Method and apparatus for producing a calibrated spectral imaging device

By using previously measured quantum efficiency and spectral transmittance information, combined with multiple stable spectral peaks, rapid calibration of spectral imaging devices is achieved, solving the long-term problem in the prior art and is suitable for large-scale production.

CN119072613APending Publication Date: 2024-12-03TEKNOLOGIAN TUTKIMUSKESKUS VTT OY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380035412.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-04-22
Filing Date
2023-04-14
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

During the calibration process, the existing spectral imaging device needs to scan a monochromator to change the wavelength of the calibration light, which takes a long time, and the spectral transmittance measurement of the Fabry-Perot interferometer takes a long time, affecting production efficiency. .

Method used

By using previously measured spectral quantum efficiency of detector pixels and spectral transmittance information of Fabry-Perot interferometers, calibrated light with multiple stable spectral peaks, combined with matching of simulated profiles with measured profiles, iteratively modified Calibrate data to speed up the calibration process and avoid using a scanning monochromator.

Benefits of technology

The calibration time of the spectral imaging device on the assembly line is significantly reduced, suitable for large-scale production, improves production efficiency, and avoids time-consuming measurements of the combined spectral response function of the Fabry-Perot interferometer and image sensor.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119072613A_ABST
    Figure CN119072613A_ABST
Patent Text Reader

Abstract

A method of calibrating a spectral imaging device (500) comprises:-providing first calibration light (LB1CAL1) having a first calibration spectrum (IMPBF) having a plurality of stable spectral peaks at a known wavelength ([lambda] 01, [lambda] 02, [lambda] 03),-coupling the first calibration light (LB1CAL1) to the spectral imaging device (500), obtaining a first measurement profile (MSPEC1) of the first calibration light (LB1CAL1) by recording the first detector pixel signal (SR) during scanning of a control parameter (Vd) of a Fabry-Perot interferometer (FPI1) of the spectral imaging device (500), obtaining a second measurement profile (MSPEC1) of the first calibration light (LB1CAL1) by recording the second detector pixel signal (SG) during scanning of the control parameter (Vd), and obtaining a second measurement profile (MSPEC1) of the second calibration light (LB1CAL1) by recording the second detector pixel signal (SG) during scanning of the control parameter (Vd). Determining a first simulated profile (SIMSPEC1) from the first calibration spectrum (IMPBF) by using a previously measured spectral quantum efficiency (QR) of the first detector pixel (P0), by using a previously measured spectral transmittance function (TFP, PRE) and by using the first calibration data (fk, gk, hk); -determining a second simulated profile from the first calibration spectrum (IMPBF) by using a previously measured spectral quantum efficiency (QR) of the second detector pixel (P1), by using a previously measured spectral transmittance function (TFP, PRE) and by using the first calibration data (fk, gk, hk), and-modifying the first calibration data (fk, gk, hk) to determine a second simulated profile from the first calibration spectrum (IMPBF), by using a previously measured spectral quantum efficiency (QR) of the second detector pixel (P1), by using a previously measured spectral transmittance function (TFP, PRE) and by using the first calibration data (fk, gk, hk). Until the simulated profile (SIMSPEC1) matches the corresponding measured profile (MSPEC1) according to one or more matching criteria.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] Some embodiments relate to a method of calibrating a spectral imaging device. Background Art

[0002] It is known to use a monochromator to provide narrowband light for calibrating a spectral imaging device including a Fabry-Perot interferometer. The wavelength of the calibration light can be changed by scanning the monochromator. A known calibration method comprises changing the wavelength of the narrowband calibration light and recording the detector signal at several different reflector gap values ​​of the Fabry-Perot interferometer of the spectral imaging device. Changing the wavelength of the narrowband light by scanning the monochromator can be time consuming. Summary of the invention

[0003] The object of the present invention is to provide a method for calibrating a spectral imaging device. The object of the present invention is to provide a method for producing a calibrated spectral imaging device. The object of the present invention is to provide an apparatus for calibrating a spectral imaging device.

[0004] According to one aspect, a method according to claim 1 is provided.

[0005] Further aspects are defined in the further claims.

[0006] The scope of protection sought by various embodiments of the present invention is set forth by the independent claims. Embodiments described in this specification that do not belong to the scope of the independent claims (if any) are to be construed as examples useful for understanding various embodiments of the present invention.

[0007] The assembled spectral imaging device includes a scanning Fabry-Perot interferometer and an image sensor. In order to perform accurate measurements, the spectral imaging device may need to be calibrated. The spectral transmittance of the Fabry-Perot interferometer is a function of two variables. The spectral transmittance of the Fabry-Perot interferometer is a function of the wavelength (λ), and the spectral transmittance of the Fabry-Perot interferometer is also a control parameter (V d ), which control parameter indicates the distance between the mirrors of the Fabry-Perot interferometer. Measuring the complete spectral transmittance function of the Fabry-Perot interferometer can be time consuming. By using previously measured information about the spectral quantum efficiency of the detector pixels, and by using previously measured information about the spectral transmittance of the Fabry-Perot interferometer, the calibration procedure can be significantly accelerated.

[0008] The assembled spectral imaging device can be calibrated at a calibration station (e.g., an assembly line). The calibration method may include:

[0009] - providing a first calibration spectrum (I MPBF) of the first calibration light (LB1CAL1), the first calibration spectrum at a known wavelength (λ 01 ,λ 02 ,λ 03 ) has multiple stable spectral peaks,

[0010] - coupling the first calibration light (LB1CAL1) to the spectral imaging device (500),

[0011] - In the control parameters (V d ) is scanned by recording the first detector pixel signal (S R ) to obtain a first measurement profile (MSPEC1) of a first calibration light (LB1CAL1),

[0012] -In the control parameters (V d ) is scanned by recording the pixel signal (S G ) obtain a second measurement profile (profile, distribution curve, distribution shape) of the first calibration light (LB1CAL1),

[0013] - by using the previously measured spectral quantum efficiency (Q R ), by using the previously measured spectral transmittance function (T FP,PRE ) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) determine a first simulation profile (SIMSPEC1),

[0014] - by using the previously measured spectral quantum efficiency (Q R ), by using the previously measured spectral transmittance function (T FP,PRE ) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) determine a second simulation profile, and

[0015] - For the first calibration data (f k , g k ,h k ) is modified until the simulated profile (SIMSPEC1) matches the corresponding measured profile (MSPEC1) according to one or more matching criteria.

[0016] The spectral imaging device can be assembled at an assembly line. The image sensor can be combined with a Fabry-Perot interferometer at an assembly line. The method allows the spectral imaging device to be calibrated by using previously obtained data on the quantum efficiency of the multispectral image sensor and by using previously obtained data on the spectral transmittance of the Fabry-Perot interferometer. The method allows calibration in mass production without providing monochromatic calibration light.

[0017] The method includes obtaining a prior spectral transmittance of a Fabry-Perot interferometer. The spectral transmittance of the Fabry-Perot interferometer may be measured at a separate facility before assembling the spectral imaging device. The measured spectral transmittance may then be transmitted to a calibration station of the assembly line (e.g., via the Internet).

[0018] The previously measured quantum efficiency and the previously measured spectral transmittance can be forwarded to, for example, a calibration station, where a final calibration of the assembled spectral imaging device can be performed. The assembly line can include a calibration station for performing calibration of the spectral imaging device. The calibration station provides a first calibration light having a plurality of stable spectral peaks for determining the spectral scale of the spectral imaging device. For example, the first calibration light can be formed by using a multi-passband filter.

[0019] By coupling a first calibration light into the spectral imaging device and by recording the detector signal while scanning the Fabry-Perot interferometer, a set of measurement profiles can be obtained.

[0020] A set of simulated profiles can be determined based on the known spectrum of the first calibration light by using previously measured quantum efficiencies, by using previously measured spectral transmittances of a Fabry-Perot interferometer, and by using calibration data. The previously measured quantum efficiencies and spectral transmittances can be considered to be somewhat representative of the assembled spectral imaging device, but not completely representative. The calibration data can be iteratively modified until the simulated profiles substantially match the measured profiles. The calibration data can be iteratively modified until the simulated profiles correspond to the measured profiles to a sufficient degree. The modified calibration data can then be used to provide accurate measurement results.

[0021] By using previously measured spectral data and stable spectral peaks of the first calibration light, the use of a scanning monochromator to measure the spectral response function of the Fabry-Perot interferometer and image sensor combination can be avoided.

[0022] In an embodiment, the method may be performed without changing the spectral position of the spectral peak of the generated calibration light. The method may significantly reduce the time required to calibrate the spectral imaging device at an assembly line. Forming a simulated profile based on previously measured spectral data may significantly reduce the time required to calibrate the spectral imaging device at an assembly line. The method may be applicable to a production line. The method may be suitable for large-scale production of multiple spectral imaging devices.

[0023] In an embodiment, additional profiles may be optionally measured by coupling a second calibration light to the spectral imaging device. The second calibration light may have a broadband spectrum. Additional measurement profiles may be used to further improve calibration accuracy.

[0024] A manufacturing batch of image sensors may have substantially similar quantum efficiencies for a given color channel. Therefore, it may be sufficient to obtain a previously measured quantum efficiency based on only a small number of image sensors. It may be sufficient to measure the quantum efficiency of, for example, less than 5% of the image sensors in a manufacturing batch. The measured quantum efficiency may then be used to represent the entire batch.

[0025] For example, the spectral quantum efficiency may be measured for several samples of image sensors from a manufacturing batch, and the same measured spectral quantum efficiency may be used for each image sensor of the manufacturing batch.

[0026] In an embodiment, the quantum efficiency may also be measured for each image sensor of a manufacturing batch. In an embodiment, the spectral transmittance may also be measured for each interferometer of a manufacturing batch. By using previously measured information about the quantum efficiency and spectral transmittance, calibration of the assembled spectral imaging device may still be performed quickly.

[0027] Interferometers of a manufacturing batch may exhibit substantially similar spectral characteristics. In particular, when produced by micro-electromechanical manufacturing methods (MEMS), interferometers of a manufacturing batch may exhibit substantially similar spectral characteristics. Therefore, it may be sufficient to obtain a previously measured spectral transmittance based on only a small number of interferometers. It may be sufficient to measure the spectral transmittance of, for example, less than 5% of the interferometers in a manufacturing batch. The measured spectral transmittance may then be used to represent the entire batch.

[0028] For example, the spectral transmittance may be measured for several samples of Fabry-Perot interferometers from a manufacturing batch, and the same measured spectral transmittance may be used for each interferometer of the manufacturing batch.

[0029] In an embodiment, the previously measured spectral transmittance may be a measured spectral transmittance of a second Fabry-Perot interferometer installed in the calibrated spectral imaging device, which is different from the first Fabry-Perot interferometer.

[0030] In an embodiment, the previously measured spectral transmittance may be the measured spectral transmittance of the same Fabry-Perot interferometer installed in the spectral imaging device. By using previously measured information about quantum efficiency and spectral transmittance, calibration of the assembled spectral imaging device can still be performed quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In the following examples, several variations will be described in more detail with reference to the accompanying drawings, in which

[0032] Figure 1a By way of example a spectral imaging device is shown,

[0033] Figure 1b By way of example the spectrum of the input light and the spectral transmittance of the Fabry-Perot interferometer are shown,

[0034] Figure 2a By way of example, changing the spectral position of the spectral passband of a Fabry-Perot interferometer is shown,

[0035] Figure 2b By way of example the spectral transmittance of a Fabry-Perot interferometer is shown as a function of wavelength and as a function of the control signal,

[0036] Figure 3a The formation of an optical image on an image sensor is shown by way of example in an axial view,

[0037] Figure 3b By way of example the spectral sensitivity of a detector pixel of an image sensor is shown,

[0038] Figure 3c The spectral sensitivities of different color channels of a spectral imaging device are shown by way of example,

[0039] Figure 3d By way of example the spectral transmittance of cut-off filters which define the spectral measurement range of a spectral imaging device are shown,

[0040] Figure 4a By way of example, a device for providing a first calibration light for a spectral imaging device is shown.

[0041] Figure 4b By way of example, a multi-passband filter is shown in cross-sectional side view,

[0042] Figure 4c By way of example, it is shown that the first calibration light is formed by filtering the light of the broadband light source using a multi-passband filter.

[0043] Figure 5a By way of example, a measurement signal profile obtained by measuring the first calibration light using a spectral imaging device is shown.

[0044] Figure 5b By way of example, a simulated signal profile calculated from the spectrum of the first calibration light is shown using a simulated spectral response of a combination of a Fabry-Perot interferometer and detector pixels,

[0045] Figure 6a By way of example, an initial measured signal profile and an initial simulated signal profile are shown,

[0046] Figure 6b By way of example the modified measured signal profile and the initial simulated signal profile are shown,

[0047] Figure 7a By way of example, a device for providing a second calibration light for a spectral imaging device is shown.

