FR3086056A1 - Spectrometer device and its management method - Google Patents

Abstract

Title: Spectrometer device and its management method Spectrometer device (10) comprising: an installation for spatial and / or temporal separation (1, 2) of wavelengths (λ1,…, λn) of an incident light ( L), a variable filter installation (3), receiving a separate light with wavelengths (λ1,…, λn) separated in space and / or over time, a detection installation (4) of a measurement quantity of the light of the filter installation (3), and comprising a spectral integration element (5) variable in time and / or in space for the measurement quantity generating an integration value of the measured variable, and a control installation (SE). Figure 1

Description

Description

Title of the invention: Spectrometer device and its management method

Field of the Invention The present invention relates to a spectrometer device comprising a separation installation for the separation in space and / or in time of the wavelengths of an incident light as well as a filter installation.

The invention also relates to a method for managing such a spectrometer device.

STATE OF THE ART [0003] In optical spectroscopy, the usual monochromators and spectrometers make it possible to analyze and characterize numerous samples. Recently, the trend is towards the miniarisation of such devices which one usually finds in laboratory, until the possibility of integrating such devices in the mobile phones. To collect spectral information in the incident light, we must basically distinguish two opposite solutions, which differ mainly in the number of detectors used. The first solution corresponds to a multiplexing in space, of information, followed by a socket with a multichannel detector. An example is the classic laboratory spectrometer, in which the incident light is dispersed by an array, which corresponds to a multiplexing in space; then it is detected using a multi-channel detector, for example, an Si imager (CCD or CMOS). In this case, we represent the incident information Ι (λ) by a function I (x).

Another possibility for extracting spectral information from incident light is to do time multiplexing, which is done with a monochromator and sequential detection, for example, with a single-channel detector. An example of a miniaturized spectrometer system based on this principle uses a tuned Fabry-Perot interferometer (FPI), followed by a photodiode. The FPI spectrometer constitutes an optically tuned filter which separates part of the incident spectrum and transmits it to a detector. Multiplexing is also done in time, that is to say that the incident information Ι (λ) gives a function I (t) and allows, by a calculation in the opposite direction, to obtain a spectrum by applying the relation known to (t). The systems and components based on this principle are, for example, described in the documents JP 2015165266 A, US 9329360, EP 2480867 AL These techniques typically use single-channel detectors and thus also apply in other spectral ranges than in the range that can be entered in silicon technique. This is how InGaAs photodiodes are used to detect light in the near infrared spectral range. Using only a single channel detector is also attractive because of the increased cost of silicon photodiodes.

Document WO 2015015493 A2 describes a spectrometer based on a linear variation filter based on the direction of random incidence of the incident light in an optical diffuser and also on the angularly dependent transmission in a linear path filter. The spectrally classified transmitted light is applied by an array of lenses to an imaging device on a chip.

Disclosure and advantages of the invention The present invention relates to a spectrometer device comprising: a separation installation for the spatial and / or temporal separation of wavelengths of an incident light, an installation filter with a filtering effect is variable, receiving a light separated by the separation installation with wavelengths separated in space and / or in time, a detection installation for detecting a quantity of measurement of light of the filter installation, the detection installation comprising an element of spectral integration variable in time and / or in space for the measurement quantity generating an integration value of the measurement quantity and an installation of control, which controls the separation installation, the filter device as well as the detection installation.

The invention also relates to a method of managing such a spectrometer device comprising the steps of providing a spectrometer device, separating in space and / or in time wavelengths of the incident light from the separation system, filter the light separated in space or over time by the filter system, detect a measured quantity of the light filtered by the filter system with the detection system and generate the integration value of the measurement quantity by the spectral integration element.

The basic idea of the present invention is that of a spectrometer device which can advantageously measure an integration value for a determined measurement quantity of light, apply a linear combination with any weighting of the lengths of wave of an incident light spectrum. Simply put, the measured value I is represented as follows (spectrum S, weighting G):

[OOW] I = f S (1) dL [0011] Advantageously, a linear combination of the wavelengths is generated and a corresponding measurement is made with the components (composition corresponding to components in the form of circuits) of the spectrometer device. In addition, we will have an improved and simplified use of spectrometric data. The spectrometer device is advantageously a spectrometer device in the general sense, that is to say a spectrometer device which is not necessarily limited to the establishment of a spectrum. Advantageously, the spectrometer device is used to determine the partial distribution corresponding to the different wavelengths (the intensities) of a determined distribution of the overall intensity.

