observation, remote sensing technology has gradually become a comprehensive technical system that covers various information acquisition technologies such as passive multispectral imaging, hyperspectral imaging, and lidar remote sensing, and it can be carried on multiple platforms such as ground-based, airborne, and spaceborne platforms for Earth observation [9,10]. It is also applied to cope with global environmental changes, rapid urban development, national defense construction, and other processes that present many difficulties and challenges. The realization of all-band, all-sky, all-weather, and high-spectral-resolution Earth observation is an important opportunity for the development of remote sensing technology, especially the integration and application of high-resolution three-dimensional information and spectral information of the target space, which is a cutting-edge scientific problem that needs to be solved when developing Earth observation technology [11].
Active and passive remote sensing detection methods have shown unique advantages over the main remote sensing technology methods. Active lidar ranging can achieve all-day and high-resolution three-dimensional information acquisition of ground objects [12]. However, the application of lidar is limited to a certain extent because of the inadequate ability of the single-wavelength detection mechanism to acquire spectral physical property information of the target [13]. Passive hyperspectral imaging can obtain rich high-resolution spectral information, but its imaging mechanism is greatly affected by weather, illumination, and other factors, and it has poor spatial information detection ability [14]. Both active and passive remote sensing techniques have outstanding advantages but face shortcomings in the simultaneous acquisition of three-dimensional spatial and spectral information. Therefore, developing new remote sensing technology methods to enhance the spectral detection ability of lidar has become an important direction for the development of remote sensing technology to achieve the integration of high-resolution three-dimensional spatial and spectral information for ground targets while retaining the advantages of airborne lidar spatial detection.
The spectral acquisition capability is improved by increasing the detection bands of lidar. The advantages of active and passive remote sensing technology are combined into one sensor, and the integrated acquisition of high-resolution three-dimensional spatial and spectral information of the target is successfully realized. Studying hyperspectral lidar technology has become a hot research direction for relevant scholars [15]. At present, the development of hyperspectral lidar prototypes is still in the ground-based testing stage. For integrated acquisition of remote sensing information of Earth observation, airborne hyperspectral lidar has become a new active remote sensing detection technology that can perform full waveform data acquisition of dozens or even hundreds of spectral segments of the ground object spectrum and is capable of all-sky, all-weather, full-spectrum, and high-spatial-resolution detection.
This research team designed and developed an airborne hyperspectral imaging lidar system that has a spectral range of 400–900 nm, a single channel spectral resolution better than 10 nm, and a total of 56 spectral detection bands. This article first introduces the imaging mechanism of airborne hyperspectral lidar and further elaborates on its overall design concept. The key unit technology in particular was analyzed in combination with the airborne application platform, with the hardware technology of the hyperspectral lidar system as the research objective. Specifically, it includes the design of the optical receiving system for hyperspectral lidar, efficient spectral coupling of wideband echo signals, and precise detection. Finally, through laboratory calibration of the hyperspectral lidar system, the theoretical design parameters of the hyperspectral lidar were validated, and the performance of the entire system was evaluated, focusing on analyzing the performance deviations that may occur in specific detection channels.
2. Design of the instrument
The airborne hyperspectral imaging lidar system consists of a supercontinuum laser source, an optical signal receiving system, a signal detection and acquisition system, and a signal synchronization online processing module. The optical signal receiving system is an important component of the hyperspectral lidar system, and its optical receiving efficiency directly affects the detection performance of the entire hyperspectral lidar system. The supercontinuum spectrum laser is limited by the emission mechanism, and the average spectral energy output is very weak. The problems of hyperspectral spectral reception and precise detection of its echo signal need to be solved. However, as the number of spectral channels increases, multichannel and wide spectral band splitting can make the optical receiving system more complex and cause significant attenuation of the echo energy. Moreover, after multiband splitting of hyperspectral lidar echo signals, a key problem that needs to be solved in the development of hyperspectral lidar systems is how to achieve efficient photoelectric detection of dozens of spectral band echo signals and ensure high-precision synchronous acquisition of echo signals from various detection channels.