[0048] Figure 7b shows the spectral intensity distribution of the second calibration light,

[0049] Figure 8a By way of example, the method steps for producing and calibrating a spectral imaging device are shown.

[0050] Figure 8b By way of example, the method steps of determining a calibration data matrix based on the modified spectral transmittance are shown.

[0051] Figure 8c By way of example, the method steps of determining calibrated spectral intensity values ​​from the measurement signal by using the elements of the calibration data matrix are shown.

[0052] Figure 9a By way of example, a measured spectrum of a first calibration light is shown, wherein the measured spectrum is calculated by using calibration data available after a first iteration cycle,

[0053] Figure 9b By way of example, a measured spectrum of a first calibration light is shown, wherein the measured spectrum is calculated by using calibration data available after a fourth iteration cycle,

[0054] Fig.9c By way of example, a measured spectrum of the first calibration light is shown, wherein the measured spectrum is calculated by using calibration data available after an eleventh iteration cycle,

[0055] Figure 9d By way of example, the real spectrum of the first calibration light is shown, and

[0056] Fig.10 Measuring the initial spectral transmittance of a Fabry-Perot interferometer is shown by way of example. DETAILED DESCRIPTION

[0057] refer to Figure 1a , the spectral imaging device 500 includes an image sensor SEN1 and an imaging optical device LNS1 to form an optical image IMG2 of the object OBJ1 on the image sensor SEN1 by focusing the light LB1 (or LB2) received from the object OBJ1. The spectral imaging device 500 includes a tunable Fabry-Perot interferometer FPI1 to provide one or more passbands for capturing the image IMG2 in a spectrally selective manner.

[0058] The Fabry-Perot interferometer FPI1 forms a transmitted light LB2 by filtering the input light LB1 received from the object OBJ1. The transmitted light LB2 may contain one or more narrow discrete spectral bands of the input light LB1, as defined by the spectral transmittance peaks of the Fabry-Perot interferometer FPI1. The transmitted light LB2 may be focused onto the image sensor SEN1.

[0059] The image sensor SEN1 may include a plurality of light detector pixels P0, P1, P2, P3 arranged in a two-dimensional array. For example, the sensor SEN1 may include a plurality of first detector pixels P0 for detecting red light (R). The sensor SEN1 may include a plurality of second detector pixels P1 for detecting green light (G). The sensor SEN1 may include a plurality of third detector pixels P2 for detecting blue light (B). The sensor SEN1 may include a plurality of fourth detector pixels P3 for detecting infrared light (IR). The signals of the detector pixels P0, P1, P2, P3 may be obtained from the sensor SEN1 as detector signals S R , S G , S B , S IR The image sensor SEN1 may convert the optical image IMG2 into a digital image representing a different color channel of the device 500. Each digital image may include a series of recorded detector signals. For example, a first digital image of a first color channel (R) may include a detector signal obtained from a first detector pixel P0. For example, a second digital image of a second color channel (G) may include a detector signal obtained from a second detector pixel P1.

[0060] The imaging optics LNS1 may include, for example, a camera objective to focus the input light LB1 and / or the transmitted light LB2 to form a focused optical image IMG2 of the object OBJ1 on the image sensor SEN1. The imaging optics LNS1 may also be, for example, a telecentric optic or an afocal optic to limit the maximum tilt of the light rays passing through the Fabry-Perot interferometer FPI1.

[0061] The image sensor SEN1 may convert the optical image IMG2 into a digital image including a plurality of detector signals S obtained from the detector pixels of the image sensor SEN1. R , S G , S B . The digital image may be a two-dimensional digital color image. The image sensor SEN1 may be arranged to capture a digital image of the object OBJ1. The spectral imaging device 500 may be arranged to form an image IMG2 of the object OBJ1 on the image sensor SEN1, and may measure the spectral intensity of two or more different parts of the object OBJ1 substantially simultaneously without changing the orientation of the device 500 relative to the object OBJ1. The spectral imaging device 500 may capture a two-dimensional image IMG2 of the object OBJ1 at two or more wavelength bands (transmittance peaks) of the Fabry-Perot interferometer FPI1.

[0062] The device 500 may include one or more spectral cut-off filters FIL1 , FIL2 for defining a minimum wavelength and a maximum wavelength of a spectral measurement range (RNG2) of the device 500 .

[0063] The spectral imaging device 500 may include a control unit CNT1 for controlling the operation of the device 500. The control unit CNT1 may be arranged to perform one or more data processing operations, for example, to calculate calibrated intensity values ​​of pixels of the spectral image from the recorded detector signals. The control unit CNT1 may perform data processing operations by using the recorded detector signals of the captured digital image. The control unit CNT1 may also be arranged to perform one or more data processing operations, for example, for calibrating and / or verifying the operation of the device 500.

[0064] By changing the distance d between the semi-transparent mirrors M1 and M2 of the interferometer FPI1 F To tune the Fabry-Perot interferometer FPI1. The interferometer FPI1 may include one or more actuators ACU1 for varying the distance d between the mirrors M1, M2. F .

[0065] In an embodiment, the mirror M1 and / or the mirror M2 may be arranged to operate as part of the actuator ACU1. For example, the mirror M1 and / or the mirror M2 may operate as part of an electrostatic actuator. The spectral position of the passband of the interferometer FPI1 may be adjusted according to the control signal V d For example, the interferometer FPI1 may include a drive unit DU1 to control the interferometer FPI1 according to the control signal V d A drive voltage HV1 is generated, wherein the drive voltage HV1 may be arranged to drive one or more actuators ACU1 .

[0066] The spectral imaging device 500 may include a user interface UIF1 for providing information to a user and / or receiving user input from a user. The user interface UIF1 may include, for example, a touch screen.

[0067] The spectral imaging device 500 may include a communication unit RXTX1 to receive and / or transmit data, for example via a communication COM1. The communication unit RXTX1 may communicate, for example, via wireless communication and / or wired communication. The spectral imaging device 500 may communicate, for example, with a database DBASE1, for example, via the Internet. For example, measured detector signals and / or calculated spectral images may be transmitted via the communication unit RXTX1.

[0068] The spectral imaging device 500 may include a memory MEM4 for storing a computer program code PROG1. The computer program code PROG1, when executed by one or more data processors (e.g., CNT1) of the device 500, may cause the device 500 to, for example, capture an image, calculate a calibration intensity value from the measured signal, perform a calibration operation and / or enable a function.

[0069] The spectral imaging device 500 may include a memory MEM1 for storing spectral scale calibration data DPAR1. The spectral scale calibration data DPAR1 may specify the control signal value V d The spectral scale calibration data DPAR1 can be expressed as a function λ0(V d ), which specifies the control parameters (V d ) is associated with the actual spectral position (λ0) of the transmittance peak of the Fabry-Perot interferometer (FPI1).

[0070] Control parameter V d It can be a control parameter of the actuator ACU1 of the Fabry-Perot interferometer FPI1. Alternatively, the control parameter V d It can indicate the measured distance d between the mirrors (M1, M2) of the Fabry-Perot interferometer FPI1 FThe Fabry-Perot interferometer FPI1 may optionally include, for example, a capacitive sensor for measuring the distance d between the mirrors (M1, M2). F , and the sensor can provide the control parameter V d As an indication measure the distance d F Output.

[0071] Spectral transmittance T of Fabry-Perot interferometer FP (λ,V d ) can be used as the control parameter V d The control parameter V d It can be, for example, a digital signal used to set the mirror gap of a Fabry-Perot interferometer. The control parameter V d The control parameter (V d ) may, for example, indicate a signal from a reflector gap sensor. The control parameter (V d ) can, for example, indicate the measurement mirror gap of a Fabry-Perot interferometer. The mirror gap refers to the distance between the semi-transparent mirrors M1, M2 of the Fabry-Perot interferometer.

[0072] The spectral imaging device 500 may include a memory MEM2 for storing spectral intensity calibration data KPAR2. The signal value S obtained from the image sensor SEN1 may be converted into IR , S R , S G , S B In particular, the data KPAR2 may include the pseudo-inverse matrix The elements of (K). Pseudo-inverse matrix The element (K) of can be determined by a calibration procedure. The spectral imaging device 500 can be arranged to determine one or more spectral intensity values ​​X(λ) as measured detector signal values ​​S by using the element (K) as a coefficient. IR , S R , S G , S B A linear combination of .

[0073] The spectral imaging device 500 may be arranged to obtain a detector signal value S from the image sensor SEN1 IR , S R , S G , S B , and by using the calibration parameters KPAR2 (e.g., the pseudo-inverse matrix The element (K) is obtained from the detector signal value S IR , S R , S G , S B Determine the intensity value X n , X n+1 , X n+2 , X n+3 The image sensor SEN1 may provide, for example, several image frames per second, and each image frame may include signal values ​​from several thousand or even several million pixels P0, P1, P2, P3 corresponding to different parts of the optical image IMG2 formed on the image sensor SEN1. Therefore, it may be necessary to determine the spectral intensity values ​​X for thousands or millions of different points of the optical image IMG2. n , X n+1 , X n+2 , X n+3 Using matrix elements (K) can provide a quick way to determine the spectral intensity values ​​X n , X n+1 , X n+2 , X n+3 way.

[0074] Control unit CNT1 may calculate calibrated spectral intensity values ​​of spectral images IMG2λ1, IMG2λ2, ...IMG2λN of object OBJ1 from recorded detector signal values ​​by using calibration data (DPAR, KPAR). Each spectral image IMG2λ1, IMG2λ2, ...IMG2λN may comprise a two-dimensional array of calibrated spectral intensity values.

[0075] The spectral imaging device 500 may include a memory (eg, MEM3) for storing the recorded detector signal S R , S G , S B , S IR .

[0076] The spectral imaging device 500 may include a memory MEM3 for storing spectral images IMG2λ1, IMG2λ2, ... IMG2λN calculated based on the detector signal. The spectral image IMG2λ1 may be a two-dimensional image of the object OBJ1, which represents the spectral components at a first wavelength (e.g., λ1). The spectral image IMG2λ2 may be a two-dimensional image of the object OBJ1, which represents the spectral components at a second different wavelength (e.g., λ2). The spectral images IMG2λ1, IMG2λ2, ... IMG2λN may be determined based on the measured detector signal by using calibration data DPAR1, KPAR2. The spectral images IMG2λ1, IMG2λ2, ... IMG2λN may together constitute a hyperspectral cube CIMG. The spectral imaging device 500 may also be referred to as, for example, a hyperspectral imager.

[0077] The spectral imaging device 500 may include an identifier ID1 to specify the identity of the spectral imaging device 500. The identifier ID1 may be used, for example, to associate unique calibration data with the identity of the spectral imaging device 500. The identifier ID1 may be stored, for example, in a read-only memory.

[0078] SX, SY, SZ represent orthogonal directions. The spectral imaging device 500 may have an optical axis AX1. The optical axis AX1 may be parallel to the direction SZ.

[0079] refer to Figure 1b , the input light LB1 received from the object OBJ1 may have a spectral intensity distribution, namely, a spectrum X(λ). The spectral transmittance TFP(λ,V d ) may have several transmittance peaks PB0, PB1, PB2, PB3 at wavelengths λ0, λ1, λ2, λ3. Each wavelength λ0, λ1, λ2, λ3 may be a control parameter V d The spectral imaging device 500 can instantly detect the input light LB1 only at the narrow discrete spectral bands specified by the wavelengths λ0, λ1, λ2, and λ3. X0 can represent the spectral intensity of the input light LB1 at the wavelength λ0. X1 can represent the spectral intensity of the input light LB1 at the wavelength λ1. X2 can represent the spectral intensity of the input light LB1 at the wavelength λ2. X3 can represent the spectral intensity of the input light LB1 at the wavelength λ3.

[0080] λ MIN and λ MAX represents the limit of the spectral measurement range RNG2 of the Fabry-Perot interferometer 500. RNG1 represents the spectral range, which includes only one transmittance peak of the Fabry-Perot interferometer 500.

[0081] In an embodiment, the spectral positions λ0, λ1, λ2, λ3 of the transmittance peaks of the Fabry-Perot interferometer FPI1 may be matched with the spectral sensitivities of the detector pixels P0, P1, P2, P3 of the image sensor SEN1, so that fast spectral analysis with high spectral resolution may be performed using the spectral imaging device 500. The spectral imaging device 500 may measure spectral intensities at, for example, two, three, four, or five wavelengths substantially simultaneously.