According to the invention, the spectrometer device comprises, as indicated, a separation installation for making the separation in space and / or in time of the wavelengths of an incident light, a filter installation whose filter effect is variable; the filter installation receives the light separated in wavelengths by the separation installation in space and / or in time; the spectrometer device also includes an installation of detectors which detects a quantity of light measurement of the installation of filter; the detection installation comprises a spectral integration element, variable in time and / or in space for the measurement quantity; this installation generates an integration value of the measured quantity.

Finally, the spectrometer device comprises a control installation which controls the separation installation, the filter device and the detection installation.

The spectrometer device comprises or constitutes a microspectrometer.

The separation installation advantageously separates the incident light into different wavelengths; advantageously, there will be the emission of the different wavelengths at different points in space and the emission will be simultaneous and offset, that is to say in different directions. The filter installation makes it possible to receive these separate wavelengths in space, advantageously on the respective radiation path (separate path). The filter installation includes, for example, a neutral density filter installation. The filter installation advantageously includes an extension in space to receive the advantageously parallel light paths for the different wavelengths in zones offset in space to advantageously apply a predefined filtering effect. For each wavelength the filter effect generates a predefined weighting for the transmission of filtered light radiation according to the respective wavelength. The light incident with this electromagnetic spectrum can also be advantageously separated into different ranges of different wavelengths according to the intensities, in space and in time.

The separation installation makes it possible to pass the different wavelengths of the incident light, in a time-shifted manner, successively on the filter installation, however without separation in space, that is to say to say all the wavelengths or certain wavelengths of the incident light advantageously arrive by the same radiation path (in an identical position in space) successively on the filter installation. The filter is designed to perform a time-shifted weighting of the filtered intensities in the range of the corresponding wavelengths. The weighting and the separation are advantageously controlled by the control installation. The weighting can also be chosen different in each time interval. In addition, for a time-shifted application, a filter with a different weighting characteristic is used.

The measurement of an integration value of the weighted linear combinations of the wavelengths is advantageously based on the consideration that for an automatic exploitation of the measured spectra (for example reflection, transmittance and absorption spectra) for operation in the direction of chemometric / multivariant automatic learning, using a projection of the spectrum on the charges determined using the analysis of the main components, one can also form the integrals of such weighted linear combinations of the measured data. We can also consider all possible automatic learning algorithms which allow variables to be pre-selected and weighted.

Vis-à-vis the usual spectrometer devices there is the advantage of a reduction in the measurement time and thus the advantage of faster detection for the same signal / noise ratio or an increase in the signal ratio / noise for the same measurement time. In addition, it is advantageously possible to reduce the conditions applied to the circuits (components) concerning, for example, the necessary taking of data and the processing resulting in a lower current consumption and / or a less space of memory. In addition, it is advantageous to use simple detectors for all types of concept (embodiment or application of the spectrometer device) so that static mechanical elements on which is based multiplexing in the information space spectral and thus requiring a network of detectors can be applied in the spectral domains for which such networks of detectors are not available at an advantageous cost. Simple detectors make it possible to advantageously reduce the cost compared to detector networks, in particular in the infrared spectral range.

The spectral integration element can advantageously be variable in space and / or in time and can be advantageously applied to a separation in time or in space of wavelengths. In the case of a variation over time, the integration element only includes a program for adapting the integration time because the detector itself measures an integration value. Thus, the integration element can advantageously integrate the output signal over a given integration time. It is also possible that the integrating element can also integrate in space and / or in space-time. For this, the detection installation which can integrate the output signal supplied by the integration element over a given integration period, is advantageously followed by a focusing element which is placed at the focus of a collecting optic. , for example, a lens. Focusing makes it possible to concentrate, in a reduced space, wavelengths separated in space and advantageously filtered to then integrate. As a variant of the detection installation with a preset focusing element, the detection installation can itself have a large surface (extension) to receive separate wavelengths in space.

