Device and method for detecting diffuse total ratio of wide field of view

Abstract

The invention discloses wide-view-field diffuse total ratio detection equipment and a wide-view-field diffuse total ratio detection method, relates to the field of environmental spectrum detection, and solves the problems of incapability of synchronous observation, incapability of considering wide view field and high precision, complex system, high cost, difficulty in deployment and calibration, poor practicability and incapability of adapting to field (such as gobi) long-term diffuse total ratio observation requirements in the existing diffuse total ratio detection technology. A diffuse total ratio detection device comprises double fisheye lenses, double hyperspectral cameras and a control part, adopts a fixed optical structure, remarkably improves the severe environment resistance of the device, and is suitable for field long-term operation. The method is suitable for detection of diffuse total ratio parameters in the fields of satellite optical load calibration, atmospheric radiation characteristic research, photovoltaic power station irradiation resource evaluation and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of environmental spectrum detection, and also relates to the field of ground spectrum detection. BACKGROUND

[0002] The diffuse-to-global ratio is a key parameter for realizing wide field of view, uninterrupted, high precision, and high cost-effective all-weather spectral data acquisition. However, the current traditional equipment for wide-angle spectral observation generally cannot adapt to the core bottleneck of synchronous observation of the entire sky target, which causes the limitation of the field of view, resulting in the inability to synchronously acquire direct light spectrum and environmental scattering spectrum data, directly restricting the diffuse-to-global ratio detection.

[0003] In the prior art for monitoring the diffuse-to-global ratio, a time-sharing observation method is used to measure the direct sunlight and scattered light data, Specifically, a mechanical rotating mechanism (such as a motor-driven lens rotating around a vertical / horizontal axis) is used to control the acquisition area of the lens, to realize time-sharing coverage of different fields of view, and to obtain full-field spectral data. This method has the following defects: (1) The time-sharing acquisition of multi-view spectral data has a time difference between the direct sunlight and scattered light data, which cannot reflect the diffuse-to-global ratio of the entire sky at the same time, affecting the accuracy of the environmental radiation transfer model inversion. (2) The dynamic environment adaptability is poor, and in the case of rapid movement of cloud layers, the correlation of the time-sharing data is invalid, further exacerbating the calculation error, and cannot support satellite payload calibration and other applications with high requirements for data continuity. (3) The mechanical rotation has cumulative error, and the daily positioning deviation of the field of view is amplified after long-term use, the illumination measurement accuracy is improved, and the precision gradually decreases. The mechanical structure is easy to wear and tear, and has weak resistance to harsh environments, so the reliability of the equipment decreases in extreme conditions (such as gobi sand and temperature fluctuations) in the field.

[0004] In order to solve the defects of the above-mentioned time-sharing acquisition observation method, the prior art uses a multi-detector integration method to realize wide-field acquisition, which effectively solves the problems existing in time-sharing acquisition, especially avoiding the reliability defects of mechanical rotation. This method integrates multiple narrow-field spectrometers arranged at specific angles together, and respectively aims at different regions of the sky, to realize synchronous measurement of direct light and scattered light. However, the integration of multiple spectrometer units in this scheme increases the volume, weight and cost of the entire system. At the same time, to ensure the light path calibration, data synchronization and consistency of the multiple spectrometer units, the system debugging and calibration are relatively complex, the production cost is relatively high, on the other hand, the system also needs to be adjusted and calibrated regularly during operation to ensure the accuracy and reliability of the system, which also leads to high product maintenance cost. This defect makes the system not suitable for long-term observation in actual field environment.

[0005] In summary, the existing diffuse total ratio detection technology has the problems of unable to observe synchronously, unable to balance wide field of view and high precision, complex system, high cost, difficult deployment and calibration, poor practicability and unable to adapt to long-term diffuse total ratio observation requirements in the field (such as Gobi). SUMMARY

[0006] The present application solves the problems of the existing diffuse total ratio detection technology, such as unable to observe synchronously, unable to balance wide field of view and high precision, complex system, high cost, difficult deployment and calibration, poor practicability and unable to adapt to long-term diffuse total ratio observation requirements in the field (such as Gobi). The present application provides the following solutions: Solution one, a diffuse total ratio detection device, comprising a double fisheye lens, a double hyperspectral camera and a control component. The double fisheye lens comprises a visible light fisheye lens and a short-wave infrared fisheye lens. The head field of view angle of the visible light fisheye lens is 180°. The head field of view angle of the short-wave infrared fisheye lens is 160°. The double hyperspectral camera comprises a visible light hyperspectral camera and a short-wave infrared hyperspectral camera. The visible light hyperspectral camera is used for detecting light radiation signals with a wave band of 340-1100 nm. The short-wave infrared hyperspectral camera is used for detecting light radiation signals with a wave band of 900-2550 nm. The visible light fisheye lens is the lens of the visible light hyperspectral camera. The short-wave infrared fisheye lens is the lens of the short-wave infrared hyperspectral camera. The control component is used for outputting control instructions to the double hyperspectral camera to control the double hyperspectral camera to perform single exposure and synchronous collection of hyperspectral image data, receiving the hyperspectral image data sent by the double hyperspectral camera, and calculating the diffuse total ratio based on the hyperspectral image data.