[0082] Figure 2a The spectrum scan of the Fabry-Perot interferometer FPI1 is shown. F With the first value d1, the spectral passbands PB0, PB1, PB2, PB3 of the interferometer FPI1 may be located at a first spectral position λ0 (V d =V1),λ1(V d =V1),λ2(V d =V1),λ3(V d = V1). Each passband has peaks FPEAK0, FPEAK1, FPEAK2, FPEAK3. For example, when the distance d F When the distance d is equal to d1, the peak FPEAK1 of the passband PB1 can be located at the first wavelength λ1. F =d1 and wavelength λ1 can be related to the first control parameter value V d =V1 is associated.

[0083] By changing the distance d between the mirrors M1 and M2 of the interferometer FPI1 F , the spectral positions λ0, λ1, λ2, λ3 of the passbands PB0, PB1, PB2, PB4 of the interferometer FPI1 can be changed.

[0084] When the distance d between the reflectors M1 and M2 F With a second different value d2, the spectral passbands PB0, PB1, PB2, PB3 of the interferometer FPI1 can be located at a second different spectral position λ0 (V d =V2),λ1(V d =V2),λ2(V d =V2),λ3(V d = V2). The distance d between the reflectors M1 and M2 F The change of the control parameter V d For example, when the distance d F When d2 is equal, the peak FPEAK1 of the passband PB1 can be located at different spectral positions λ1(V d =V2). Distance d F =d2 and wavelength λ2 can be related to the second control parameter value V d=V2 is associated.

[0085] The passband may define the spectral position and bandwidth used to capture one or more images using the image sensor SEN1 .

[0086] refer to Figure 2b , the spectral transmittance T of the Fabry-Perot interferometer FPI1 FP (λ,V d ) are two variables (i.e. wavelength λ and control parameter V d ). By scanning the wavelength λ and controlling the parameter V d Each combination of complete spectral transmittance function T FP (λ,V d ) can be very time-consuming. When using this method, there is no need to scan the wavelength λ and control parameter V at the calibration site. d every combination of .

[0087] After obtaining the spectral transmittance T of the Fabry-Perot interferometer FPI1 FP (λ,V d ) can be obtained by tracking the position of the transmittance peak in the plane defined by the wavelength dimension and the control parameter dimension according to the spectral transmittance T FP (λ,V d ) determines the spectral calibration function λ0(V d ),λ1(V d ),λ2(V d )、λ3(V d ).

[0088] refer to Figure 3a The imaging optics LNS1 may form an optical image IMG2 of the object OBJ1 on the image sensor SEN1. The spectral imaging device 500 may be arranged to form an image IMG2 of the object OBJ1 on the image sensor SEN1 and may measure spectral intensities of two or more different portions of the object OBJ1 substantially simultaneously.

[0089] The spectral bandwidth and spectral position of the optical image IMG2 may be defined by the spectral passbands PB0, PB1, PB2, PB3 of the interferometer FPI1. The spectrometer may capture a two-dimensional image IMG2 of the object, for example, at two or more narrow wavelength bands (ie, transmittance peaks) of the Fabry-Perot interferometer FPI1.

[0090] The image sensor SEN1 may include a first detector pixel P0 that is spectrally selectively sensitive to a first color. The image sensor SEN1 may include a second detector pixel P1 that is spectrally selectively sensitive to a second color. The image sensor SEN1 may include a third detector pixel P2 that is spectrally selectively sensitive to a third color. For example, the first detector pixel P0 may be sensitive to red light (R), the second detector pixel P1 may be sensitive to green light (G), and the third detector pixel P2 may be sensitive to blue light (B).

[0091] The image sensor SEN1 may further include a fourth detector pixel P3 that is spectrally selectively sensitive to a fourth color. For example, the first detector pixel P0 may be sensitive to red light (R), the second detector pixel P1 may be sensitive to green light (G), the third detector pixel P2 may be sensitive to blue light (B), and the fourth detector pixel P3 may be sensitive to infrared light (IR).

[0092] The optical image IMG2 may cover a plurality of detector pixels P0, P1, P2 simultaneously in order to analyze spatial variations of the spectrum at different points of the two-dimensional image IMG2. The image IMG2 may cover, for example, four or more adjacent pixels P0, P1, P2 in the direction SX and four or more adjacent pixels in the direction SY.

[0093] The image sensor SEN1 may also be realized by using several detector pixel arrays, wherein the transmitted light LB2 may be distributed to different arrays, for example by means of beam splitters and / or spectrally selective reflectors.

[0094] refer to Figure 3b , pixel P0 can have a spectral quantum efficiency Q R (λ). The pixel P1 may have a spectral quantum efficiency Q G (λ). The pixel P2 may have a spectral quantum efficiency Q B (λ). The pixel P3 may have a spectral quantum efficiency Q IR (λ). Efficiency Q G (λ) can be compared with Q R (λ) is different. Efficiency Q B (λ) can be related to Q R (λ) is different from Q G (λ) is different. Efficiency Q IR (λ) can be compared with Q R (λ) is different from Q G (λ) is different from Q B (λ) is different.

[0095] refer to Figure 3c, the spectral quantum efficiency of the combination of pixel P0 and Fabry-Perot interferometer FPI1 can be formed as a function of the spectral transmittance T of interferometer FPI1 FP (λ,V d ) and the spectral quantum efficiency Q of the detector pixel P0 R The spectral quantum efficiency of the combination of pixel P1 and interferometer FPI1 can be formed as the spectral transmittance function T FP (λ,V d ) and spectral quantum efficiency Q G The spectral quantum efficiency of the combination of pixel P2 and interferometer FPI1 can be formed as a spectral transmittance function T FP (λ,V d ) and spectral quantum efficiency Q B The spectral quantum efficiency of the combination of pixel P3 and interferometer FPI1 can be formed as the spectral transmittance function T FP (λ,V d ) and spectral quantum efficiency Q IR (λ).

[0096] refer to Figure 3d , the spectral imaging device 500 may include one or more cut-off filters FIL1, FIL2 to define the spectral measurement range RNG2 of the spectral imaging device 500. The optical components of the spectral imaging device 500 may have a combined spectral transmittance T FIL (λ). Combined spectral transmittance T FIL (λ) may represent the spectral transmittance of other optical components in the apparatus 500 except for the Fabry-Perot interferometer FPI1 and the image sensor SEN1 .

[0097] refer to Figure 4a , the calibration device 1000 may include a calibration light source LS1 to provide a first calibration light. The device 1000 may include a control unit CNT2 for performing a data processing operation of the calibration.

[0098] The device 1000 may comprise a communication unit RXTX2 to receive and / or transmit data, for example via the communication COM1. The device 1000 may communicate with a database DBASE1, for example via the Internet. For example, measured detector signals and / or calculated spectral images may be transmitted via the communication unit RXTX2.

[0099] The calibration device 1000 may include:

[0100] - a communication unit RXTX2 for obtaining first spectral data DATA1 specifying the spectral quantum efficiency (Q) of the detector pixels (P0, P1, P2, P3) of the image sensor (SEN1);IR , Q R , Q G , Q B ),

[0101] - a communication unit RXTX2 for obtaining second spectral data DATA2 specifying a spectral transmittance function T for the Fabry-Perot interferometer FPI1 FP,PRE (λ,V d ),

[0102] a calibration light source LS1 for forming a first calibration light LB1CAL1 having a first calibration spectrum including a plurality of spectral peaks (CALPEAK1, CALPEAK2, CALPEAK3), and

[0103] - A data processor CNT2 for determining calibration data from the acquired data and the measured data.

[0104] The calibration light source LS1 may include a primary light source LED0 and a multi-passband filter MPBF1. The primary light source LED0 may provide a broadband primary light LB0. The multi-passband filter MPBF1 may form a first calibration light LB1CAL1 by filtering the broadband primary light LB0. The primary light source LED0 may be, for example, a light emitting diode or a halogen tungsten lamp. The calibration light source LS1 may optionally include an optical (spatial) integrator INT1 to provide uniform illumination over the field of view of the spectral imaging device 500. The integrator INT1 may include, for example, an integrating sphere and / or an optical diffuse surface.

[0105] The first calibration light can be provided by using a calibration light source LS1, which comprises a multi-passband filter MPBF1 and a broadband LED or halogen lamp LED0. The multi-passband filter can have multiple spectral passbands. For example, the number of passbands in the range of 400nm to 1000nm can be, for example, in the range of 3 to 20, advantageously in the range of 7 to 12.

[0106] The spectral imaging device 500 may be assembled at an assembly line. The image sensor SEN1 may be combined with the Fabry-Perot interferometer FPI1 at an assembly line. The method may include forming a combination CMB1 including the image sensor SEN1 and the Fabry-Perot interferometer FPI1.

[0107] The spectral imaging device 500 may be arranged to capture an image (IMG2) of the calibration light source LS1. The spectral imaging device 500 may be arranged to capture an image of an area within the output aperture APE1 of the calibration light source LS1. The spectral imaging device 500 may be arranged to measure the spectrum of the calibration light LB1CAL1 by scanning the Fabry-Perot interferometer FPI1. The image sensor SEN1 may convert the intensity values ​​of the image IMG2 into a detector signal S IR , S R , S G , S B The apparatus 1000 may receive the measured detector signal S from the spectral imaging device 500, for example via the communication units RXTX1, RXTX2. IR , S R , S G , S B .

[0108] The device 1000 may be arranged to:

[0109] - By scanning the control parameter V of the Fabry-Perot interferometer d During this period, the detector pixel signal S is recorded IR , S R , S G , S B To obtain the measurement profile S of the first calibration light LB1CAL1 IR,MEAS (V d ), S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ),

[0110] - By using the spectral quantum efficiency (Q IR , Q R , Q G , Q B ), and by using the spectral transmittance function (T FP (λ,V d )), and by using the first calibration data (f k (λ), g k (λ), h k (V d )), according to the spectrum I of the first calibration light LB1CAL1 MPBF (λ) Determine the simulation profile S IR,SIM,k (V d ), S R,SIM,k (V d ), S G,SIM,k (Vd ), S B,SIM,k (V d ),

[0111] - For the first calibration data (f k (λ), g k (λ), h k (V d )) Modify until the simulated profile S IR,SIM,k (V d ), S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) according to one or more matching criteria and the corresponding measured profile S IR,MEAS (V d ), S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) are matched (e.g., until the error indicator value is less than a predetermined limit), and

[0112] - The modified first calibration data (f k (λ), g k (λ), h k (V d )) is stored in the memory.

[0113] The device 1000 may include a memory (eg, MEM21) for storing first spectral data DATA1 specifying previously measured spectral quantum efficiencies Q of detector pixels P0, P1, P2, P3. IR , Q R , Q G , Q B .

[0114] The device 1000 may include a memory (eg, MEM21) for storing second spectral data DATA2, the second spectral data DATA2 specifying a previously measured spectral transmittance T of a Fabry-Perot interferometer. FP,PRE (λ,V d ).

[0115] The device 1000 may include a memory (eg, MEM21) for storing data specifying the spectrum I of the calibration light LB1CAL1. MPBF (λ).

[0116] The device 1000 may include a device for storing calibration data fk (λ), g k (λ), h k (V d ) of the memory (e.g., MEM21).

[0117] The apparatus 1000 may also be arranged to determine the calibration data f k (λ), g k (λ), h k (V d ) to calculate the calibration parameters KPAR2, DPAR1. The calibration parameters KPAR2, DPAR1 may be stored in the memory (MEM1, MEM2) of the spectral imaging device 500. The calibration parameters KPAR2, DPAR1 may be transmitted to the spectral imaging device 500, for example, via the communication units RXTX2, RXTX1.

[0118] The calibration data DPAR1 can specify the control signal value V d The relationship associated with the actual wavelength λ.

[0119] The calibration parameter KPAR2 may comprise, for example, a matrix element K. The spectral imaging device 500 may be arranged to determine one or more spectral intensity values ​​X(λ) as measured detector signal values ​​S by using the element (K) as a coefficient IR , S R , S G , S B A linear combination of .

[0120] In particular, the calibration apparatus 1000 may be arranged to:

[0121] - After using the modified first calibration data (f k , g k ,h k ), according to the quantum efficiency (Q R , Q G ) and according to the spectral transmittance function (T FP,PRE (λ,V d )) forms a matrix element (U) of a first matrix (M), wherein the matrix element (U) of the first matrix is ​​used to convert the detector signals (S) of different color channels R , S G , S B ) is determined as the spectral intensity value (X) of the input light (LB1) at the spectral position (λ0, λ1, λ2) of the spectral transmittance peak of the Fabry-Perot interferometer (FPI1) n , X n+1 , X n+2 ),

[0122] - Forming the second matrix So that the second matrix is the inverse or pseudo-inverse of the first matrix (M), and

[0123] - The second matrix The matrix elements (K) are output and / or stored in a memory (MEM2).

[0124] The device 1000 may comprise a memory MEM22 for storing a computer program PROG2. The computer program PROG2, when executed by the processor CNT2, may cause the calibration device 1000 to perform one or more steps of a calibration method.