As shown in Fig. 1(a), the hyperspectral lidar optical receiving system consists of optical components such as a scanning rotating mirror, an off-axis parabolic mirror, a reflector, a small aperture, a collimating mirror, a grating, and a telecentric focusing lens. The off-axis parabolic mirror does not have a central obstruction, which can fully utilize the effective receiving area of the mirror to reflect the echo signal and can solve the problems of weak echo signal and small field-of-view reception in hyperspectral lidar systems. The surface of the off-axis parabolic mirror is evaporated with a highly reflective film, further improving the efficiency of echo signal reception. The small aperture can limit the receiving field of view of the hyperspectral lidar system and play a certain limiting role in the sky background light and system stray light. The grating spectrometer structure of the hyperspectral lidar system is composed of a transmissive grating splitter and a telecentric focusing lens. The design of the telecentric focusing lens needs to be consistent with the numerical aperture of the hyperspectral lidar optical receiving system to further suppress stray light signals.
To achieve synchronous detection of different spectral bands in hyperspectral lidar, existing technologies mainly adopt a multi-photosensitive surface detector array scheme, which utilizes a highly integrated linear or planar detector array composed of multiple photosensitive detection surfaces for detection [2,3]. It can minimize the volume of the lidar system as much as possible but also has obvious disadvantages. For example, a linear array or surface array detector array cannot independently adjust the gain of each detection channel. The nonlinear effect of broad-spectrum laser sources leads to uneven distribution of the output energy of supercontinuum laser in different spectral bands, especially in the ultraviolet band where the laser energy is much weaker than existing single wavelength lidars, and the echo signal intensity even reaches the level of a single photon. The increase in the number of spectral bands detected by hyperspectral lidar and the detection of weak echo signals mean that the bandwidth, sensitivity, and gain of traditional line array and surface array detectors cannot effectively meet the detection and processing requirements of hyperspectral lidar echo signals [16]. The hyperspectral lidar system prototype is shown in Fig. 1(b). Its detection and acquisition module uses a single tube detector, which solves the problem of the bandwidth, sensitivity, and gain of traditional line array and surface array detectors being unable to satisfy the spectral data detection and processing of hyperspectral imaging lidar echo signal. It can flexibly set the corresponding performance parameters according to the actual strength of the echo signal of each detection channel, thus expanding the spectral detection range of the hyperspectral lidar system and improving the spectral detection accuracy. The detection module of the hyperspectral lidar system consists of a photomultiplier tube (PMT) detector and an APD detector, and the echo signal coupling module and circuit control unit of the APD detector are independently designed [17].
The energy of each pulse emitted by the supercontinuum laser may dither because of the limitation of the luminescence mechanism of the supercontinuum laser. The influence of the energy jitter of the supercontinuum laser on the measurement results needs to be eliminated to obtain the spectral information of ground objects accurately. As shown in Fig. 1(a), although most of the light emitted by the laser will change direction and incident on the surface of the scanning rotating mirror after passing through the reflector, a small amount of light will still pass through the reflector. The wide spectrum laser spectrum is monitored in real time so that a small amount of transmitted light can be collected and used as reference light of the hyperspectral lidar system. We designed a coupling lens at the back end of the reflector. Through this lens, a small amount of transmitted light is focused into the fiber into the hyperspectral lidar optical receiving system and the energy change of each laser pulse of the supercontinuum laser is monitored in real time.
The reference light signal is coupled to the hyperspectral lidar optical receiving system through fiber, and the hyperspectral lidar system itself can complete real-time detection of the wide laser spectrum. The design method simplifies the structure of the hyperspectral lidar system, thus allowing the hyperspectral lidar system to complete real-time monitoring of the energy of the supercontinuum laser source without other auxiliary equipment. The length of the fiber is set to 50 m, which ensures the detection accuracy of the hyperspectral lidar system, to separate the near-field stray light signal reflected by the rotating mirror from the reference light signal in the time domain.