[0007] Further, in an embodiment of the present application, the optical axes of the visible light hyperspectral camera and the short-wave infrared hyperspectral camera are parallel, and the short-wave infrared hyperspectral camera is located lower than the visible light hyperspectral camera.

[0008] Further, in an embodiment of the present application, the control component is internally embedded with a spectrum processing unit, a data fusion unit and a diffuse total ratio calculation unit. The spectrum processing unit is used for performing dark background correction, flat field correction, spectrum correction and spectrum demodulation on the obtained hyperspectral image data to obtain corresponding visible light original data and short-wave infrared original data. a data fusion unit configured to fuse the visible light raw data and the short-wave infrared raw data through a fusion algorithm to obtain a full-band spectral curve; a diffuse-to-total ratio calculation unit configured to obtain a diffuse-to-total ratio based on the full-band spectral curve.

[0009] Further, in an embodiment of the present application, the spectral processing unit comprises a dark background correction subunit, configured to superimpose the 10-minute collected multi-frame hyperspectral image data to obtain an adaptive main dark field every 10 minutes; and configured to subtract the adaptive main dark field from the current time hyperspectral image data to obtain dark background corrected hyperspectral image data.

[0010] Further, in an embodiment of the present application, the spectral processing unit comprises a flat field correction subunit configured to perform flat field correction on the dark background corrected hyperspectral image data according to a vignetting compensation coefficient to obtain flat field corrected hyperspectral image data.

[0011] Further, in an embodiment of the present application, the spectral processing unit comprises a spectral correction subunit configured to correct the flat field corrected hyperspectral image data based on a spectral correction coefficient matrix to obtain spectral corrected hyperspectral image data.

[0012] Further, in an embodiment of the present application, the spectral processing unit comprises a spectral demodulation subunit configured to demodulate the spectral corrected hyperspectral image data through a preset model; the preset model being a mapping relationship model obtained based on a standard light source and different scattering conditions under a plurality of constraint conditions; the plurality of constraint conditions including a basic physical fact constraint that a spectral intensity value is non-negative, a smoothness constraint, a typical form prior constraint of a solar direct spectrum, and a typical form prior constraint of an atmospheric scattering spectrum.

[0013] Further, in an embodiment of the present application, the data fusion unit comprises the following subunits, a matching subunit configured to clock register and pixel align the visible light raw data and the short-wave infrared raw data, and further accurately pair and standardize the visible light raw data and the short-wave infrared raw data for an overlapping band of 900-1100 nm to obtain preliminarily matched visible light raw data and short-wave infrared raw data; a fusion coefficient acquisition subunit configured to take inherent physical characteristics of a solar spectrum and an atmospheric spectrum in the overlapping band as a strong constraint condition, and perform feature fusion on the preliminarily matched visible light raw data and short-wave infrared raw data through an optimal estimation algorithm to obtain a fusion coefficient; The spectrum curve acquisition subunit is configured to acquire a full-band spectrum curve based on the fusion coefficient.

[0014] Further, in an embodiment of the present application, the diffuse-total ratio calculation unit comprises the following subunits: The total irradiance acquisition subunit is configured to perform radiometric integration on the full-band spectrum curve to acquire total solar irradiance; The spectrum curve acquisition subunit is configured to subtract a spectrum curve of a direct sunlight component from the full-band spectrum curve to acquire a spectrum curve of a scattering component; The scattering irradiance acquisition subunit is configured to perform radiometric integration on the spectrum curve of the scattering component to acquire scattering irradiance; The diffuse-total ratio acquisition subunit is configured to acquire a diffuse-total ratio by dividing the scattering irradiance by the total solar irradiance.

[0015] Scheme II, a diffuse-total ratio detection method, which is realized based on an apparatus comprising a double fisheye lens, a double hyperspectral camera and a control component; The double fisheye lens comprises a visible light fisheye lens and a short-wave infrared fisheye lens; The visible light fisheye lens has a head field of view angle of 180°; The short-wave infrared fisheye lens has a head field of view angle of 160°; The double hyperspectral camera comprises a visible light hyperspectral camera and a short-wave infrared hyperspectral camera; The visible light hyperspectral camera is configured to detect light radiation signals in a waveband of 340-1100 nm; The short-wave infrared hyperspectral camera is configured to detect light radiation signals in a waveband of 900-2550 nm; The visible light fisheye lens is a lens of the visible light hyperspectral camera; The short-wave infrared fisheye lens is a lens of the short-wave infrared hyperspectral camera; The control component is embedded with a computer program-implemented diffuse-total ratio detection method, The method comprises the following steps: Step S31, issuing a control instruction, controlling the double hyperspectral camera to set a frequency for data acquisition each time an image is collected, and controlling the double hyperspectral camera to perform single exposure and synchronous collection each time an image is collected; Step S32, collecting hyperspectral image data output by the double hyperspectral camera; Step S33, performing dark background correction, flat field correction, spectrum correction and spectrum demodulation on the hyperspectral image data to acquire corresponding visible light original data and short-wave infrared original data; Step S34, the visible light raw data and short wave infrared raw data are fused through a fusion algorithm to obtain a full waveband spectral curve; Step S35, a diffuse total ratio is obtained based on the full waveband spectral curve.