According to a preferred embodiment, the spectrometer device comprises a signal processing installation which processes the integration value.

The information contained in the spectra is used and allows conclusions to be drawn such as the nutrient content of a food, the identification of a material and the like. This is usually done with an algorithm using a chemometric method of data analysis, for example, a principal component analysis, a linear discrimination analysis followed by a regression or classification (and other machine learning method, for example , neural networks). Such processes extracted from the complete spectral information, the characteristic parts give information concerning this reduced part of the spectrum. Under these conditions, it is possible advantageously to renounce the use of the largest part of the spectrum received and to save measurement time, since, advantageously, we limit ourselves to the parts of the spectrum necessary for the algorithms which can be taken with a longer integration time instead of the whole spectrum with the ranges not concerned. In other words, instead of a multi-variant quantity, such as an intensity distribution with spectral resolution, the exploitation can be done only on a quantity of reduced dimensions, which makes it possible to avoid the additional dimensions.

For processing, the signal transmitted by the detection installation to the signal processing installation is processed by calculation, which is advantageously the measured integration value. This makes it possible to advantageously use the chemometric algorithms mentioned above which calculate from a multi-variant quantity, the measurement quantity sought using a result of the machine learning process. Also, the preparatory steps for more complex machine learning algorithms (eg, neural networks) often rely on dimension reduction methods (such as principal component analysis) applying the above method. These automatic learning processes are not done during the measurement, but before the measurement of quantities. From the results of the automatic learning previously obtained, we can then make a modified processing of the measurement data. The result of a machine learning operation can thus be copied directly into the circuit and use previous learning methods.

According to a preferred embodiment of the spectrometer device, the separation installation comprises a diffraction grating, a network of static Fabry-Perot filters or, in general, interference filters or a filter element to linear variation.

The diffraction grating, static Fabry-Perot filters or a linear variation filter element allow a separation in space of the wavelengths of the incident light. Advantageously, these are concepts of static mechanics (components).

According to a preferred embodiment of the spectrometer device, the separation installation comprises a Fabry-Perot filter with variable wavelengths or a diffraction grating rotating with a slot.

The Fabry-Perot variable wavelength filter or the diffraction grating rotating with a slit make it possible to make a separation in time of the wavelengths of the incident light. These are mechanically implemented concepts (components).

According to a preferred embodiment of the spectrometer device, the filter device comprises a gray filter element with a variable filter effect in space and a network of diaphragms of different dimensions.

According to a preferred embodiment of the spectrometer device, the filter device comprises a gray filter element with a time-varying filter effect with electrically controlled liquid crystals, a network of mirror projectors, a network of shutters MEMS or an electrically controlled LCD network.

The network of mirror projectors comprises, for example, a set of DLP devices (digital light processors). The expression network represents, within the framework of the present description, a field, a group, a layout each time a set of corresponding components.

According to a preferred embodiment, the spectrometer device is a filter device which is variable in time and / or in space.

According to a preferred development, the spectrometer device comprises an optical focusing element between the filter installation and the detector installation.

According to the invention, and as already described, the method for managing a spectrometer device consists in providing or using a spectrometer device according to the invention, to be separated in space and / or in time, wavelengths of the light incident by the separation installation, to filter the light separated in space and / or in time with the filter installation, detecting a measurement quantity of the light filtered by the filter installation by the detector installation and generate an integration value of the measurement quantity by the spectral integration element.

The method is advantageously distinguished by the implementation of the combination of the characteristics of the spectrometer device and the advantages and vice versa.

According to a preferred embodiment of the method, the integration element for the detector channels of the detector installation has different or identical weights of the measured variables to generate the integration value.

The weighting advantageously relates to the filter effect.

According to a preferred embodiment of the method, the measurement quantity comprises the intensity of the light in the wavelength is detected.