3. Calibration methods
Hyperspectral imaging lidar is a new Earth observation technology. An important measurement purpose of hyperspectral lidar, unlike the traditional single-wavelength lidar, is to obtain backscatter spectral information of ground objects [18,19]. However, before the spectral information measured by the hyperspectral lidar system is applied in ground object classification, the spectral intensity of the hyperspectral lidar system itself needs to be corrected [20,21]. Calibrating the hyperspectral lidar system is important for the practical application of echo signal data [22].
3.1 Spectral calibration
The spectral resolution of this hyperspectral imaging lidar system is different from the uniform distribution of spectra in various channels of traditional passive hyperspectral measurement instruments. The hyperspectral lidar system adopts multi-fiber splitting technology, ensuring that the spectral resolution of each channel can be designed according to the application requirements of ground object classification. The spectral resolution of a hyperspectral lidar system is determined by the passband width of each spectral channel, which is usually described by the half-peak width. During spectral calibration in the laboratory, according to GB/T 30697-2014 “Performance Test Method of Spaceborne Large Field of View multi-spectral Camera [23],” a halogen lamp was used as the light source for spectral calibration, and the light emitted from the calibration light source was concentrated to the center of the monochromator slit through the subsequent shaping light path. First, a monochromator was used to scan the spectrum of each channel, the spectral scanning width was 20 nm, and the scanning wavelength interval of the monochromator was 0.2 nm. As shown in Eq. (1), the spectral response of each channel band of the all-day active hyperspectral lidar system is calculated according to the measurement data, and the spectral response is represented by $R(\lambda )$
As shown in Fig. 2, spectral calibration is mainly divided into two steps—spectral scanning of the hyperspectral lidar system and self-calibration of the halogen lamp—to obtain variables $V{(\lambda )_{LiDAR}}$ and $V{(\lambda )_{Std}}$ in Eq. (1). First, the monochromator is used to complete the scanning of all the detection channels of the hyperspectral lidar system, thus determining the spectral coverage of each channel. As shown in Fig. 2(a), we direct the light emitted by the monochromator incident onto the hyperspectral lidar optical receiving system. After being split by the hyperspectral lidar system itself, the light emitted by the monochromator is coupled to the corresponding detection channel. Then, a standard detector with a known relative spectral response is used to complete the self-calibration of the halogen source to eliminate the influence of energy difference on spectral calibration accuracy caused by halogen light source splitting by the monochromator. The whole spectrum calibration work is completed in a dark room.
3.2 Radiometric calibration
According to GB/T 30697-2014 “Performance Testing Methods for Spaceborne Large Field of View Multispectral Cameras,” the radiation calibration response is the response value of the channel divided by the equivalent radiance. The spectral responsivity of a hyperspectral lidar system is measured by an integrating sphere source. The integrating sphere light source is directly aligned with the receiving window of the hyperspectral lidar system so that the light of the main optical axis of the hyperspectral lidar prototype is emitted from the center of the light plane through the integrating sphere light source. The amplifier gain is set to ensure that the output signal intensity reaches between 50% and 90% of the full scale. The output voltage value of the hyperspectral lidar prototype is repeatedly collected 30 times in each channel. A spectral radiometer is used to measure the output radiance value of the integrating sphere source. The ratio ${L_\textrm{e}}(\lambda )$ of equivalent radiance is calculated by using the spectral calibration results as shown in Eq. (2)
As shown in Fig. 3, the spectroradiometer used in the laboratory relative radiation calibration process is a large-aperture integrating sphere source, and the light that it emits needs to completely fill the receiving field of the hyperspectral lidar system. During radiometric calibration, the equivalent radiance level of the spectral segment corresponding to the hyperspectral lidar needs to be determined. Therefore, the spectral radiance of the reference radiation source needs to be determined first. The reference radiation source is the integrating sphere source in Fig. 3. Then, the electrical signal intensity of each channel of the hyperspectral lidar system is measured. We directly couple the light emitted by the integrating sphere source to the photodetector and store the spectral radiation intensity of the integrating sphere source through a data acquisition card. The calibration coefficients of each channel obtained by radiometric calibration are shown in Eq. (3)