[0016] The wide field of view diffuse total ratio detection device and method provided by the application is realized based on a double fisheye lens, a double hyperspectral camera and a control component, effectively solves the problems of the existing diffuse total ratio detection technology, such as unable to observe synchronously, unable to consider wide field of view and high precision, complex system, high cost, difficult deployment and calibration, poor practicability and unable to adapt to long-term diffuse total ratio observation requirements in the field (such as the Gobi), and has specific beneficial effects, including: 1, the fisheye lens provided by the application adopts a wide-angle fisheye lens with a head field of view angle of 180° and 160°, can cover the whole sky hemisphere at one time, can synchronously capture the direct sunlight and scattered light in different directions without adjusting the direction of the lens, realizes millisecond-level synchronous acquisition of total irradiance-scattered radiation data, and provides full field of view data basis for accurate and real-time calculation of the diffuse total ratio parameter.

[0017] 2, the diffuse total ratio detection device provided by the application does not have a mechanical transmission mechanism, adopts a fixed optical structure of the fisheye lens equipped with a hyperspectral modulation demodulation imaging component, and combines a matching data processing algorithm, optimizes the diffuse total ratio calculation from time-sharing observation to real-time calculation, completely solves the problem of disconnection of full-sky hemisphere spectral data, and realizes the detection of the diffuse total ratio.

[0018] 3, the diffuse total ratio detection device provided by the application adopts a fixed optical structure, significantly improves the anti-adverse environment capability of the device, and is suitable for long-term operation in the field. Compared with the traditional mechanical method, the application can add a communication module in actual application, which is used for uploading the collected hyperspectral image data and detection results to the cloud, a remote mobile terminal or a remote management center, etc. in real time, realizes intelligent management and application of the diffuse total ratio detection device.

[0019] 4, the diffuse total ratio detection device provided by the application has small volume and is easy to carry, and can be conveniently integrated with existing devices and systems that need to obtain the diffuse total ratio.

[0020] 5, it is verified through experiments that, in terms of cumulative error and precision decay, the wavelength precision of the application is stably controlled within ±0.6nm, the wavelength repeatability is ±0.1nm, the illumination measurement precision decay is less than or equal to 0.3% in a long-term observation (8 months of fault-free period), is much better than the precision decay level of the traditional device, the matching degree of the full field of view detection result and the standard spectrometer is 98.33%, and the system is stable in the range of-40℃ to 70℃ in the prototype test.

[0021] 6、In terms of observation efficiency and data continuity, the present application can synchronously obtain total irradiance and multi-directional scattering radiation data of the full waveband in a single observation, the observation efficiency is improved by more than 10 times compared with traditional equipment; and the synchronous data has no time difference, the calculation error of the diffuse total ratio in a dynamic scene is less than or equal to 3%, which fundamentally eliminates the data mismatching problem caused by time difference in the traditional time-sharing observation method, so that the calculation result of the diffuse total ratio can truly reflect the sky radiation state at the same time (millisecond level), which can accurately support the scenes with high requirements for data continuity such as atmospheric radiation transfer model inversion and satellite load calibration, and is significantly better than the traditional equipment.

[0022] The detection device described in the present application is suitable for detecting the diffuse total ratio parameter in the fields of satellite optical load calibration, atmospheric radiation characteristic research, photovoltaic power station irradiation resource evaluation, etc. BRIEF DESCRIPTION OF DRAWINGS

[0023] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description, taken in conjunction with the accompanying drawings, in which: Figure 1 is the overall schematic diagram of the diffuse total ratio detection device of the first embodiment.

[0024] Figure 2 is the structural assembly schematic diagram of the diffuse total ratio detection device of the first embodiment.

[0025] Figure 3 is the comparison diagram of the inversion spectrum and the original spectrum of the detected solar spectrum of the ninth embodiment, wherein the blue line is the inversion solar spectrum, and the orange line is the original solar spectrum.

[0026] Figure 4 is the comparison diagram of the inversion spectrum and the original spectrum of the detected atmospheric spectrum of the ninth embodiment, wherein the blue line is the original atmospheric spectrum, and the orange line is the inversion atmospheric spectrum.

[0027] Figure 5 is the diffuse total ratio spectrum curve of the ninth embodiment.

[0028] Figure 6 is the flowchart of the diffuse total ratio detection method of the tenth embodiment.

[0029] REFERENCE SIGNS: Upper shell 1; visible light fisheye lens 2; short-wave infrared fisheye lens 3; waterproof rubber ring 4; bottom shell 5; rubber pad 6; high-reflectance glass 7; high-reflectance glass mounting back shell 8. DETAILED DESCRIPTION

[0030] The various embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments described by reference to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0031] Implementation Method 1: The diffuse total ratio detection device described in this implementation method, such as... Figure 1 and Figure 2 As shown, it includes a double fisheye lens, a double hyperspectral camera, and control components; The dual fisheye lens includes a visible light fisheye lens and a short-wave infrared fisheye lens; The visible light fisheye lens has a head field of view of 180°; The head field of view of the short-wave infrared fisheye lens is 160°; The dual hyperspectral cameras include a visible light hyperspectral camera and a short-wave infrared hyperspectral camera; The visible light hyperspectral camera is used to detect light radiation signals in the 340 to 1100 nm band. The short-wave infrared hyperspectral camera is used to detect optical radiation signals in the 900 to 2550 nm band. The visible light fisheye lens is the lens of the visible light hyperspectral camera; The short-wave infrared fisheye lens is the lens of the short-wave infrared hyperspectral camera; The control unit is used to output control commands to the dual hyperspectral cameras to control the dual hyperspectral cameras to perform single exposure and synchronous acquisition of hyperspectral image data; it is also used to receive hyperspectral image data sent by the dual hyperspectral cameras, and to calculate the diffuse-total ratio based on the hyperspectral image data.