[0125] At a later stage, the spectral imaging device can be used to measure the unknown spectrum X(λ). The calibration parameters KPAR2, DPAR1 can be retrieved from the memory (MEM1, MEM2) and can be adjusted according to the detector signal S by using the previously stored calibration parameters KPAR2, DPAR1. IR , S R , S G , S B The unknown spectrum X(λ) is calculated. In particular, the spectral imaging device 500 can calculate the unknown spectrum X(λ) by using the calibration parameter KPAR2 according to the measured detector signal value S IR , S R , S G , S B The spectral images IMG2λ1, IMG2λ2, ..., IMG2λN are calculated.

[0126] refer to Figure 4b , the multi-passband filter MPBF1 can be an interference filter. The multi-passband filter MPBF1 can be realized by depositing multiple material layers on the substrate SUB1. The multi-passband filter MPBF1 may include, for example, a first filter FIL11 implemented on the first main surface of the substrate, and a second filter FIL12 implemented on the second main surface of the substrate. The first filter FIL11 can be arranged to provide several spectral passbands in the first spectral range (for example, in the range of 450nm to 650nm). The second filter FIL12 can be arranged to provide several spectral passbands in the second spectral range (for example, in the range of 625nm to 850nm). The substrate may include, for example, fused silica. The material layers may be, for example, silicon dioxide (SiO2) and titanium dioxide (TiO2).

[0127] Figure 4c It is shown that the first calibration light LB1CAL1 is formed by filtering the broadband primary light LB0 using the multi-passband filter MPBF1. Figure 4cThe uppermost curve shows by way of example the spectral intensity distribution I of the broadband primary light LB0 LED (λ).

[0128] from Figure 4c The second curve from the top shows by way of example the spectral transmittance T of the multi-passband filter MPBF1. MPBF (λ). The multi-passband filter MPBF1 may have a stable spectral position λ 01 , 02 , 03 ......Multiple passbands PB01, PB02, PB03...... at ......

[0129] Figure 4c The lowermost curve shows by way of example the spectral intensity distribution I of the first calibration light LB1CAL1 formed by filtering the primary light LB0 using the filter MPBF1. MPBF (λ). Spectral intensity distribution I MPBF (λ) can be formed as the product I LED (λ)·T MPBF (λ). Spectral intensity distribution I MPBF (λ) may also be referred to as a first calibration spectrum CALSPEC1. The first calibration spectrum CALSPEC1 may include a stable spectral position λ 01 , 02 , 03 ......Multiple calibration peaks CALPEAK1, CALPEAK2, CALPEAK3... at ......

[0130] In an embodiment, the multi-passband filter can be optionally tilted to provide additional spectral peaks for the calibration light. By rotating the multi-passband filter, additional spectral peaks can be provided to perform calibration. The multi-passband filter can be tilted relative to the propagation direction of the light. Calibration measurements can be selectively performed at several different tilt angles of the multi-passband filter. Changing the tilt angle can provide additional narrow spectral bands for calibration. By changing the tilt angle, the number of different spectral bands used for calibration can be, for example, in the range of 50 to 500.

[0131] refer to Figure 5a The first calibration light LB1CAL1 can be coupled to the spectral imaging device 500, and the spectral imaging device 500 can be configured to be a spectral imaging device by changing the distance d between the mirrors M1 and M2 of the Fabry-Perot interferometer. F and by converting the signal of one or more detector pixels (S R , S G , S B , S IR ) is recorded as the control parameter Vd The function provides the measured spectrum MSPEC1. The control parameter V d Can indicate distance d F .

[0132] Figure 5a The uppermost curve of shows by way of example the measurement signal S obtained from the detector pixel of the first color channel (red) of the spectral imaging device 500 R,MEAS (V d ) as the control parameter V d The measurement is the control parameter V d The function signal S R,MEAS (V d ) can also be called the measured spectral profile MSPEC1. R,MEAS (V d ) may include multiple peaks MPEAK1, MPEAK2, MPEAK3...MPEAK6.

[0133] The calibration light LB1CAL1 may have a spectral intensity distribution I MPBF (λ). Profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ), S IR,MEAS (V d ) can be formed as follows:

[0134]

[0135] λ MIN and λ MAX represents the lower limit and upper limit of the spectrum measurement range RNG2 of the spectrum imaging device 500. The range RNG2 includes all spectral components that can be detected by the spectrum imaging device 500. The measurement range RNG2 can also be referred to as MIN to λ MAX spectral integration range.

[0136] I MPBF (λ) represents the spectral intensity distribution of the first calibration light. R,A (λ), Q G,A (λ), Q B,A (λ), Q IR,A (λ) represents the actual spectral quantum efficiency of detector pixels P0, P1, P2, and P3. FP,A (λ,V d ) represents the actual spectral transmittance of the Fabry-Perot interferometer. FIL(λ) represents other optical components of the spectral imaging device 500 (including the limit λ defined). MIN and λ MAX The combined spectral transmittance of one or more cutoff filters FIL1, FIL2).

[0137] The first color channel of the spectral imaging device 500 may have a combined spectral response function Q R,500 (λ), which is equal to the product Q R,A (λ)·T FP,A (λ,V d )·T FIL (λ). The detector signal S of the first color channel R,MEAS (V d ) is formed into the product Q R,500 (λ)·I MPBF (λ) Integrated over the measurement range RNG2.

[0138] from Figure 5a The second curve at the top of the graph shows by way of example the measurement signal S obtained from the detector pixel of the second color channel (green) of the spectral imaging device 500. G,MEAS (V d ) as the control parameter V d function.

[0139] from Figure 5a The third curve from the top shows by way of example the measurement signal S obtained from the detector pixel of the third color channel (blue) of the spectral imaging device 500 B,MEAS (V d ) as the control parameter V d function.

[0140] Figure 5a The lowermost curve shows by way of example the measurement signal S obtained from the detector pixel of the fourth color channel (infrared) of the spectral imaging device 500 IR,MEAS (V d ) as the control parameter V d function.

[0141] refer to Figure 5b , the calibration apparatus 1000 may be arranged to calibrate the detector pixels by using the previously measured spectral quantum efficiency Q R (λ) and the spectral transmittance T of the Fabry-Perot interferometer 500 measured previously FP,PRE (λ,V d ), according to the known spectrum I MPBF (λ) Calculate the simulated profile S R,SIM,k (V d ).

[0142] Function Q of previous measurements R (λ), Q G (λ), Q B (λ), Q IR (λ), T FP,PRE (λ,V d ) can represent the assembled spectral imaging device 500 to some extent, but is not accurate.

[0143] The simulated spectral profiles of the different color channels of the spectral imaging device may initially be (index k=0):

[0144]

[0145] Figure 5b The uppermost curve shows by way of example that the calibration light coupled into the spectral imaging device has a known spectral intensity distribution I MPBF The simulated profile S of the detector pixels of the first color channel (red) of the spectral imaging device under the assumption that R,SIM,k (V d )(SIMSPEC1). Simulated S R,SIM,k (V d )(SIMSPEC1) has multiple peaks SIMPEAK1, SIMPEAK2, SIMPEAK3...SIMPEAK6.

[0146] RNG3 represents the control parameter V d of the scan range.

[0147] from Figure 5b The second curve from the top shows by way of example the simulated profile S of the detector pixels of the second color channel (green). G,SIM,k (V d ).

[0148] from Figure 5b The third curve from the top shows by way of example the simulated profile S of the detector pixels of the third color channel (blue). B,SIM,k (V d ).

[0149] Figure 5b The lowermost curve shows by way of example a simulated profile S of a detector pixel for the fourth color channel (infrared) IR,SIM,k (V d ).

[0150] The spectral scale of the spectral imaging device 500 can be determined, for example, by the function λ0(V d ) is defined, the function λ0(V d ) defines the control parameter value Vd The relationship between the wavelength λ0 corresponding to the peak transmittance of the Fabry-Perot interferometer. d ) can be done, for example, by Figure 5a One or more measured spectral profiles of Figure 5b The measured spectral profile of a color channel may be compared with the simulated spectral profile of the color channel. R,MEAS (V d ) and S R,SIM,k (V d ) can be compared. G,MEAS (V d ) and S G,SIM,k (V d ) can be compared. B,MEAS (V d ) and S B,SIM,k (V d ) can be compared. IR,MEAS (V d ) and S IR,SIM,k (V d ) for comparison.

[0151] refer to Figure 6a , the initial simulation profile SIMSPEC1 may deviate significantly from the corresponding measured profile MSPEC1.

[0152] The initial simulation profile SIMSPEC1 may contain several peaks SIMPEAK1, SIMPEAK2, ... SIMPEAK6. The simulation profile SIMSPEC1 may represent the simulation detector signal S R,SIM,k=0 (V d ) as the control signal V d function.

[0153] The measurement profile MSPEC1 may contain several peaks MPEAK1, MPEAK2, ... MPEAK6. The measurement profile MSPEC1 may represent the measurement detector signal S R,MEAS (V d ) as the control signal V d function.

[0154] refer to Figure 6b , the calibration data can be modified iteratively until the simulated profile (e.g., S R,SIM,k ) according to one or more matching criteria with the corresponding measurement profile (e.g., S R,MEAS (V d ))match.

[0155] After determining the calibration data, the peaks SIMPEAK1, SIMPEAK2, ... SIMPEAK6 of the simulation profile SIMSPEC1 may coincide with the peaks MSPEC1, MSPEC2, ... MSPEC6 of the measurement profile MSPEC1.

[0156] When simulating a profile (e.g., S R,SIM,k ) according to one or more matching criteria with the corresponding measurement profile (e.g., S R,MEAS (V d )) match, the iteration can stop. In an embodiment, the contours do not need to match perfectly. A small deviation between the contours can be allowed. ΔV d Represents the difference between the positions of corresponding peaks (eg, SIMPEAK6, MPEAK6).

[0157] The method may include modifying the estimated spectral transmittance T of the Fabry-Perot interferometer FP,MOD,k (λ,V d ) until the simulated profile corresponds sufficiently to the measured profile.

[0158] The method may include iteratively modifying the estimated spectral transmittance T of the Fabry-Perot interferometer FP,MOD,k (λ,V d ), until the simulated profile sufficiently corresponds to the measured profile. The symbol k may represent the index of the iterative loop.

[0159] Find the best estimate of the spectral transmittance T of the Fabry-Perot interferometer FP,FIT (λ,V d ) can then determine and store the best estimate T FP,FIT (λ,V d ) and use them to calculate calibration spectral intensity values ​​based on the measured detector signals obtained from the image sensor.

[0160] refer to Figure 7a The method may optionally include providing second calibration light LB1CAL2 having a wide spectral bandwidth. The second calibration device 1002 may be arranged to provide the second calibration light LB1CAL2. The second calibration device 1002 may provide the second calibration light LB1CAL2 without using a multi-passband filter MPBF1.

[0161] The second calibration light LB1CAL2 may be used to determine the combined spectral response of the different color channels with improved accuracy.

[0162] The calibration device 1002 may include, for example Figure 4a, wherein the calibration device 1002 may optionally include an actuator SERVO1 for positioning the multi-passband filter MPBF1 to the optical path and / or for positioning the multi-passband filter MPBF1 away from the optical path. Thus, the calibration device 1002 may be arranged to provide the first calibration light or the second calibration light depending on the position of the multi-passband filter MPBF1. The first calibration light may be changed to the second calibration light by positioning the multi-passband filter MPBF1 away from the optical path with the actuator SERVO1.

[0163] The light source LS2 of the device 1002 may include a primary light source LED0, an integrator INT1, and a multi-passband filter MPBF1. The light source LS2 may be arranged to provide a first calibration light with the multi-passband filter MPBF1 during a first time period, and the light source LS2 may be arranged to provide a second calibration light without the multi-passband filter MPBF1 during a second time period.

[0164] refer to Figure 7b , the spectral intensity distribution I of the second calibration light LB1CAL2 CAL2 (λ) can be compared with the spectral intensity distribution of the primary light source LED0 LED (λ) is basically similar. Spectral intensity distribution I LED (λ) may also be referred to as the second calibration spectrum CALSPEC2.

[0165] The second calibration light LB1CAL2 can be coupled to the spectral imaging device 500, and the spectral imaging device 500 can be coupled to the spectral imaging device 500 by changing the distance d between the mirrors M1 and M2 of the Fabry-Perot interferometer. F and by converting the signal of one or more detector pixels (S R , S G , S B , S IR ) is recorded as the control parameter V d The function is used to provide a second set of measurement profiles MSPEC2.