where $V(\lambda )$ is the electrical signal intensity corresponding to each channel during the radiation calibration of hyperspectral lidar.4. Results and discussion
The airborne hyperspectral lidar adopts focal plane spectroscopy technology, which uses the telecentric lens to focus the wide-spectrum echo signal to the focal plane. The echo signals of different wavelengths at the focal plane of the telecentric lens are coupled to the corresponding detection channels by using a fiber optic array [24]. As shown in Fig. 4(a), the optical array consists of 168 optical fibers. The core diameter of a single optical fiber is 0.4 mm, and the interval between adjacent optical fibers is 0.57 mm. In the process of theoretical derivation, with the spectral channel that has a center wavelength of 650 nm taken as an example, the diameter of the monochromatic spot is about 0.47 mm according to the design parameters of the hyperspectral lidar optical receiving system, basically filling the 0.4 mm core diameter of a single fiber. As shown in Fig. 4(b), the red spot represents the spot size of monochromatic light of 650 nm after passing through the hyperspectral lidar optical receiving system. The fiber end face is directly opposite the focal plane of the telecentric focusing lens. Therefore, the single-channel spectral responsivity curve can be obtained through theoretical simulation, as shown in Fig. 4(c).
Notably, the detection performance of the hyperspectral lidar can be improved by using the array composed of 168 optical fibers in the hyperspectral lidar optical receiving system. According to the spectral scattering characteristics of the ground object, the arrangement of the optical fiber array can also be optimized for the central wavelength and bandwidth of each detection channel. In the actual development process of hyperspectral lidar, we coupled three adjacent optical fibers into one detection channel, obtaining a total of 56 detection channels. According to the actual detection needs of ground object targets, the center wavelengths of each channel are determined, and different optical fibers are combined and coupled to the corresponding detectors so that center wavelengths and spectral bandwidths can be selected flexibly for different detection channels.
We finished calibrating the hyperspectral lidar system before the airborne flight experiment in the Key Laboratory of General Optical Calibration and Characterization Technology of the Chinese Academy of Sciences. As shown in Fig. 5, the spectral calibration of the hyperspectral lidar system in the laboratory is mainly divided into two parts: spectral band scanning and data synchronous acquisition and processing. The light emitted from the monochromator enters the hyperspectral lidar system directly through the window of the telescope, and after being split by the grating spectrometer, the monochromatic light of different wavelengths is incident to the corresponding fiber. The wavelength accuracy of the monochromator is 0.2 nm, and the spectral resolution during the band scanning is also set to 0.2 nm. Each detection channel is progressively scanned by controlling the monochromator, and the other end of the fiber is directly coupled to the detector. The acquisition card transmits the collected signal to the computer for real-time storage. Through time domain control, the light output of the monochromator and the acquisition signal of the acquisition card are synchronized.
Spectral calibration is mainly used to determine the central wavelength and bandwidth of each detection channel of the hyperspectral lidar system. The bandwidth obtained by spectral calibration is also the spectral resolution of each channel of the hyperspectral lidar prototype, which is usually described by the semi-peak width. The narrower the spectral resolution of the hyperspectral lidar system, the more easily the information of the ground object can be distinguished and recognized by the hyperspectral lidar system. The spectral data obtained from the calibration were processed, and the spectral curve as shown in Fig. 6 was obtained. Figure 6 shows only the relative spectral responsivity curves of some channels of the hyperspectral lidar system, and the measured spectral responsivity curves are basically consistent with the theoretical results.