[0032] In this embodiment, the dual fisheye lens is preferably mechanically connected and optically aligned with the hyperspectral imaging module behind it via a standard C / CS interface to ensure that the light signal can be completely converged onto the photosensitive surface of the sensor.

[0033] In this embodiment, the dual fisheye lens acquires the light radiation signal to be detected and sends the light radiation signal to the dual hyperspectral camera.

[0034] In this embodiment, the photosensitive surfaces of the dual hyperspectral cameras are precisely located at the focal plane of their respective fisheye lenses, and are preferably electrically connected to the data processing unit of the diffuse-total ratio detection system via a data cable (such as MIPI-CSI).

[0035] In this embodiment, the control component is preferably connected to the spectral detection unit and power supply module of the diffuse-total ratio detection system via a plug-in interface, and is responsible for the workflow scheduling, data processing, and communication control of the entire device. For example, it controls the data acquisition frequency of the dual hyperspectral cameras. This acquisition frequency can be set according to the actual environment and needs, taking into account the amount and capacity of data processing, such as once, ten times, or one hundred times per minute. It also controls the single exposure and synchronous acquisition of the dual hyperspectral cameras, with the synchronization time difference between the two cameras controlled at the millisecond level.

[0036] In the embodiment, the diffuse total ratio detection device adopts optical window in terms of material and protection, and the optical window adopts high-purity quartz glass to ensure high transmittance in full wave band of 340-2550 nm.

[0037] The diffuse total ratio detection device in the embodiment is a diffuse total ratio detection device with wide field of view. The double fisheye lens can cover the full sky hemisphere at one time, and can synchronously capture direct sunlight and scattered light in different directions without adjusting the direction of the lens, so as to realize millisecond-level synchronous acquisition of total irradiance-scattered radiation data. The double hyperspectral cameras realize high-precision and all-weather capture of full-sky spectrum through differential band coverage and collaborative work.

[0038] The visible light camera is responsible for the wave band of 340-1100 nm, and is equipped with an ultra-wide field angle fisheye lens to realize complete coverage of the sky hemisphere field of view, and to ensure measurement of total irradiance and scattered irradiance in the core visible light and near-infrared wave band.

[0039] The short-wave infrared camera focuses on the wave band of 900-2550 nm, and the matching fisheye lens has a slightly smaller view, and the optical axis is coaxial with the visible light lens. The height difference is used to avoid field of view obstruction between the lenses, and the two fields of view are complementary to each other, so as to ensure that the core sky areas measured by the two are coincident, and to realize collaborative observation of spectral wave bands.

[0040] A specific structure of the diffuse total ratio detection device in the embodiment is shown in FIGS. 1-4. Figure 1 2 As shown in FIGS. 1-4, the double fisheye lens and the double hyperspectral cameras are fixed inside a closed box composed of an upper shell 1 and a bottom shell 5. The top of the upper shell 1 is provided with a rectangular window, and the inside of the window is covered with high-reflection glass 7. A rubber pad 6 is fixed between the high-reflection glass 7 and the upper shell 1. The lower part of the high-reflection glass 7 is provided with a high-reflection glass mounting back shell 8, which is a rectangular frame with the same shape as the rectangular window. The high-reflection glass mounting back shell 8 is used to be fixedly connected with the upper shell 1 to fix the high-reflection glass 7. A rubber pad 6 is arranged between the high-reflection glass 7 and the high-reflection glass mounting back shell 8. The rubber pad 6 is rectangular and has the same shape as the rectangular window, and is used to protect and seal the high-reflection glass 7. A waterproof rubber ring 4 is arranged at the connection position of the upper shell 1 and the bottom shell 5 to realize sealing, waterproofing and dustproofing. The visible light fisheye lens 2 of the double fisheye lens is arranged towards the window. The high-reflection glass 7 protects the double fisheye lens. The side wall of the upper shell 1 is provided with a connection port for embedding a connector to realize transmission of electrical signals between the double hyperspectral cameras and the outside.

[0041] ​To ensure the airtightness of the enclosed enclosure consisting of the upper shell 1 and the bottom shell 5, a sealing ring or sealing adhesive is provided between the connector and the upper shell 1.

[0042] The detection device described in this embodiment is small in size, easy to carry, and suitable for field operations. To make it more suitable for long-term field operations, the upper shell 1 and the bottom shell 5 can be made of aluminum alloy material with excellent corrosion resistance, and then subjected to hard anodizing treatment, which can improve surface hardness and wear resistance and corrosion resistance, making it less prone to corrosion and damage when in the field for a long time.

[0043] Implementation Method 2: This implementation method further defines the diffuse total ratio detection device described in Implementation Method 1. In this implementation method, the optical axes of the visible light hyperspectral camera and the short-wave infrared hyperspectral camera are parallel, and the short-wave infrared hyperspectral camera is positioned lower than the visible light hyperspectral camera.