[0166] Calibration apparatus 1000 (and / or 1002) may be arranged to calibrate based on previously measured data Q R (λ), Q R (λ), Q R (λ), Q R (λ), T FP,PRE (λ,V d ) and based on the known spectrum I MPBF (λ) and / or I CAL2 (λ) to simulate the profile S R,SIM,k (V d ), S G,SIM,k (Vd ), S B,SIM,k (V d ), S IR,SIM,k (V d ) for calculation.

[0167] The calibration apparatus 1000 (and / or 1002) may be arranged to modify the calibration parameters such that the simulated profile sufficiently corresponds to the measured profile.

[0168] The calibration apparatus 1000 (and / or 1002) may be arranged to modify the calibration parameters such that the simulated profile sufficiently corresponds to the measured profile.

[0169] The calibration apparatus 1000 (and / or 1002) may be arranged to iteratively modify the calibration parameters until the simulated profile sufficiently corresponds to the measured profile.

[0170] Next, the calibration method will be described in detail for three color channels. For example, calibration can be performed for a red channel, a green channel, and a blue channel. However, the calibration method can also be used when the number of color channels is, for example, in the range of 2 to 10. The number of color channels can be, for example, 2, 3, 4, or 5. The color channels can be specified, for example, by the symbol c∈{R, G, B, IR}. The spectral imaging device may include detector pixels for providing additional color channels, for example, a fourth detector pixel for detecting an infrared band. For example, calibration can be performed for red (R), green (G), blue (B), and infrared (IR) color channels. The spectral imaging device may include detector pixels for providing a greater number of color channels, for example, a fourth detector pixel for detecting a first infrared band, and a fifth detector pixel for detecting a second infrared band.

[0171] When capturing an image of a spatially uniform light source, several pixels may provide detector signals representing the same spectrum. One or more first detector pixels may provide a first detector signal for a first color channel. One or more second detector pixels may provide a second detector signal for a second color channel. One or more third detector pixels may provide a third detector signal for a third color channel.

[0172] According to the above reference Figure 5a The equations (1a), (1b), (1c) discussed form the measured spectral profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ).

[0173] The measured spectral profile specifies the detector signal S c,MEASThe measured amplitude is used as the control parameter V d Function of Spectral Profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) can be achieved by changing the control signal V d , and by recording the amplitude of the detector signal of the first color channel R as the control parameter V d The function of V is obtained by recording the amplitude of the detector signal of the second color channel G as the control parameter V d function of the detector signal of the third color channel B as the control parameter V d function to measure.

[0174] The scanning of the Fabry-Perot interferometer consists of varying the control signal V d , where changing the control signal V d Associated with the mechanical movement of the mirrors of the Fabry-Perot interferometer.

[0175] Measure the spectral profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) can cover the control signal V d The scanning range of RNG3 is from the minimum value V d,MIN To the maximum value V d,MAX The scanning range RNG3 may be divided into a plurality of scanning steps, wherein the number of scanning steps covering the entire range RNG3 may be, for example, in the range of 100 to 10000. Spectral profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) can be determined at each control signal value V determined by the scanning step of the range RNG3 d Measurements are taken at

[0176] The scanning of the Fabry-Perot interferometer takes some time, but can be achieved by changing only one parameter at the calibration site, namely the control signal V d ) to measure the spectral profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ).

[0177] When the present calibration method is used, there is no need to spend time scanning the wavelength λ of the monochromatic calibration light at the assembly line.

[0178] The simulated spectral profiles of different color channels of the spectral imaging device can be formed, for example, according to equations (2a), (2b), (2c),

[0179]

[0180] Reference Figure 5b These (same) equations are also mentioned.

[0181] The measured spectral profile specifies the detector signal S c,MEAS The measured amplitude is used as the control parameter V d function.

[0182] The previously measured spectral transmittance function T FP,PRE Specify the transmittance of the Fabry-Perot interferometer as two variables λ and V d For variables λ and V d The function T of each combination of FP,PRE The measurement is very time-consuming. Thanks to the method, there is no need to measure the function T at the assembly line. FP,PRE The spectral transmittance data T FP,PRE (λ,V d ) and the quantum efficiency data Q R (λ), Q G (λ), Q B The Fabry-Perot interferometer can even be separated from the image sensor when performing (λ) measurements.

[0183] Obtain the spectral transmittance data T measured previously separately FP,PRE (λ,V d ) and quantum efficiency data Q R (λ), Q G (λ), Q B (λ) and subsequent calibration by using the first calibration light can significantly speed up the production of assembled spectral imaging devices.

[0184] Spectral transmittance data T FP,PRE (λ,V d ) can be performed, for example, by using a first measuring device (1003, see Fig.10 ) is obtained. The first measuring device may be separated or far away from the assembly line for assembling the spectral imaging device. The spectral transmittance function T FP,PRE The wavelength λ of the monochromatic calibration light can be swept, for example, over the spectral range RNG2, and the control signal V can be swept over the scanning range RNG3. dThe range RNG2 may be divided into a plurality of scanning steps, wherein the number of scanning steps covering the entire range RN2 may be, for example, in the range of 100 to 100000. The range RNG3 may be divided into a plurality of scanning steps, wherein the number of scanning steps covering the entire range RNG3 may be, for example, in the range of 100 to 10000. FP,PRE (λ,V d ), it is possible to determine at each wavelength λ determined by the scanning step of the range RNG2, and at each control signal value V determined by the scanning step of the range RNG3. d Spectral transmittance T FP,PRE (λ,V d ) for measurement.

[0185] Q R (λ), Q G (λ), Q B (λ) represents the (previously measured) quantum efficiency of the different color channels c of the spectral imaging device. Quantum efficiency data Q R (λ), Q G (λ), Q B (λ) can be obtained, for example, by using a second measurement device in a measurement laboratory. The second measurement device can be separated or remote from the assembly line used to assemble the spectral imaging device. The quantum efficiency data Q R (λ), Q G (λ), Q B (λ) can be obtained, for example, by scanning the wavelength λ of the monochromatic calibration light over the spectral range RNG2. The range RNG2 can be divided into a plurality of scanning steps, wherein the number of scanning steps covering the entire range RN2 can be, for example, in the range of 100 to 100000. The quantum efficiency Q can be measured at each wavelength λ determined by the scanning steps of the range RNG2 R (λ), Q G (λ), Q B (λ).

[0186] T FIL (λ) represents the combined spectral transmittance of other optical components in the spectral imaging device except the Fabry-Perot interferometer. In particular, the spectral transmittance function T FIL (λ) represents one or more passband limiting filters of the spectral imaging device.

[0187] I MPBF (λ) represents the (previously measured) spectral intensity distribution of the calibration light formed by the multi-passband filter. MIN to λ MAXThe intensity of the first calibration light within the entire measurement range RNG2 can be referred to as a function of the wavelength λ. The range RNG2 can be divided into a plurality of scanning steps, wherein the number of scanning steps covering the entire range RNG2 can be, for example, in the range of 100 to 100000. The spectrum I can be known at each wavelength λ determined by the scanning steps of the range RNG2. MPBF (λ).

[0188] The actual quantum efficiency Q of the spectral imaging device R (λ), Q G (λ), Q B (λ) and the actual spectral transmittance function T FP,PRE (λ,V d ) may have a slight error. Initial transmittance function T FP,PRE The modified transmittance function T FP.MOD,k Replace , so as to iteratively minimize the error.

[0189] Modified transmittance function T FP.MOD,k This can be done by, for example, using the helper function f k (λ), g k (λ), h k (V d ) to form the multiplication function f k (λ) can modify the transmittance function T FP.MOD,k The spectral amplitude of . Spectral correction function g k (λ) can provide a modified wavelength λ MOD,k .

[0190] Control correction function h k (V d ) can provide a modified control signal value V d,MOD,k .

[0191] λ MOD,k =g k (λ) (3a)

[0192] V d,MOD, k=h k (V d ) (3b)

[0193] T FP,MOD,k =f k ·T FP,PRE (λ MOD,k , V MOD,k ) (3c)

[0194] T FP,MOD,k =f k ·T FP,PRE (g k (λ), hk (V d )) (3d)

[0195] Multiplication function f k (λ) can be, for example, a polynomial function, a piecewise defined polynomial function (e.g., a polynomial spline function), a power function, or a sum of power functions. A power function is the product of a coefficient and a variable (the variable is raised to a fixed real number). The spectral correction function g k (λ) can be, for example, a polynomial function, a piecewise defined polynomial function, a power function, or a sum of power functions. The control correction function h k (V d ) can be, for example, a polynomial function, a piecewise defined polynomial function, a power function, or a sum of power functions.

[0196] Function f k (λ), g k (λ), h k (V d ) can be implemented, for example, as a polynomial parameterized by polynomial coefficients, optionally including fractional powers and / or a lookup table with points at selected positions in the function domain, optionally followed by interpolation of intermediate values. Function f k (λ), g k (λ), h k (V d ) can be specified by calibration data. The calibration data can be iteratively modified until the simulated profile matches the measured profile. The parameters in the function definition (coefficients, lookup table values, etc.) can be found by an iterative optimization procedure.

[0197] The iterative optimization can be started (k=0) by using the following function:

[0198] f k=0 (λ)=1 (3e)

[0199] g k=0 (λ)=λ (3f)

[0200] h k=0 (V d )=V d (3g)

[0201] T FP,MOD,k=0 =T FP,PRE (λ,V d ) (3h)

[0202] Function f k (λ), g k (λ) and / or h k (V d ) can be modified iteratively until the error index value meets the predetermined standard.

[0203] After the kth iteration (k=1, 2, 3, 4, ...), the simulated spectral profiles of different color channels of the spectral imaging device are:

[0204]

[0205] For example, the error indicator value ErrS k The average deviation or mean squared deviation between the measured and simulated signal amplitudes can be expressed as:

[0206]

[0207] Error index value ErrS k The contribution of all color channels (c) and the d,MIN To V d,MAX All peaks of the profile are measured within the range.

[0208] For example, the error indicator value ErrP k The average deviation or mean squared deviation between the measured and simulated positions of the spectral peaks of the calibration light formed by the multi-passband filter can be expressed as:

[0210] oeLh k =∑ c ∑ p (λ c,p,PEAK,MEAS -λ c,p,PEAK,SIM ) 2 (5b)

[0211] Error indicator value ErrP k The contribution of all color channels can be considered as well as the d,MIN To V d,MAX All peaks of the profile are measured within the range.

[0212] Error index value ErrS k 、ErrP k It may also be called, for example, an error metric or an error metric value.

[0213] λ c,p,PEAK,MEAS represents the wavelength of the peak of the measurement profile, where the peak is associated with the index p, and where the measurement profile is associated with the color c.

[0214] λ c,p,PEAK,SIM represents the wavelength of the peak of the simulated profile, where the peak is associated with the index p, and where the measured profile is associated with the color c.

[0215] From V d,MIN To V d,MAX The first measurement profile S within the range R,MEAS (V d) is associated with a different index p=1, 2, 3, 4, ... PMAX.

[0216] PMAX represents the d,MIN To V d,MAX The number of peaks within the range.

[0217] From V d,MIN To V d,MAX The second measurement profile S G,MEAS (V d ) is associated with a different index p=1, 2, 3, 4, ... PMAX.

[0218] From V d,MIN To V d,MAX The third measurement profile S within the range B,MEAS (V d ) is associated with a different index p=1, 2, 3, 4, ... PMAX.

[0219] The relative weights of different colors and / or different peaks can be equal, as shown in equations (5a) and (5b). k and / or ErrP k Different colors and / or different peaks may also have different relative weights when performing calculations.

[0220] The error index value can also be formed as the error index value ErrS k and ErrP k A linear combination of .

[0221] Error metric (ErrS k 、ErrP k ) can be formulated together with an optimization algorithm to find the function f that minimizes the error metric k (λ), g k (λ) and / or h k (V d ) parameter. You can use the continuous (for f k (λ), g k (λ)) and discrete (for h k (V d ))Global optimization methods for both, but global optimization is not required.

[0222] An example of such an algorithm is the family of "genetic algorithms" for global optimization.

[0223] The definition of the input to the optimizer algorithm can be, for example, the following:

[0224] -The state variable is defined as the function f k(λ), g k (λ) and h k (V d ) parameters.

[0225] - Initial conditions are for example as described in equations (3e), (3f), (3g).

[0226] -The function used for minimization is a subroutine that computes ErrS when given a state variable k or ErrP k The subroutine can be configured to calculate ErrS for a given state variable. k or ErrP k The subroutine can convert the data set (S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ), Q R (λ), Q G (λ), Q B (λ), T FP,PRE (λ,V d ), I MPBF (λ)) as additional inputs, which are used to simulate the profile S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) is calculated and then used for the error metric ErrS k and / or ErrP k Calculation.