The linear dispersion coefficient of this hyperspectral lidar optical receiving system is 5.53 nm/mm, so the spectral bandwidth of a single channel is about 8.5 nm. In the actual spectral calibration process, the exit slit of the monochromator has a certain width; thus, the outgoing spot has a certain spectral width. The defocus and aberration of the optical receiving system also affect the spectral calibration results. Therefore, the actual bandwidth will be slightly larger than the theoretical value.
As shown in Table 1, spectral calibration obtained the central wavelength and bandwidth of each channel of the hyperspectral lidar system. The spectral resolution of most receiving channels in the hyperspectral lidar system is around 9 nm. The theoretical calculation values are basically consistent with the actual test results. In this calibration process, the optical signals of three optical fibers, each of which had a spectral resolution of about 3 nm, were coupled to the target surface of a detector. In the actual airborne flight test, the wavelength can be selected according to the actual situation of ground objects [25], and the bandwidth and central wavelength of each channel can be set flexibly. The spectral resolution of a single 3 nm fiber can accurately classify the spectrum of most ground objects. For the hyperspectral lidar system operating during daytime when the sky background light is strong, the signal-to-noise ratio of each channel can be improved further by coupling three optical fibers to a detection channel, thus ensuring that the hyperspectral lidar system can realize full-time detection.
Table 1. Spectral calibration results of hyperspectral lid
The spectral calibration method of the hyperspectral imaging lidar system in the laboratory was evaluated according to the measurement uncertainty evaluation and expression specification issued by the national unit of measurement [26]. The calibration device and the detector are among the main sources of uncertainty in the calibration process. The uncertainty of the calibration device is mainly due to the wavelength and bandwidth accuracy of the monochromator light source, while the sources of uncertainty in detectors mainly include factors such as responsiveness, stability, and nonlinearity. The uncertainty evaluation of the entire calibration process is shown in Table 2.
Table 2. Accuracy analysis of spectral calibration
The spectral calibration accuracy of the hyperspectral lidar system is calculated according to the general uncertainty evaluation formula, as shown in formula (4) [27].
According to the formula, the error during the calibration process of the hyperspectral lidar system is 0.49 nm. During spectral calibration, the output wavelength of the light source is determined by the monochromator, with a wavelength accuracy of 0.4 nm. The uncertainty evaluation found that the uncertainty introduced by the calibration light source is dominant. Improving the performance of the light source is expected to further reduce the uncertainty of spectral calibration in hyperspectral lidar systems.
On the basis of the spectral calibration results of the hyperspectral lidar system in Table 1, we obtain the curve of the central wavelength coverage of all detection channels of the hyperspectral lidar, as shown in Fig. 7. The results show that the spectral coverage of the hyperspectral lidar system is 400–900 nm, and the bandwidth of all detection channels of the hyperspectral lidar is less than 9 nm. The machining accuracy of adjacent fibers results in small differences in bandwidth between individual detection channels. The bandwidth differences between the various detection channels are also caused by transmissive grating used in the hyperspectral lidar system and by the processing accuracy of each optical component of the optical receiving system.
However, the relative spectral responsivity curves obtained by all detection channels are not consistent with the theory. Comparing Figs. 8 and 6, we find that the peak point of a single detection channel in the ultraviolet band shows an approximate upward trend. This result occurred because the hyperspectral lidar system in the airborne flight test requires each band detector to have a high quantum conversion efficiency so that the supercontinuum laser in all detection channels has an echo signal response. Therefore, a single detector model will have difficulty simultaneously covering the spectral range of 400–900 nm. A PMT with higher quantum efficiency in the short wavelength is used to detect extremely weak signals. In the ultraviolet band in particular, Hamamatsu’s PMT detector H10721-210 was selected. The quantum conversion efficiency of the detector in the ultraviolet band is sharply increased, which is the main reason for the increase in the peak point of the single detection channel of the hyperspectral lidar system in the ultraviolet band.