[0044] In this embodiment, the optical axis spacing and lens height difference are determined by the following formula:

[0045] To initially determine the minimum safety parameters. In the formula, α The included angle is the angle between the line connecting the side of the visible light hyperspectral camera closest to the short-wave infrared hyperspectral camera and the optical axis of the short-wave infrared hyperspectral camera, and the optical axis of the short-wave infrared hyperspectral camera, expressed in degrees (d). v The diameter of the visible light fisheye lens is in mm. D is the horizontal distance between the centers of the dual fisheye lenses, in mm; h is the vertical height difference between the optical axes of the dual hyperspectral cameras, in mm; θ It is 80°, which is half of the total field of view of 160°.

[0046] Based on the relationship between the parameters defined by the above formula, the minimum safe optical axis spacing and lens height difference can be preliminarily determined. In actual scenarios, the lens is not an ideal geometric point; its outer shell has a certain thickness, and stray ambient light will enter the lens at a small angle. Furthermore, achieving perfect parallelism during installation is difficult, resulting in slight tilt errors. To fully avoid these interfering factors, in actual installation, 1.5-2 times the minimum safe spacing is taken as the actual safe optical axis spacing and lens height difference.

[0047] The embodiment is a further limitation of the dual hyperspectral camera, and the structure design of the dual hyperspectral camera is illustrated.

[0048] Embodiment three, the embodiment is a further limitation of the diffuse-to-total ratio detection device of embodiment one, in the embodiment, the control component is embedded with a spectral processing unit, a data fusion unit and a diffuse-to-total ratio calculation unit. The spectral processing unit is used for dark background correction, flat field correction, spectral correction and spectral demodulation of the obtained hyperspectral image data, to obtain corresponding visible light original data and short wave infrared original data. The data fusion unit is used for fusing the visible light original data and the short wave infrared original data through a fusion algorithm to obtain a full waveband spectral curve. The diffuse-to-total ratio calculation unit is used for obtaining the diffuse-to-total ratio based on the full waveband spectral curve.

[0049] In the embodiment, the hyperspectral image data contains two-dimensional spatial information and one-dimensional spectral information.

[0050] The embodiment is a further limitation of the diffuse-to-total ratio detection device, and the diffuse-to-total ratio detection system is illustrated, which is composed of a spectral processing unit, a data fusion unit and a diffuse-to-total ratio calculation unit, to realize full field of view one-time imaging and spectral calculation reconstruction.

[0051] Embodiment four, the embodiment is a further limitation of the diffuse-to-total ratio detection device of embodiment three, in the embodiment, the spectral processing unit includes a dark background correction subunit, is used for superimposing the 10-minute collected multiple frames of hyperspectral image data every 10 minutes to obtain an adaptive main dark field; and is also used for subtracting the adaptive main dark field from the hyperspectral image data at the current time to obtain the dark background corrected hyperspectral image data.

[0052] In the embodiment, the non-negative spectral intensity condition and other physical prior constraints are preferably used to reduce random noise in the dark background correction process.

[0053] The embodiment is a further limitation of the total diffuse ratio detection device, and the dark background correction is illustrated. In the start of the total diffuse ratio detection device and every 10 minutes period, the spectral processing unit synchronously collects multiple frames of hyperspectral image data as dark field data with ultra-low exposure setting and fisheye lens; and the adaptive main dark field is generated by superimposing multiple frames. In real-time observation, the spectral processing unit dynamically strips the adaptive main dark field from the original data stream. This method effectively suppresses system thermal noise and dark current, improves the overall signal-to-noise ratio of the system, effectively avoids correction lag caused by environmental fluctuations, and ensures high purity of radiation spectrum data in rapidly changing scenes.

[0054] Embodiment five, the embodiment is a further limitation of the total diffuse ratio detection device of embodiment three. In the embodiment, the spectral processing unit includes a flat field correction subunit for flat field correction of the dark background corrected hyperspectral image data according to the vignetting compensation coefficient, to obtain the flat field corrected hyperspectral image data.

[0055] In the embodiment, the vignetting compensation coefficient is obtained in a laboratory environment before the detection device is put into the field, which is a kind of calibration process of the device. In the process, the reference standard value is obtained by using the total diffuse ratio detection device to shoot a uniform white board under the condition of uniform illumination in the laboratory; based on the reference standard value, the vignetting compensation coefficient is obtained through the field of view angle adaptive mechanism and the response difference of all pixel points in the field of view.

[0056] Under the condition of uniform illumination in the laboratory, the uniform white board is shot by using the total diffuse ratio detection device, and the vignetting compensation coefficient corresponding to all pixel points is calculated in real time according to the response of each pixel and the reference standard value (here, it is the laboratory calibration link, so the vignetting compensation coefficient can be calculated pixel by pixel. Because it has not been put into application at this time, it does not need to be concerned about time and algorithmic constraints. In real application, the system only needs to directly call the coefficient for calculation without long time and high algorithmic power). The flat field corrected hyperspectral image data is obtained by correcting the dark background corrected hyperspectral image data; The vignetting compensation coefficient is dynamically adjusted according to the response difference of different regions in the field of view (here, it refers to all pixel points, because each pixel is an independent field of view angle) through the field of view angle adaptive mechanism.

[0057] The vignetting compensation coefficient obtained by the above method is applied to flat field correction, which can realize dynamic adjustment according to the response difference of different regions in the field of view through the field of view angle adaptive mechanism, and finally ensure the consistency and accuracy of full field of view irradiance measurement.