[0227] - Information about the expected changes in different parameters can be introduced as nonlinear constrained inequalities. For example, the algorithm can be provided with a subroutine that computes the left side of the following inequality:

[0228] |f k (λ)-1|-ε1(λ)<0 (5c)

[0229] |g k (λ)-λ|-ε2(λ)<0 (5d)

[0230] |h k (V d )-V d |-ε3(V d )<0 (5e)

[0231] Functions ε1(λ), ​​ε2(λ) and ε3(V d) represents the spectral transmittance of each Fabry-Perot interferometer and the previously measured reference data TF P , PRE (λ,V d ) between manufacturing batches.

[0232] Upper and lower limits on the state variables can be specified based on knowledge of the corresponding spectral changes, which limit the range of values ​​that the optimization algorithm searches. Typically, these limits are chosen to be less restrictive than equations (5c), (5d), and (5e), ​​but they can help the optimization algorithm in finding values ​​of the state variables that satisfy (5c), (5d), and (5e).

[0233] The iterative optimization procedure can produce a function f k (λ), g k (λ), h k (V d ), T FP.MOD,k , which in turn provides the best fit, i.e. the minimum error index value. Function f k (λ), g k (λ), h k (V d ), T FP.MOD,k Can be renamed to f FIT (λ), g FIT (λ), h FIT (V d ), T FP.FIT (λ). The most accurate estimate of the spectral transmittance function of a Fabry-Perot interferometer is

[0234] T FP,FIT (λ,V d ) = f FIT ·T FP,PRE (g FIT (λ), h FIT (V d )) (6a)

[0235] Function T FP,FIT (λ,V d ) will control the signal value V d is associated with the corresponding wavelength λ of the transmittance peak of the Fabry-Perot interferometer. The best fit spectral transmittance function T FP,FIT (λ,V d ) allows to determine the control signal V d A fairly accurate relationship between the corresponding wavelengths λ0, λ1, λ2, ... of the spectral transmittance peaks of the Fabry-Perot interferometer. One or more spectral calibration functions λ0(V d ),λ1(V d )、λ2(Vd )...... can be calculated based on the spectral transmittance function T FP,FIT (λ,V d ) is determined. The spectral calibration function λ0(V d ),λ1(V d ),λ2(V d )......Establish control signal V d The relationship between the corresponding wavelengths λ0, λ1, λ2, ... of the spectral transmittance peaks of the Fabry-Perot interferometer.

[0236] Transmittance function T FP,FIT (λ,V d ) can be calculated by the following first spectral calibration function according to the control signal value V d Sure

[0237] λ0=λ0(V d ) (6b)

[0238] With the transmittance function T FP,FIT (λ,V d ) corresponds to the control signal value V d It can be determined according to the following inverse spectral calibration function

[0239] V d =V d (λ0) (6c)

[0240] in

[0241] λ0=λ0(V d (λ0)) (6c)

[0242] The detector signals for different color channels can be estimated with high accuracy as follows

[0243]

[0244] S R,FIT (V d ), S G,FIT (V d ), S B,FIT (V d ) represents the measured spectral profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) best fit to the simulated detector signal.

[0245] The spectral imaging device can be fully calibrated by using a first calibration light having a stable known spectrum I MPBF (λ), the spectrum has multiple spectral peaks.

[0246] In addition to using the first calibration light, the calibration method may optionally include using a second broadband calibration light. Spectral intensity information may be obtained from the second calibration light alone. The second calibration light may be obtained from a broadband light source without using a multi-passband filter. The first calibration light may be used to calibrate the spectral scale, and the second calibration light may be used to calibrate the spectral intensity value. The first calibration light may be used to calibrate the error indicator value ErrP k The second calibration light can be used to minimize the error indicator value ErrS k When the second broadband calibration light is used, it is not necessary to accurately know the spectral intensity value of the first calibration light, but only the spectral position of the spectral peak of the first calibration light needs to be known.

[0247] Therefore, the method may optionally include coupling a second calibration light to the spectral imaging device. The second calibration light may have a broadband spectrum I CAL2 (λ). When the second calibration light is used, the second set of analog signals S can be formed as follows R,SIM,k , S R,SIM,k , S R,SIM,k :

[0248]

[0249] When the second calibration light is coupled to the spectral imaging device, a second set of detector signals S for different color channels can be measured. R,MEAS , S G,MEAS , S B,MEAS , and can be iteratively modified according to the known broadband spectrum I CAL2 (λ) The calculated analog signal S R,SIM,k , S R,SIM,k , S R,SIM,k , until the error indicator value ErrS k In addition to using the first calibration light, using the second calibration light can also facilitate determining the amplitude correction function f FIT (λ).

[0250] After the spectral imaging device is produced and calibrated, it can then be used to measure the unknown spectrum X(λ) of the input light. When measuring the unknown input spectrum X(λ), the formation of the detector signals of the different color channels R, G, B can be approximated as follows:

[0251]

[0252] From λ MINto λ MAX The spectral range of may include several transmittance peaks of the Fabry-Perot interferometer. Each transmittance peak may transmit a spectral component of the unknown spectrum X(λ) at the wavelength of the transmittance peak. Therefore, the detector signal of the color channel may be formed as a linear combination of the spectral component contributions of several discrete wavelengths.

[0253] When the number of transmittance peaks is less than or equal to the number of color channels, then the values ​​of the input spectrum at the wavelengths corresponding to each of the transmittance peaks may be calculated.

[0254] The relative contribution of the input spectrum at discrete wavelengths λ0, λ1, λ2 to the detector signal for a given color channel can be approximated using the matrix elements U. The discrete wavelengths λ0, λ1, λ2 are the wavelengths of the Fabry-Perot interferometer at a given control signal value V. d The matrix element U can be defined, for example, as follows:

[0255]

[0256] The width of the integrated spectral range RNG1 can be chosen, for example, so that the range RNG1 includes at least 95% of the spectral transmittance peak in question, but does not include more than 5% of the area of ​​any other spectral transmittance peak. The width of the spectral range RNG1 can be in the range of 20% to 50% of the free spectral range of the Fabry-Perot interferometer.

[0257] Detector signal S R , S G , S B Can be formed as intensity value X n , X n+1 , X n+2 A linear combination of the intensity values ​​X n is the intensity of the first transmittance peak at wavelength λ0, and the intensity value X n+1 is the intensity at wavelength λ1 of the second transmittance peak, and the intensity value X n+2 is the intensity at wavelength λ2 of the third transmittance peak. Wavelengths λ0, λ1, and λ2 are control signal values ​​V d function.

[0258] Detector signal S R , S G , S B The formation of can be described by the following matrix equation

[0259]

[0260] Where X n =X(λ0(V d ))、Xn+1 =X(λ1(V d ))、X n+2 =X(λ2(V d ))、S R =S R,MEAS (V d ), S G =S G,MEAS (V d ), and S B =S B,MEAS (V d ).

[0261] The matrix M can be composed of the first set of elements U

[0262]

[0263] therefore

[0264]

[0265] Strength value X n , X n+1 , X n+2 This can be done by using the pseudo-inverse matrix of the matrix M From the detector signal S R , S G , S B Solved in.

[0266]

[0267] Pseudo-inverse matrix The calculation is as follows

[0268]

[0269] Among them, M T is the transpose of the matrix, ‖·‖ is the matrix norm, I is the identity matrix, (·) -1 is the inverse matrix. μ is a scalar constant, with a typical value of μ = 10 -3 It is used to control the size of the result, and thus ultimately the pixel noise in the final processed image, especially when the rows of the matrix M are linearly dependent or numerically close to linearly dependent, that is, when the matrix M is not invertible or numerically close to being not invertible. If the matrix M is invertible and μ = 0, then That is, the method of the old patent is included as a special case.

[0270] Pseudo-inverse matrix It can be composed of the second set of matrix elements K

[0271]

[0272] Matrix M and / or pseudo-inverse matrix Provides accurate calibration spectrum I MPBF (λ) and the measured detector signal value.

[0273] The matrix M enables the calibration spectrum I to be MPBF The exact value of (λ)I MPBF (λ0(V d ))、I MPBF (λ1(V d ))、I MPBF (λ2(V d )) is mapped to the measured value S of the spectral profile R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ).

[0274] Pseudo-inverse matrix It is possible to obtain the measured value S of the spectral profile from R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) is mapped to the calibration spectrum I MPBF The exact value of (λ)I MPBF (λ0(V d ))、I MPBF (λ1(V d ))、I MPBF (λ2(V d )).

[0275] At a later stage, when the unknown spectrum X(λ) is measured using a spectral imaging device, the pseudo-inverse matrix It is possible to obtain the measured value S of the spectral profile from R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) is mapped to the exact value X(λ0(V d ))、X(λ1(V d ))、X(λ2(V d )).

[0276] The number of different color channels may be, for example, in the range of 2 to 5. The number of spectral transmittance peaks within the passband of the spectral imaging device may be, for example, in the range of 1 to 5. The number of spectral transmittance peaks is less than or equal to the number of different color channels. When the number of transmittance peaks is less than or equal to the number of color channels in the spectral imaging device, the value of the input spectrum at the wavelength corresponding to each of the transmittance peaks may be recovered.

[0277] The number of spectral transmittance peaks within the spectral measurement range (RNG2) may be less than or equal to the number of color channels (c).

[0278] For example, the number of different color channels may be two, while the number of transmittance peaks may be one.

[0279] For example, the number of different color channels may be three, while the number of transmittance peaks may be one.

[0280] For example, the number of different color channels may be four, while the number of transmittance peaks may be one.

[0281] For example, the number of different color channels may be two, and the number of transmittance peaks may be two.

[0282] For example, the number of different color channels may be three, while the number of transmittance peaks may be two.

[0283] For example, the number of different color channels may be 4, while the number of transmittance peaks may be 2. The color channels may be, for example, red (R), green (G), blue (B), and infrared (IR).

[0284] In this case, the matrix M can be

[0285]

[0286] The symbol n represents the wavelength λ0(V d ) is associated with the transmittance peak at wavelength λ1 (V d ) is associated with the transmittance peak at .

[0287] The corresponding pseudo-inverse matrix Can be

[0288]

[0289] For example, the number of different color channels may be three, and the number of transmittance peaks may be three.

[0290] For example, the number of different color channels may be four, while the number of transmittance peaks may be three.

[0291] For example, the number of different color channels may be four, and the number of transmittance peaks may be four.

[0292] The spectral characteristics of individual components of a spectral imaging device may exhibit statistically random variations that differ from previously measured representative characteristics. Manufacturing tolerances can be calculated using the function f k (λ), g k (λ), h k (V d ) is at least partially compensated. Within the range of manufacturing batch variations, we intuitively have

[0293] f k (λ)≈1 (12a)

[0294] g k (λ)≈λ (12b)

[0295] h k (V d )≈V d (12c)

[0296] In other words, when the multiplication function f k When (λ) is approximately equal to one, but not exactly equal to one, the simulated signal can usually correspond to the measured signal. In other words, when the spectral correction function gk(λ) is approximately equal to λ, but not exactly equal to λ, the simulated signal can usually correspond to the measured signal. In other words, when the control correction function h k (V d ) is approximately equal to V d , but not exactly equal to V d When , the simulated signal can usually correspond to the measured signal. More strictly, the above can be written as

[0297] f k (λ)-1|<ε1(λ) (12d)

[0298] |g k (λ)-λ|<ε2(λ) (12e)

[0299] |h k (V d )-V d |<ε3(V d ) (12f)

[0300] Functions ε1(λ), ​​ε2(λ) and ε3(V d ) represents the spectral transmittance of each Fabry-Perot interferometer and the previously measured reference data T FP,PRE (λ,V d ) between manufacturing batches.

[0301] Alternatively, if the spectral quantum efficiency Q of the detector pixel is known R (λ), Q G (λ), Q B (λ) varies significantly, a different multiplication function f can be used for each color channel R,k (λ), f G,k (λ), f B,k (λ) to replace the common multiplication function f k (λ). Multiplication function f R,k (λ), f G,k (λ), f B,k (λ) may take into account possible variations of the transmittance value of the current Fabry-Perot interferometer cell and the spectral quantum efficiency of the color channel c compared to the reference Fabry-Perot interferometer transmittance at each wavelength. Since the contributions of the Fabry-Perot interferometer transmittance and quantum efficiency are combined into the assembled spectral imaging device, the individual variations of the characteristics of the Fabry-Perot interferometer cell and the individual variations of the characteristics of the image sensor may be combined into the same function.