According to a preferred embodiment of the method, after the supply, a reference measurement is made with an incident light of known intensity and a light-shadow measurement, the spectrometer device having a reference light source. The reference measurement is advantageously done with the light reflected on an object irradiated by a known light source (reference light source) advantageously by the spectrometer device itself.

According to a preferred embodiment of the method, the integration value is processed by the signal processing installation and a machine learning method is calculated and applied.

The spectrometer device can itself represent the result of an automatic learning process and apply the learned step to the measurement signal.

Usually, the reflectivity R is calculated from the first three spectra measured according to which the reference measurement R is given by the following formula: R = (Iample-Shadow) / (Iref-Shadow), Iref advantageously represents the measurement reference, Isombre represents the measurement of the dark current, the sample corresponds to the intensity measured by a measurement made on the examined sample.

To determine the measurement value sought in an application (for example, the fat content of a cheese), it is possible to use, for the usual measurements, a scientific process with computer assistance:

Chemometric algorithms are used as supports which, starting from a multivariant quantity calculate the measurement quantity sought using a machine learning process. For this, we often apply a transformation of the axes of the multidimensional quantity (reflection spectrum) and we generally arrive at the value by a regression. Finally, one calculates the unknown value score (result) which corresponds to the measurement value sought or is in direct relation, for example, in proportional with this one, by a scalar multiplication of the spectrum of reflection and the vector of loading according to the Score = Normalized spectrum x loading vector relation: in the above embodiments it is advantageous to remove the scalar multiplication and use another significantly simpler calculation. The normalized spectrum is as follows: R = (Iample-Shadow) / (Iref-Shadow), this spectrum corresponds to the intensity referenced and at the bottom of the corrected spectrum, the reflexivity of a sample.

Without varying the measurement interval, the following calculation can be used:

Score = Normalized spectrum x loading vector: = ((S-Smoyen) / o-MC) x L, in this relation S represents the spectrum (in particular here R);

Smoyen is the average value of the spectrum;

Σ is the standard deviation, [0048] MC is the mean central vector and L is the loading vector.

For the measurement of the reflection, the following formula is advantageously applied: S = (I-Isombres) / (Iref-Isombres) = (R-Rsombres) / (Ref-Rsombres).

Substitute in the global equation (the scalar product already multiplied) we get [Score] = l / o SxL - l / σ Smoyen xL - MC xL = [0053] = l / o Z (IiLi / ( Iref, i-Isombres, i)) - l / o Z (Isombres, iLi / (Iref, i-Isombres, i)) [0054] -l / (oN) Zli / (Iref, i-Isombres, i) ZLj -l / (oN) Z Isombres, i / (Iref, i-Isombres, i)) Z (Ej- (MC xL)).

This result corresponds to scalar multiplication and the deletion of parentheses. There are advantageously here the four terms for calculating starting from the measurements: 1. operand as the sample with filter, [0058] 2. operand as Isombres with filter, [0059] 3. operand as sample without multiplied filter by the known filter transmission, and 4. operand as Isombre without filter multiplied by the known filter transmission.

In a simple case for which the measurement conditions can remain constant (distance between the spectrometer device and the object, ambient light, light source) the parameters Iref and Isombres are advantageously constant and as a first approximation, we can not not use normalization.

The signal is then calculated as follows:

Score = Normalized spectrum x loading vector = ((S-mean) -MC) xL, In this formula S = I = ZliLi-ZliZLj- (MC xL).

This advantageously gives the value sought by the addition of two measurements because the last operand is a known constant for a given chemometric model. If you want to have a variation of distance, you have to calculate at least approximately the standard deviation o. For this, we can perform measurement number 4 (also for the different wavelengths without a filter element). There is a huge advantage of time and cost especially in the application of the above embodiments to a Fourier transform spectrometer. Instead of a complex transformation calculation (Fourier transformation) we obtain the solution directly by a circuit.

Presentation of the drawings The present invention will be described below, in more detail with the aid of exemplary embodiments represented in the appended drawings in which:

[Fig-1] diagram of a spectrometer device according to an exemplary embodiment of the invention, [fig68] [fig.2] diagram of a spectrometer device according to another exemplary embodiment of the invention , [Fig.3] transmission curve of a filter installation as a function of wavelength and time, [fig.4] flowchart of an example of the method of the invention.