[0058] The flat field correction defined in the embodiment is a challenge facing lens vignetting and sensor response unevenness, and the correction in the flat field correction process is performed by a vignetting compensation coefficient. The vignetting correction is essentially a correction of the brightness distortion of the received image, because the image brightness value directly affects the interpretation of spectral information, and the correction process can overcome the distortion defects in the large field of view image obtained by the fisheye lens. The embodiment realizes the integrated correction of brightness distortion and radiation response by combining the high-throughput light signal capture capability of the fisheye lens of the total diffuse ratio detection device with the spatial and spectral information provided by the hyperspectral modulation-demodulation type imaging chip of the dual hyperspectral camera. The spatial and spectral information provided by the hyperspectral modulation-demodulation type imaging chip is used to realize the integrated correction of geometric distortion and radiation response.

[0059] In the embodiment, the correction process of the image does not include the correction of the image texture and distortion, thereby reducing the data processing amount and improving the data processing speed. At the same time, because the image texture and distortion do not excessively affect the spectral information, the correction of the image texture and distortion will not affect the subsequent calculation results of the total diffuse ratio.

[0060] Embodiment six, the embodiment is a further limitation of the total diffuse ratio detection device of embodiment three. In the embodiment, the spectral correction is that the spectral processing unit includes a spectral correction subunit, which is used to correct the hyperspectral image data after the flat field correction based on a spectral correction coefficient matrix to obtain hyperspectral image data after spectral correction.

[0061] In the embodiment, the spectral correction coefficient matrix is specifically measured by using a standard light source with absolute spectral radiance brightness to measure the original response of the dual hyperspectral camera under the entire working waveband and the full field of view angle; and the original response is compared with the prior true spectral curve of the standard light source to obtain the spectral correction coefficient matrix.

[0062] The embodiment is a further limitation of the total diffuse ratio detection device, and the spectral correction is exemplified. In the field measurement, the data after the first two steps of correction is operated by using the spectral correction coefficient matrix of the embodiment, so as to compensate for the spectral response deviation caused by factors such as lens dispersion and filter transmittance, and finally to obtain high-fidelity hyperspectral image data from the mixed signal.

[0063] Embodiment seven, the embodiment is a further limitation of the total diffuse ratio detection device of embodiment three. In the embodiment, the spectral processing unit includes a spectral demodulation subunit, which is used to demodulate the hyperspectral image data after the spectral correction by using a preset model. The preset model is a mapping relationship model obtained under multiple constraint conditions based on a standard light source and different scattering conditions. The various constraints include the fundamental physical fact constraint that the spectral intensity value is non-negative, the smoothness constraint, the a priori constraint of the typical form of the direct solar spectrum, and the a priori constraint of the typical form of the atmospheric scattering spectrum.

[0064] This embodiment further defines the diffuse-to-total ratio detection device and provides an example of spectral demodulation. Based on the principles of physical priors and computational spectroscopy, this embodiment achieves accurate extraction of the effective signal by establishing and solving physical prior constraints. During real-time spectral demodulation in the field, no complex iterations are required; rapid calculations can be performed, successfully removing atmospheric scattering noise from the mixed signal and ultimately outputting high signal-to-noise ratio hyperspectral image data synchronously. This process is computationally efficient and possesses significant advantages in terms of high precision, strong robustness, and high efficiency.

[0065] Multiple constraints include the fundamental physical fact constraint that the spectral intensity value is non-negative, which serves as a hard condition to exclude unreasonable results; smoothness constraints reduce the violent fluctuations of the spectral curve, making it conform to natural laws; and a priori constraints on the typical forms of direct solar spectrum and atmospheric scattering spectrum, which guide the algorithm to converge to the physically most probable result.

[0066] Implementation Method Eight: This implementation method further defines the diffuse total ratio detection device described in Implementation Method Three. In this implementation method, the data fusion unit includes the following sub-units: The matching subunit is used to perform clock registration and pixel alignment on the raw visible light data and raw short-wave infrared data, and to further perform precise matching and standardization on the overlapping band of 900-1100nm to obtain the raw visible light data and raw short-wave infrared data after preliminary matching. The fusion coefficient acquisition subunit is used to take the inherent physical characteristics of the solar spectrum and atmospheric spectrum in the overlapping band as strong constraints, and perform feature fusion on the raw visible light data and raw shortwave infrared data after the initial matching through the optimal estimation algorithm to obtain the fusion coefficient. The spectral curve acquisition subunit is used to obtain a full-band spectral curve based on the fusion coefficient.

[0067] The embodiment is a further limitation of the diffuse global ratio detection device, and the fusion algorithm is exemplified. The key link of the embodiment is the data processing unit, and the core purpose is to solve the data barrier caused by the hardware difference between the two hyperspectral cameras of visible light and short wave infrared, and to generate a continuous, consistent and physically meaningful spectral curve in the full waveband range of 340-2550 nm. The matching subunit ensures that the data from different frequency bands of the two different cameras is comparable in space, time and radiation scale. The fusion coefficient acquisition subunit can effectively compensate for the subtle differences in spectral response function and calibration coefficient between the two independent optical systems, eliminate the spectral jump at the joint, and achieve seamless splicing. The spectral curve acquisition subunit uses the fused fusion coefficient as a reliable reference to extrapolate in the short wave and long wave directions respectively, and generates a complete and smooth full waveband spectral curve. This curve has high resolution in the visible light band and specific absorption information in the short wave infrared band, providing a reliable data source for subsequent irradiance integral calculation.