[0302] The method may include:

[0303] - Get the previously measured quantum efficiency function Q R , Q G , Q B ,

[0304] - Obtain the previously measured spectral transmittance function T of the Fabry-Perot interferometer FP,PRE (λ,V d ),

[0305] - Assembling a spectral imaging device by combining a Fabry-Perot interferometer with an image sensor,

[0306] - coupling a first calibration light to a spectral imaging device, wherein the first calibration light comprises a plurality of spectral calibration peaks,

[0307] - When the spectral imaging device detects the first calibration light, by changing the control signal value V d To measure the first set of spectral profiles, and record the detector signals of several color channels as control signals V d The function of

[0308] - By using the calibration functions f(λ), g(λ), h(V d ), according to the previously measured quantum efficiency function Q R , Q G , Q B , according to the previously measured spectral transmittance function T FP,PRE (λ,V d), and according to the spectral intensity distribution I of the first calibration light MPBF (λ), forming the second set of simulated spectral profiles,

[0309] - compare the measured spectral profile with the simulated spectral profile,

[0310] - By adjusting the calibration functions f(λ), g(λ), h(V d ) is iteratively optimized to determine the calibration parameters so that the simulated spectral profile corresponds to the measured spectral profile.

[0311] - By using the determined calibration functions f(λ), g(λ), h(V d ) to determine a first set of matrix elements (U), wherein the first set of matrix elements (U) constitute a first matrix M,

[0312] - Determine a second set of matrix elements (K) such that the second set of matrix elements (K) constitutes a pseudo-inverse matrix of the first matrix M

[0313] - Storing the second set of matrix elements (K) in a memory of the spectral imaging device.

[0314] The second set of matrix elements (K) may then be used to determine an intensity value X from the measured detector signal.

[0315] refer to Figure 8a , a method for producing and calibrating a spectral imaging device may include, for example, the following steps.

[0316] The spectral quantum efficiency Q of a given previously measured detector pixel P0, P1, P2 can be obtained R (λ), Q G (λ), Q B The first spectrum data DATA1 of (λ) is obtained (step #801).

[0317] The spectral transmittance T of a given previously measured Fabry-Perot interferometer can be obtained FP,PRE (λ,V d )'s second spectrum data DATA2 (step #802).

[0318] The Fabry-Perot interferometer may be combined with the image sensor (step #805). A combination CMB1 including the Fabry-Perot interferometer FPI1 and the image sensor SEN1 may be formed. In particular, the spectral imaging device 500 including the combination CMB1 may be assembled at an assembly line.

[0319] The first calibration light is formed by using the multi-passband filter MPBF1 (step #810).

[0320] The control parameter value V can be obtained by coupling the first calibration light into the spectrometer and by scanning the control parameter value V over the scanning range RNG3. d The detector signal is measured to obtain the measurement profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d )(Step #815). The first calibration light LB1CAL1 can be coupled to the combined light CMB1 and can be used to calibrate the profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) for measurement.

[0321] The known spectrum I of the first calibration light can be obtained by using equations (2a)-(4c). MPBF (λ) to simulate the profile S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) is calculated (step #851). For example, for the first simulated profile S R,SIM,k (V d ) calculations include the use of simulated spectral response Q R (λ)·T FP,PRE (λ,V d ). For example, by using the function f k (λ), g k (λ), h k (V d ), based on the quantum efficiency Q of the first calibration light LB1CAL1 R (λ), spectral transmittance T FP,PRE (λ,V d ) and spectrum I MPBF (λ) to simulate the profile S R,SIM,k (V d ) for calculation.

[0322] At the assembly line, only the control parameter V can be scanned d There is no need to change the spectral position of the spectral peak of the first calibration light at the assembly line, since this time-consuming operation (=scanning of the wavelength λ) can be omitted by using the simulated response of the assembled spectral imaging device.

[0323] One or more error indicator values ​​(ErrS) may be calculated, for example, by using equations (5a) and / or (5b). k, ErrP k )(Step 861).

[0324] The method may optionally include coupling a second broadband calibration light into the spectral imaging device (step #820). The second calibration light LB1CAL2 may be formed without using a multi-passband filter MPBF1. The method may include coupling the second calibration light into the spectrometer and scanning the control parameter value V over a scanning range RNG3. d The detector signal is measured to obtain a second set of measurement profiles S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d )(Step #825). The known spectrum I of the second calibration light can be used to CAL2 (λ) for the second set of simulated profiles S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) is calculated (step #852). This can be done by using the function f k (λ), g k (λ), h k (V d ), based on the quantum efficiency Q of the second calibration light LB1CAL2 R (λ), spectral transmittance T FP,PRE (λ,V d ) and spectrum I CAL2 (λ) for the second set of simulated profiles S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) for calculation.

[0325] The error index value (ErrS) of the second calibration light can be calculated, for example, by using equation (5a). k ) is calculated (step #862). Error index value (ErrS k ) can indicate the measurement profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) and the simulated profile S R,SIM,k (V d ), S G,SIM,k (V d ), SB,SIM,k (V d ) between them. Error index value (ErrS k ) can indicate the measurement profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) and the simulated profile S R,SIM,k (V d ), S G,SIM,k (V d ), S B,SIM,k (V d ) between the two.

[0326] The method may include examining one or more error indicator values ​​(ErrS k , ErrP k ) satisfies one or more criteria (step #870).

[0327] Spectral transmittance function T FP,MOD,k (λ,V d ) can be modified iteratively until one or more error indicator values ​​(ErrS k , ErrP k ) satisfies one or more criteria (step #875).

[0328] Function f k (λ), g k (λ) and / or h k (V d ) can be modified iteratively until one or more error indicator values ​​(ErrS k , ErrP k ) satisfies one or more criteria. Function f k (λ), g k (λ) and / or h k (V d ) can be modified iteratively until the error index value ErrS k is less than the first predetermined limit LIM1. Function f k (λ), g k (λ), h k (V d ) can be modified iteratively until the error index value ErrP k is less than the second predetermined limit LIM2. Function f k (λ), g k (λ), h k (V d ) can be modified iteratively until the error index value ErrS kis less than the first predetermined limit LIM1, and the error indicator value ErrP k is smaller than the second predetermined limit value LIM2.

[0329] Function f k (λ), g k (λ), h k (V d ) can be modified iteratively until the error index value ErrS k and / or ErrP k within the predetermined tolerance margin.

[0330] The method may comprise comparing the measured profile MSPEC2 with the simulated profile SIMSPEC2 and modifying the calibration data based on a result of the comparison.

[0331] The modified spectral transmittance function T FP,MOD,k (λ,V d )'s final version T FP,FIT (λ,V d ) is stored in the memory (step #880). The specified function T FP,FIT (λ,V d ) is stored in memory. The specified function f k (λ), g k (λ), h k (V d )’s calibration data is stored in memory.

[0332] refer to Figure 8b , we can use equations (9a)-(9i) to solve the elements (U Rn ......U Bn+2 ) is calculated (step #910).

[0333] It can be obtained, for example, by using equation (10e) according to the elements (U R,n ......U Bn+2 ) for the pseudo-inverse matrix Elements (K Rn ......K Bn+2 ) is calculated (step #920).

[0334] The pseudo-inverse matrix can be Elements (K Rn ......K Bn+2 ) is stored in the memory of the spectral imaging device (step #930). In particular, the pseudo-inverse matrix Elements (K Rn ......K Bn+2) is stored in the memory (MEM2) of the spectral imaging device 500.

[0335] Figure 8c The method steps for measuring the unknown spectrum X(λ) of the input light LB1 are shown. When measuring the unknown spectrum X(λ), the input light LB1 can be coupled to the spectral imaging device, and at the selected control signal value V d Under this condition, a set of detector signals S is obtained from the image sensor SEN1. R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d )(Step #1110).

[0336] Pseudo-inverse matrix The matrix elements (K Rn ......K Bn+2 ) can be used to measure the spectral signal S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) for one or more spectral intensity values ​​(X n (λ0(V d )),(X n+1 (λ1(V d )),(X n+2 (λ2(V d )) performs calculations (step #1120).

[0337] The one or more spectral intensity values ​​(X n (λ0(V d )),(X n+1 (λ1(V d )),(X n+2 (λ2(V d )) output and / or stored in memory (step #1130).

[0338] This can be done, for example, by using the spectral calibration function λ0(V d ) determines the control signal value V d The corresponding wavelength λ0 is given by equation (6b).

[0339] This can be done, for example, by using the inverse spectral calibration function V d (λ0) to determine the control signal value V corresponding to λ0 d , see equation (6c).

[0340] In an embodiment, the control signal V d For a set of contours S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) is measured. The pseudo-inverse matrix The matrix elements (K Rn ......K Bn+2 ) can be used to measure the profile S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) for one or more spectral intensity values ​​(X n (λ0(V d ))、(X n+1 (λ1(V d ))、(X n+2 (λ2(V d )) is calculated. The calculated spectral intensity value (X n (λ0(V d ))、(X n+1 (λ1(V d ))、(X n+2 (λ2(V d )) can be expressed as the control signal V d or the wavelength (λ0, λ1 or λ2) of the transmittance peak of the Fabry-Perot interferometer.

[0341] In particular, the pseudo-inverse matrix can be obtained by using The matrix elements (K Rn ......K Bn+2 ), according to a set of measured profiles S R,MEAS (V d ), S G,MEAS (V d ), S B,MEAS (V d ) is used to calculate the spectral intensity distribution (X(λ)) of the input light LB1.

[0342] One or more calculated spectra (X n (λ0(V d ))、(X n+1 (λ1(V d ))、(X n+2 (λ2(V d )) output and / or stored in memory (step #1130).

[0343] A computer program (PROG2) which, when executed by at least one data processor (CNT2), can cause a device (1000, 1002) to perform one or more steps of the method.

[0344] A computer program product embodied on a non-transitory computer readable medium (MEM22) may include computer program code (PROG2) for performing one or more steps of the method.

[0345] The calibrated spectral imaging device 500 can be used, for example, in remote sensing applications, where it can have high transmittance at the wavelength band selected for imaging. The device 500 can be used, for example, to monitor spatial changes in the color of the object OBJ1. The spectrometer 500 can be used, for example, for absorption measurements, where the wavelength λ0 of the first transmittance peak matches the absorption band, and the wavelength λ1 of the second transmittance peak matches the reference band. The device 500 can be used, for example, for absorption measurements, where the wavelength λ0 of the first transmittance peak matches the fluorescence, and the wavelength λ1 of the second transmittance peak matches the illumination light that induces the fluorescence.

[0346] Figures 9a to 9c The measured spectrum X(λ) of the input light LB1 in the case where calibration light LB1CAL1 is used as the input light LB1 coupled to the spectral imaging device 500 is shown by way of example.

[0347] Figures 9a to 9c The spectrum X(λ) is obtained by using the calibration data (f k , g k ,h k ) according to the measured detector signal (S R,MEAS (V d ), S R,MEAS (V d ), S B,MEAS (V d ), S IR,MEAS (V d ))Calculated by .

[0348] Calibration data (f k , g k ,h k ) is modified iteratively until each simulation profile (S R,SIM,k , S G,SIM,k , S B,SIM,k , S IR,SIM,k ) according to one or more matching criteria and the corresponding measured profile (S R,MEAS (V d ), S R,MEAS (V d ), S B,MEAS (V d ), SIR,MEAS (V d Each completed iteration cycle considers four measured profiles (S R,MEAS (V d ), S R,MEAS (V d ), S B,MEAS (V d ), S IR,MEAS (V d )).For example, Figure 5a and Figure 5b Examples of these profiles are shown.

[0349] Used for calculation Figure 9a The calibration data (f k , g k ,h k ) is obtained by executing one iterative loop (k=1).

[0350] Used for calculation Figure 9b The calibration data (f k , g k ,h k ) is obtained by executing four iteration cycles (k=4).

[0351] Used for calculation Fig.9c The calibration data (f k , g k ,h k ) is obtained by executing eleven iteration loops (k=11).

[0352] Figure 9d The actual spectrum I of the first calibration light LB1CAL1 is shown MPBF (λ), the spectrum is measured by an accurate reference instrument traceable to an international standard. The reference instrument may include, for example, a scanning monochromator and a detector.

[0353] Fig.9c The measured spectrum and Figure 9d Comparison with the actual spectra of shows that performing eleven iteration cycles (k=11) is sufficient to provide calibration data (f k , g k ,h k ), the calibration data makes Fig.9c The measured spectrum X(λ) is related to Figure 9d The actual spectrum I MPBF The RMS deviation of (λ) is less than 0.5%. RMS stands for root mean square.

[0354] Increasing the number of iteration cycles can further improve the accuracy of the measured spectrum. The number of iteration cycles (k) can be, for example, greater than 10. The number k can be, for example, greater than 100, or even greater than 1000. The number k can be, for example, in the range of 10 to 100. The number k can be, for example, in the range of 10 to 1000.

[0355] The unit of the horizontal scale of the spectrum X(λ) may be, for example, nm. The unit of the vertical scale of the spectrum X(λ) may be, for example, W·sr. -1 ·m -2 ·nm -1 W represents watt. sr represents steradian. m represents meter. nm represents nanometer. The spectral intensity value of the spectrum X(λ) may represent, for example, the radiance per unit wavelength of the surface of the object OBJ1.