The bandwidth of some abnormal detection channels in Table 1 is around 5 nm, which may be caused by fiber breakage. As shown in Fig. 9, the relative spectral response curves of these five detection channels were fitted. These detection channels have only two peak inflection points. The two peak inflection points in channel 53 showed an obvious downward trend; the reason for this condition was consistent with the ultraviolet band. The quantum conversion efficiency of the detector also shows a decreasing trend in the near-infrared band.
We analyzed the reasons for bandwidth anomalies by using channel 17 as an example. As shown in Fig. 10, a comparative analysis was conducted by combining the spectral responsivity fitting curves of adjacent channels 16 and 18. The central wavelength of channel 16 is 526.5 nm with a bandwidth of 8.9 nm, that of channel 17 is 534.2 nm with a bandwidth of 6 nm, and that of channel 18 is 544.9 nm with a bandwidth of 8.7 nm. The coupling of three optical fibers in a single detection channel caused the abnormal bandwidth because one of the optical fibers broke. Comparison and analysis of adjacent channels show that the abnormal bandwidth of channel 17 is due to the fracture of the fiber optic wire in the infrared band direction, which is the abnormal fiber optic wire in red in Fig. 10. In the one-dimensional configuration of 168 optical fibers, this abnormal fiber has the 51st serial number.
As shown in Fig. 11, the light emitted by the large-aperture integrating sphere source is directly coupled into the hyperspectral lidar optical receiving system during the laboratory relative radiation calibration. The light emitted by the integrating sphere source needs to not only fill the receiving field of the hyperspectral lidar but also completely cover the incident end face of each fiber. The original radiation intensity is expressed as a dimensionless value; thus, the relative radiation calibration results also normalize the measured value to a dimensionless value. Consistency analysis of relative radiometric calibration parameters can further quantify the accuracy of echo signal intensity. The intensity and pulse width of the incident pulse of waveform data obtained during airborne hyperspectral lidar flight are not exactly the same. The echo waveform data need to be corrected to the same level to consider the initial state of the transmitted pulse and the effect of atmospheric attenuation on the echo waveform.
According to formula (7), the relative radiation calibration coefficient is obtained as shown in Table 3. The radiation calibration constants of each channel of the hyperspectral lidar system are basically on the same order of magnitude, indicating good consistency of the relative radiation calibration constants of the hyperspectral lidar system. Radiometric calibration results can further guide the design and calibration of hyperspectral lidar prototypes. Radiation calibration can eliminate the impact of laser emission energy differences in different bands caused by the special emission mechanism of the supercontinuum spectrum laser on the final airborne flight operation results. The radiation calibration results can also be corrected for the impact of atmospheric transmission on the accuracy of echo signals in airborne flight tests.
Table 3. Relative radiometric calibration results
As shown in Table 4, the radiometric calibration accuracy error mainly consists of the error ${\mu _R}$ introduced by the calibration of the spectral radiometer, the error ${\mu _S}$ introduced by the stability of the integrating sphere light source, and the error ${\mu _A}$ introduced by the repeatability of multiple measurements in the hyperspectral lidar system. The synthetic calibration accuracy error is shown in formula (5):
where the responsivity error of the spectral radiometer and the stability error of the integrating sphere source are calibrated by the National Bureau of Metrology. The final relative radiometric calibration accuracy of the hyperspectral lidar system was 4.2%.Table 4. Precision analysis of radiometric calibration
5. Conclusions
In the face of major development requirements in Earth observation, hyperspectral imaging lidar technology combines the technical advantages of high-precision ranging and hyperspectral imaging of lidar, thus realizing real-time, all-day, full-spectrum monitoring of ground objects with high precision and resolution. This technology has become a hot spot in the development of remote sensing technology at home and abroad. In this paper, a set of hyperspectral imaging lidar systems is designed and developed based on the focal-plane spectroscopy technology of optical fiber array. Through the optical fiber array, the broad spectral echo signal is coupled to the corresponding single-tube detector, solving the problem of hyperspectral imaging under a weak light signal and realizing accurate detection of ground objects. The spectral range of the hyperspectral lidar system covers 400–900 nm, and the spectral resolution of a single fiber is better than 3 nm. A calibration method for hyperspectral imaging lidar system is also proposed. Spectral calibration results show that the theoretical design parameters of the hyperspectral lidar system are consistent with the actual detection performance. The anomalies of spectral bandwidth of individual channels are also analyzed. Radiometric calibration results show that the calibration coefficients of each detection channel of the hyperspectral lidar system have strong consistency.