[0068] Embodiment nine, the embodiment is a further limitation of the diffuse global ratio detection device of embodiment three. In the embodiment, the diffuse global ratio calculation unit comprises the following subunits: The total irradiance acquisition subunit performs radiometric integration on the full waveband spectral curve to obtain the total solar irradiance; The spectral curve acquisition subunit subtracts the spectral curve of the direct sunlight component from the full waveband spectral curve to obtain the spectral curve of the scattering component; The scattering irradiance acquisition subunit performs radiometric integration on the spectral curve of the scattering component to obtain the scattering irradiance; The diffuse global ratio acquisition subunit obtains the diffuse global ratio by dividing the scattering irradiance by the total solar irradiance.

[0069] After field deployment and testing, the test effect is as shown in Figures 3 to 5 It can be seen from Figure 3 The inversion spectrum and the original spectrum of the detected solar spectrum, Figure 4 The inversion spectrum and the original spectrum of the detected atmospheric spectrum, the curves are basically consistent, and the detection effect is remarkable. Figure 5 The diffuse global ratio spectral curve.

[0070] The embodiment is a further limitation of the diffuse-global ratio detection device, and illustrates the process of obtaining the diffuse-global ratio. The radiometric integral is an integral of the radiation flux arriving in all directions within the full field of view and is again based on prior knowledge to determine and strip noise. The physical essence is to measure the total solar irradiance received on the horizontal plane. On the basis of accurately obtaining the diffuse irradiance and the total solar irradiance, the diffuse-global ratio obtained intuitively represents the proportion of scattered radiation in the total radiation, and is a core index for analyzing atmospheric conditions, evaluating photovoltaic system performance, and other applications.

[0071] The system internally adopts a layered, zoned, and modular wiring strategy. All cables are regularly laid and fixed with tape through special cable slots, effectively avoiding cable loosening, wear, or mutual electromagnetic interference caused by vibration. In the selection and processing of connectors, the principles of high reliability and environmental adaptability are strictly followed, and industrial-grade connectors with locking structures are selected. The surface of the contact piece is gold-plated to improve corrosion resistance and electrical stability. After completing the plug-in, a specific type of elastic sealant is applied to the interface of each connector for point fixing.

[0072] Embodiment ten, the method for detecting the diffuse-global ratio, the method flow chart is as shown in Figure 6 The method is realized based on the following device, the device includes a double fish-eye lens, a double hyperspectral camera and a control component; The double fish-eye lens includes a visible light fish-eye lens and a short-wave infrared fish-eye lens; The head visual field angle of the visible light fish-eye lens is 180°; The head visual field angle of the short-wave infrared fish-eye lens is 160°; The double hyperspectral camera includes a visible light hyperspectral camera and a short-wave infrared hyperspectral camera; The visible light hyperspectral camera is used for detecting light radiation signals with a wave band of 340-1100 nm; The short-wave infrared hyperspectral camera is used for detecting light radiation signals with a wave band of 900-2550 nm; The visible light fish-eye lens is a lens of the visible light hyperspectral camera; The short-wave infrared fish-eye lens is a lens of the short-wave infrared hyperspectral camera; The control component internally embeds a computer program to realize the method for detecting the diffuse-global ratio, The method includes the following steps: Step S31, a control instruction is sent to control the double hyperspectral camera to collect data at a set frequency, and the double hyperspectral camera is controlled to perform single exposure and synchronous collection each time an image is collected; Step S32, hyperspectral image data output by the double hyperspectral camera is collected; Step S33, dark background correction, flat field correction, spectral correction and spectral demodulation are performed on the hyperspectral image data to obtain corresponding visible light original data and short wave infrared original data; Step S34, the visible light original data and the short wave infrared original data are fused through a fusion algorithm to obtain a full waveband spectral curve; Step S35, a diffuse total ratio is obtained based on the full waveband spectral curve.

Claims

1. A wide field-of-view, total-to-rough ratio detection device, characterized in that, The double fisheye lens, the double hyperspectral camera and the control component are comprised; The double fisheye lens comprises a visible light fisheye lens and a short-wave infrared fisheye lens; The head visual field angle of the visible light fisheye lens is 180°; The head visual field angle of the short-wave infrared fisheye lens is 160°; The double hyperspectral camera comprises a visible light hyperspectral camera and a short-wave infrared hyperspectral camera; The visible light hyperspectral camera is used for detecting light radiation signals in the wave band of 340-1100 nm; The short-wave infrared hyperspectral camera is used for detecting light radiation signals in the wave band of 900-2550 nm; The visible light fisheye lens is the lens of the visible light hyperspectral camera; The short-wave infrared fisheye lens is the lens of the short-wave infrared hyperspectral camera; The control component is used for outputting control instructions to the double hyperspectral camera, controlling single-exposure and synchronous acquisition of hyperspectral image data by the double hyperspectral camera, receiving hyperspectral image data sent by the double hyperspectral camera, and calculating albedo based on the hyperspectral image data.