[0356] refer to Fig.10 , initial spectral transmittance function T FP,PRE (λ,V d ) can be measured, for example, by using an apparatus 1003, which includes a light source LED3, a tunable filter FIL3, a representative Fabry-Perot interferometer FPIO, a detector DET3 and a control unit CNT3. The tunable filter FIL3 can provide monochromatic light LB3 by filtering the broadband light LBLED3 obtained from the light source LED3 or by filtering the broadband light transmitted through the representative Fabry-Perot interferometer FPIO. The tunable filter FIL3 can be, for example, a scanning monochromator. The filter FIL3 can also be located between the representative Fabry-Perot interferometer FPIO and the detector DET3.

[0357] The method may include scanning the wavelength λ of the monochromatic light LB3 transmitted through the representative Fabry-Perot interferometer FPI0 LB3 , by scanning the control parameter V of the representative Fabry-Perot interferometer FPI0 d , and by recording the intensity of monochromatic light LB3 as wavelength λ LB3 The function and record as the control parameter V d The function to measure the initial spectral transmittance function T FP,PRE (λ,V d ). The recorded data can cover both RNG2 and RNG3 scanning ranges.

[0358] The transmittance can be determined by dividing the recorded intensity function by a reference intensity function recorded without using the interferometer FPIO.

[0359] The tunable filter FIL3 can form a tunable wavelength λ by filtering the broadband light LBLED3 obtained from the light source LED3. LB3The detector DET3 can detect the light LB3 that has been transmitted through the tunable filter FIL3 and passed through the Fabry-Perot interferometer FPIO. The detector DET3 can provide a signal SD indicating the intensity of the detected light (LB3). ET3 The control unit CNT3 and the detector DET3 can record the spectral intensity of the light LB3 transmitted through the tunable filter FIL3 and the Fabry-Perot interferometer FPI1. The control unit CNT3 can change the tunable wavelength λ LB3 and control parameter V d , to convert the spectral transmittance T FP,PRE (λ,V d ) is recorded as two parameters (λ LB3 , V d ) function. The recorded spectral transmittance T FP,PRE (λ,V d ) can be stored in the memory MEM33. The recorded spectral transmittance T FP,PRE (λ,V d ) can then be transmitted to, for example, the calibration device 1000, 1002 and / or the spectral imaging device 500.

[0360] In the initial spectral transmission function T FP,PRE (λ,V d ) after measurement, the Fabry-Perot interferometer FPI1 of the spectral imaging device (500) can be combined with the image sensor SEN1.

[0361] Initial spectral transmission function T FP,PRE (λ,V d ) can basically represent the Fabry-Perot interferometer FPI1 of the assembled spectral imaging device 500.

[0362] The representative Fabry-Perot interferometer FPI0 may be the same interferometer as the Fabry-Perot interferometer FPI1 of the spectral imaging device 500 or an interferometer different from the Fabry-Perot interferometer FPI1 of the spectral imaging device 500. The measured initial spectral transmittance function T of the representative Fabry-Perot interferometer FPI0 is FP,PRE (λ,V d ) can be used to represent the initial spectral transmittance function T of the Fabry-Perot interferometer FPI1 of the spectral imaging device 500 FP,PRE (λ,V d ).

[0363] The assembled spectral imaging device 500 may include a first Fabry-Perot interferometer FPI1 and a first image sensor SEN1. The previously measured quantum efficiency may be obtained by measuring the characteristics of a second different image sensor. The previously measured spectral transmittance may be obtained by measuring the characteristics of a second different Fabry-Perot interferometer FPI0. The previously measured quantum efficiency and spectral transmittance may be considered to be representative of the assembled spectral imaging device 500 to some extent, but not completely representative.

[0364] It will be apparent to those skilled in the art that modifications and variations of the systems, products, devices, apparatuses and methods of the present invention are understandable. These figures are schematic. The specific embodiments described above with reference to the accompanying drawings are illustrative only and are not intended to limit the scope of the present invention as defined by the appended claims.

Claims

1. A method for calibrating (500) a spectral imaging device, the method comprising: - obtaining first spectral data (DATA1) specifying the spectral quantum efficiency (Q) of the detector pixels (P0, P1, P2) of the image sensor (SEN1); R , Q G , Q B ), - obtaining second spectral data (DATA2) specifying the spectral transmittance function (T FP,PRE (λ,V d )), - providing the spectral imaging device (500), the spectral imaging device comprising the Fabry-Perot interferometer (FPI1) ​​and the image sensor (SEN1), the image sensor (SEN1) comprising a first detector pixel (P0) and a second detector pixel (P1), - providing a first calibration light (LB1CAL1), wherein the first calibration light (LB1CAL1) has a first calibration spectrum (I MPBF ), the first calibration spectrum includes a plurality of spectral peaks (CALPEAK1, CALPEAK2, CALPEAK3), - coupling the first calibration light (LB1CAL1) to the spectral imaging device (500), - In the control parameters (V d ) is scanned by recording the first detector pixel signal (S R,MEAS ) to obtain a first measurement profile (S) of the first calibration light (LB1CAL1) R,MEAS (V d )), - In the control parameter (V d ) is scanned by recording the pixel signal (S G,MEAS ) to obtain a second measurement profile (S G,MEAS (V d )), - By using the spectral quantum efficiency (Q R ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) to determine the first simulation profile (S R,SIM,k ), - By using the spectral quantum efficiency (Q G ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) to determine the second simulation profile (S G,SIM,k ), - for the first calibration data (f k , g k ,h k ) is modified until the simulation profile (S R,SIM,k , S G,SIM,k ) according to one or more matching criteria and the corresponding measured profile (S R,MEAS (V d ), S G,MEAS (V d )) matches, and - The modified first calibration data (f k , g k ,h k ) is stored in the memory (MEM21).

2. The method according to claim 1, wherein: The spectral transmittance function (T FP,PRE (λ,V d )) specifies the spectral transmittance of the Fabry-Perot interferometer (FPI1) ​​as a function of wavelength (λ) and also as the control parameter (V d ), wherein the control parameter (V d ) indicates the distance (d) between the mirrors (M1, M2) of the Fabry-Perot interferometer (FPI1) F ).

3. The method according to claim 1 or 2, wherein: The number of color channels (c) of the spectral imaging device (500) is in the range of 2 to 5.

4. The method according to any one of claims 1 to 3, wherein: The spectral transmittance function (T FP,PRE (λ,V d )) is within the spectral measurement range (RNG2) of the spectral imaging device (500) and is in the range of 2 to 5, wherein the number of spectral transmittance peaks within the spectral measurement range (RNG2) is less than or equal to the number of the color channels (c).

5. The method according to any one of claims 1 to 4, wherein: The number of passbands (PB01, PB02) of the multi-passband filter (MPBF1) is within the spectrum measurement range (RNG2) of the spectrum imaging device (500) and is in the range of 3 to 20.

6. The method according to any one of claims 1 to 5, comprising: - providing a second broadband calibration spectrum (I CAL2 ), - coupling the second calibration light (LB1CAL2) to the spectral imaging device (500), - In the control parameters (V d ) is scanned by recording the first additional detector pixel signal (S R,MEAS ) to obtain a first additional measurement profile (S) of the second calibration light (LB1CAL2) R,MEAS (V d )), - In the control parameter (V d ) is scanned by recording the pixel signal (S G,MEAS ) to obtain a second additional measurement profile (S G,MEAS (V d )), - By using the spectral quantum efficiency (Q R ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the second calibration spectrum (I CAL2 ) to determine the first additional simulation profile (S R,SIM,k ) - By using the spectral quantum efficiency (Q G ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the second calibration spectrum (I CAL2 ) to determine the second additional simulation profile (S G,SIM,k ), - for the first calibration data (f k , g k ,h k ) is modified until the additional simulation profile (S R,SIM,k , S G,SIM,k ) according to one or more matching criteria and the corresponding additional measurement profile (S R,MEAS (V d ), S G,MEAS (V d ))match.

7. The method according to any one of claims 1 to 6, comprising: - After using the modified first calibration data (f k , g k ,h k ), according to the quantum efficiency (Q R , Q G , Q B ) and according to the spectral transmittance function (T FP,PRE (λ,V d )) to form a matrix element (U) of a first matrix (M), wherein the matrix element (U) of the first matrix is ​​used to convert the detector signals (S R , S G , S B ) is determined as the spectral intensity value (X) of the input light (LB1) at the spectral position (λ0, λ1, λ2) of the spectral transmittance peak of the Fabry-Perot interferometer (FPI1) n , X n+1 , X n+2 ), - Forming the second matrix So that the second matrix is the inverse or pseudo-inverse of the first matrix (M), and - The second matrix The matrix elements (K) are output and / or stored in a memory (MEM2).

8. The method according to claim 7, comprising: - forming an image (IMG2) of the object (OBJ1) on the image sensor (SEN1) by coupling input light (LB1) received from the object (OBJ1) to the spectral imaging device (500), - Obtaining a detector signal (S) from the image sensor (SEN1) R , S G , S B ), - By using the second matrix The matrix elements (K) of R , S G , S B ) to calculate one or more calibrated spectral intensity values ​​(X n , X n+1 , X n+2 ).

9. The method according to any one of claims 1 to 8, wherein: The distance (d) between the mirrors (M1, M2) of the Fabry-Perot interferometer (FPI1) F ) is adjustable, wherein the image sensor (SEN1) comprises a plurality of first detector pixels (P0) and a plurality of second detector pixels (P1), the detector pixels (P0, P1) being arranged to detect light (LB1) transmitted through the Fabry-Perot interferometer (FPI1), the first detector pixels (P0) having a first spectral quantum sensitivity (Q R (λ)), the second detector pixel (P1) has a different second sensitivity function (Q G (λ)), the Fabry-Perot interferometer (FPI1) ​​has a first transmittance peak (FPEAK0) and a second transmittance peak (FPEAK1), and the first detector pixel (P0) is arranged to detect light (LB1) at a spectral position (λ0) of the first transmittance peak (FPEAK0) and a spectral position (λ1) of the second transmittance peak (FPEAK1).

10. The method according to any one of claims 1 to 9, comprising: By scanning the wavelength (λ) of monochromatic light (LB3) transmitted through a representative Fabry-Perot interferometer (FPIO) LB3 ) to the initial spectral transmittance function (T FP,PRE (λ,V d )) is measured, wherein the initial spectral transmittance function (T FP,PRE (λ,V d )) comprises measuring the control parameter (V d ) is scanned and the intensity of the monochromatic light (LB3) is recorded as the wavelength (λ LB3 ) and the control parameter (V d ), wherein the initial spectral transmittance function (T FP,PRE (λ,V d )) Thereafter, the Fabry-Perot interferometer (FPI1) ​​of the spectral imaging device (500) is combined with the image sensor (SEN1).

11. A device (1000) for calibrating (500) a spectral imaging device, the spectral imaging device comprising a combination (CMB1) of a Fabry-Perot interferometer (FPI1) ​​and an image sensor (SEN1), The device (1000) comprises: - a communication unit (RXTX2) for obtaining first spectral data (DATA1) specifying the spectral quantum efficiency (Q) of the detector pixels (P0, P1, P2) of the image sensor (SEN1) R , Q G , Q B ), - a communication unit (RXTX2) for obtaining second spectral data (DATA2) specifying a spectral transmittance function (T FP,PRE (λ,V d )), a first calibration light source (LS1) for forming a first calibration light (LB1CAL1), the first calibration light having a first calibration spectrum (CALSPEC1), the first calibration spectrum comprising a plurality of spectral peaks (CALPEAK1, CALPEAK2, CALPEAK3), Wherein, the device (1000) is arranged as follows: - In the control parameters (V d ) is scanned by recording the first detector pixel signal (S R,MEAS ) to obtain a first measurement profile (S) of the first calibration light (LB1CAL1) R,MEAS (V d )), - In the control parameter (V d ) is scanned by recording the pixel signal (S G,MEAS ) to obtain a second measurement profile (S G,MEAS (V d )), - By using the spectral quantum efficiency (Q R ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) to determine the first simulation profile (S R,SIM,k ), - By using the spectral quantum efficiency (Q G ), by using the spectral transmittance function (T FP,PRE (λ,V d )) and by using the first calibration data (f k , g k ,h k ), according to the first calibration spectrum (I MPBF ) to determine the second simulation profile (S G,SIM,k ), - for the first calibration data (f k , g k ,h k ) is modified until the simulation profile (S R,SIM,k , S G,SIM,k ) according to one or more matching criteria and the corresponding measured profile (S R,MEAS (V d ), S G,MEAS (V d )) matches, and - The modified first calibration data (f k , g k ,h k ) is stored in the memory (MEM21).