In the future, airborne hyperspectral lidar will play a more important role in Earth observation applications, relying on the advantages of fast, wide-ranging, and high-precision application of airborne platforms. In particular, airborne hyperspectral lidar will achieve more refined inversion of specific parameters with multi-channel full waveform acquisition capabilities.
Funding
Natural Science Foundation of Anhui Province (2208085UQ01); University Natural Science Research Project of Anhui Province (2023AH050930); Foundation of Key Laboratory of Science and Technology Innovation of Chinese Academy of Sciences (CXJJ-23S028); Hefei Institutes of Physical Science, Chinese Academy of Sciences (HFIPS) Director's Fund (YZJJ202205-CX); National Natural Science Foundation of China (41875033).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Q. Wang, L. B. Bu, L. Tian, et al., “Validation of an airborne high spectral resolution Lidar and its measurement for aerosol optical properties over Qinhuangdao,” Opt. Express 28(17), 24471–24488 (2020). [CrossRef]
2. U. Okyay, J. Telling, C. L. Glennie, et al., “Airborne lidar change detection: An overview of Earth sciences applications,” Earth-Sci. Rev. 198, 102929 (2019). [CrossRef]
3. W. Gong, S. Shi, and L. Du, “Development and prospect of hyperspectral LiDAR for earth observation,” National Remote Sensing Bulletin. 25(1), 501–513 (2021). [CrossRef]
4. C. J. Gleason and J. Im, “Forest biomass estimation from airborne LiDAR data using machine learning approaches,” Remote Sens. Environ. 125, 80–91 (2012). [CrossRef]
5. J. Mascaro, M. Detto, G. P. Asner, et al., “Evaluating uncertainty in mapping forest carbon with airborne LiDAR,” Remote Sens. Environ. 115(12), 3770–3774 (2011). [CrossRef]
6. Z. Y. Hui, Y. J. Hu, and Y. Z. Yevenyo, “Road centerline extraction from airborne LiDAR point cloud based on hierarchical fusion and optimization,” ISPRS J. Photogramm. Remote Sens. 118, 22–36 (2016). [CrossRef]
7. D. P. Yuan, Z. H. Mao, P. Chen, et al., “Remote sensing of seawater optical properties and the subsurface phytoplankton layer in coastal waters using an airborne multiwavelength polarimetric ocean lidar,” Opt. Express 30(16), 29564–29583 (2022). [CrossRef]
8. J. T. Bausell and R. M. Kudela, “Modeling hyperspectral normalized water-leaving radiance in a dynamic coastal ecosystem,” Opt. Express 29(15), 24010–24024 (2021). [CrossRef]
9. S. Shi, W. Gong, and M. S. Liao, “Development and Application of Airborne Hyperspectral LiDAR Imagine Technology,” Guangxue Xuebao 42(12), 1200002 (2022). [CrossRef]
10. J. Ilinca, S. Kaasalainen, T. Malkamäki, et al., “Improved waveform reconstruction and parameter accuracy retrieval for hyperspectral lidar data,” Appl. Opt. 58(35), 9624–9633 (2019). [CrossRef]
11. J. Q. Xu, B. Witschas, P. G. Kabelka, et al., “High-spectral-resolution lidar for measuring tropospheric temperature profiles by means o Rayleigh-Brillouin scattering,” Opt. Lett. 46(13), 3320–3323 (2021). [CrossRef]