2. The albedo detection device according to claim 1, wherein the optical axes of the visible light hyperspectral camera and the short-wave infrared hyperspectral camera are parallel, and the short-wave infrared hyperspectral camera is located lower than the visible light hyperspectral camera. The control component is internally embedded with a spectral processing unit, a data fusion unit and an albedo calculation unit; 3. The total-to-differential probe device of claim 1, wherein, The spectral processing unit is used for performing dark background correction, flat field correction, spectral correction and spectral demodulation on the obtained hyperspectral image data to obtain corresponding visible light original data and short-wave infrared original data; The data fusion unit is used for fusing the visible light original data and the short-wave infrared original data through a fusion algorithm to obtain a full-wave band spectral curve; The albedo calculation unit is used for obtaining albedo based on the full-wave band spectral curve. The spectral processing unit comprises a dark background correction subunit, 4. The total-to-differential probe device of claim 3, wherein, The dark background correction subunit is used for superimposing a plurality of frames of hyperspectral image data collected in every 10 minutes to obtain an adaptive main dark field; The dark background correction subunit is also used for subtracting the adaptive main dark field from the hyperspectral image data at the current time to obtain dark background corrected hyperspectral image data. The spectral processing unit comprises a flat field correction subunit, which is used for performing flat field correction on the dark background corrected hyperspectral image data according to a vignetting compensation coefficient to obtain flat field corrected hyperspectral image data.

5. The total-to-differential probe device of claim 3, wherein, The spectral processing unit comprises a spectral correction subunit, which is used for correcting the flat field corrected hyperspectral image data based on a spectral correction coefficient matrix to obtain spectral corrected hyperspectral image data.

6. The total-to-differential probe device of claim 3, wherein, The spectral processing unit comprises a spectral demodulation subunit, which is used for demodulating the spectral corrected hyperspectral image data through a preset model; 7. The total-to-differential probe device of claim 3, wherein, The preset model is a mapping relationship model obtained based on standard light sources and different scattering conditions under a plurality of constraint conditions; The plurality of constraint conditions comprise a basic physical fact constraint that spectral intensity values are non-negative, a smoothness constraint, a typical form prior constraint of a solar direct spectrum and a typical form prior constraint of an atmospheric scattering spectrum. ​ 8. The total-to-differential probe device of claim 3, wherein, The data fusion unit comprises the following sub-units, a matching sub-unit, configured to clock register and pixel align the visible light raw data and the short-wave infrared raw data, and further accurately pair and standardize the visible light raw data and the short-wave infrared raw data for an overlapping waveband of 900-1100 nm to obtain the visible light raw data and the short-wave infrared raw data after preliminary matching; a fusion coefficient acquisition sub-unit, configured to take inherent physical features of the sunlight spectrum and the atmospheric spectrum in the overlapping waveband as a strong constraint condition, and perform feature fusion on the visible light raw data and the short-wave infrared raw data after preliminary matching by an optimal estimation algorithm to obtain a fusion coefficient; a spectrum curve acquisition sub-unit, configured to obtain a full-waveband spectrum curve based on the fusion coefficient.

9. The total-to-differential probe device of claim 3, wherein, The diffuse-to-global ratio calculation unit comprises the following sub-units: a total irradiance acquisition sub-unit, configured to perform radiometric integration on the full-waveband spectrum curve to obtain total solar irradiance; a spectrum curve acquisition sub-unit, configured to subtract a spectrum curve of a direct sunlight component from the full-waveband spectrum curve to obtain a spectrum curve of a scattering component; a scattering irradiance acquisition sub-unit, configured to perform radiometric integration on the spectrum curve of the scattering component to obtain scattering irradiance; a diffuse-to-global ratio acquisition sub-unit, configured to obtain a diffuse-to-global ratio by dividing the scattering irradiance by the total solar irradiance.

10. A wide field of view diffuse total ratio detection method, characterized in that, The method is implemented based on an apparatus comprising a double fisheye lens, a double hyperspectral camera and a control component; the double fisheye lens comprises a visible light fisheye lens and a short-wave infrared fisheye lens; a head field of view angle of the visible light fisheye lens is 180°; a head field of view angle of the short-wave infrared fisheye lens is 160°; the double hyperspectral camera comprises a visible light hyperspectral camera and a short-wave infrared hyperspectral camera; the visible light hyperspectral camera is configured to detect light radiation signals in a waveband of 340-1100 nm; the short-wave infrared hyperspectral camera is configured to detect light radiation signals in a waveband of 900-2550 nm; the visible light fisheye lens is a lens of the visible light hyperspectral camera; the short-wave infrared fisheye lens is a lens of the short-wave infrared hyperspectral camera; the control component internally embeds a diffuse-to-global ratio detection method implemented by a computer program, the method comprises the following steps: step S31, issuing a control instruction to control the double hyperspectral camera to collect data at a set frequency, and controlling the double hyperspectral camera to perform single exposure and synchronous collection each time an image is collected; step S32, collecting hyperspectral image data output by the double hyperspectral camera; step S33, performing dark background correction, flat field correction, spectrum correction and spectrum demodulation on the hyperspectral image data to obtain corresponding visible light raw data and short-wave infrared raw data; step S34, fusing the visible light raw data and the short-wave infrared raw data by a fusion algorithm to obtain a full-waveband spectrum curve; step S35, obtaining a diffuse-to-global ratio based on the full-waveband spectrum curve.

Images