12. Z. P. Li, J. T. Ye, X. Huang, et al., “Single-photon imaging over 200 km,” Optica 8(3), 344–349 (2021). [CrossRef]
13. G. G. Taylor, D. Morozov, N. R. Gemmell, et al., “Photon counting LIDAR at 2.3µm wavelength with superconducting nanowires,” Opt. Express 27(26), 38147–38158 (2019). [CrossRef]
14. X. Y. Liu, Z. Q. Yu, S. H. Zheng, et al., “Residual image recovery method based on the dual-camera design of a compressive hyperspectral imaging system,” Opt. Express 30(11), 20100–20116 (2022). [CrossRef]
15. T. F. Wang, D. Liu, Z. S. Xue, et al., “Spectral missing color correction based on an adaptive parameter fitting model,” Opt. Express 31(5), 8561–8574 (2023). [CrossRef]
16. L. Y. Qian, D. C. Wu, X. J. Zhou, et al., “Radiation calibration and ground object information acquisition based on high spectral imaging lidar system,” Guangxue Xuebao 40(11), 1128001 (2020).
17. L. Y. Qian, D. C. Wu, D. Liu, et al., “Infrared detector module for airborne hyperspectral LiDAR: design and demonstration,” Appl. Opt. 62(8), 2161–2167 (2023). [CrossRef]
18. L. Y. Qian, D. C. Wu, X. J. Zhou, et al., “Optical system design for a hyperspectral imaging lidar using supercontinuum laser and its preliminary performance,” Opt. Express 29(11), 17542–17553 (2021). [CrossRef]
19. R. Ceolato, N. Riviere, and L. Hespel, “Reflectances from a supercontinuum laser-based instrument: hyperspectral, polarimetric and angular measurements,” Opt. Express 20(28), 29413–29425 (2012). [CrossRef]
20. B. D. Cook, L. A. Corp, R. F. Nelson, et al., “NASA Goddard's LiDAR, Hyperspectral and Thermal (G-LiHT) Airborne Imager,” Remote Sens. 5(8), 4045–4066 (2013). [CrossRef]
21. S. Kaasalainen, T. Lindroos, and J. Hyyppa, “Toward hyperspectral lidar: Measurement of spectral backscatter intensity with a supercontinuum laser source,” IEEE Geosci. Remote. Sens. Lett. 4(2), 211–215 (2007). [CrossRef]
22. W. Li, Z. Niu, G. Sun, et al., “Deriving backscatter reflective factors from 32-channel full-waveform LiDAR data for the estimation of leaf biochemical contents,” Opt. Express 24(5), 4771–4785 (2016). [CrossRef]
23. GB/T 30697-2014.Test methods of characteristics for spaceborne multispectral camera with wide field of view[S]. China, (2014).
24. L. Y. Qian, D. C. Wu, D. Liu, et al., “Design and demonstration of airborne hyperspectral imaging LiDAR system based on optical fiber array focal plane splitting,” Opt. Commun. 534, 129331 (2023). [CrossRef]
25. S. Song, W. Gong, B. Zhu, et al., “Wavelength selection and spectral discrimination for paddy rice, with laboratory measurements of hyperspectral leaf reflectance,” ISPRS J. Photogramm. Remote Sens. 66(5), 672–682 (2011). [CrossRef]
26. “JJF 1059.1-2012 Evaluation and Expression of Uncertainty in measurement,” [s] (2012).
27. A. Borraccino, M. Courtney, and R. Wagner, “Generic Methodology for Field Calibration of Nacelle-Based Wind Lidars,” Remote Sens. 8(11), 907 (2016). [CrossRef]