In the different figures we will use the same references to designate the same elements or similar function.

Description of the embodiment of the invention [0073] Figure 1 is a diagram of an exemplary embodiment of the spectrometer device according to the invention.

The spectrometer device 10 advantageously comprises a separation installation 1 for spatially separating wavelengths λΐ, ..., Λη 1 from the incident light L; a filter installation 3, the filter effect of which is variable; the filter installation 3 receives the light separated in its wavelengths λΐ, ..., λη in space by the separation installation 1; a detector installation 4 which detects a quantity of measurement of the light supplied by the filter installation 3; the detector installation 4 comprises a spectral integration element 5 which is variable in time and / or in space for the measurement quantity; this integration element generates an integration value of the measurement value. The installation also includes a control installation SE for controlling the separation installation 1, the filter installation 3 and the detector installation 4. In particular, the filter effect will be controlled at different points in space. predefined for filter installation.

The spectrometer device 10 advantageously comprises a signal processing installation 6 connected to the detector installation 4 (5) and which processes the integration value. The spectrometer device 10 further comprises an optical focusing element 7 between the filter installation 3 and the detector installation 4, which focuses the light intensities of the different separate wavelengths or ranges of wavelengths in a small area in space, which advantageously reduces the necessary size of the detector installation.

We can weight the different remote detection channels in space by a gray linear variation filter (filter installation). If this filter is additionally variably adjustable, we can have different weights. The right part of figure re presents the space transmission function (according to the extension (d) of the filter installation) of the filter as TA loading function. Depending on the wavelength, the intensity is individually weighted to extract characteristic information from the global spectrum. The light is then collected by the optical focusing element 7 (in general a focusing lens); alternatively also, using a concave mirror) the measurement can be made with a detector and advantageously extract only one parameter which corresponds to the measurement value. This avoids an additional complex extraction process to obtain the information.

Figure 2 is a diagram of the spectrometer device according to another embodiment of the invention.

FIG. 2 shows the spectrometer device 10 of FIG. 1 with, however, a separation installation 2 which is a time separation installation performing the separation in time of the wavelengths λΐ, ..., Λη incident light L (this means that the vertical representation corresponds to the course over time). Advantageously, in a limited spatial range; filter installation 3 has a variable filter effect over time.

The light is divided in time and not in space. The time axis is shown in the right image. The spectrum divided in time can be weighted by a filter adjustable in time (filter control) so that the detector can, here again, provide the sum of the intensities directly as the quantity sought. This eliminates or reduces the use of complex electronic operating procedures or algorithms / programs. The transmission function TA in this case advantageously varies as a function of time t.

Figure 3 shows the transmission of the filter installation as a function of wavelength and time.

The figure shows the transmission of a filter installation (at the top) as well as the transmission wavelength of the separation installation (below) as a function of the wavelength and the time.

In the case of a time variation of the filter effect of the filter installation, the transmission TA of the filter as a function of time t is done as shown in FIG. 3. In principle, it is possible to also consider a single constant filter element which itself has the transmission function as a function of the wavelength TA of the corresponding load. The use of a constant filter element has the advantage of not requiring modulation (control) of the filter and this, neither in time nor in space, but in requiring an exact design (self-forming) of such a filter element. In addition, in this case, the application can always be limited to a measurement variable and thus to an application case.

The installation of a variable filter (for example, the use of a liquid crystal polarizer) has the advantage of being able to carry out any number of application cases and thus possibilities of determining the measurement quantity in a component. Only the TA transmission function required for the respective measured variable will be saved and used.

A characteristic measurement can be made according to the diagram and be advantageously oriented according to the characteristic sequence of spectral reflection measurements with an integrated light source:

Determining Iref: first of all, a reference measurement is carried out with a measurement standard. This measure is often called a clear measure. This measurement guarantees that the reflection of the object to be measured will be normalized to the value 1. Often Teflon or Spectralon (registered trademark) is used as the measurement standard having a practically constant reflectivity equal to 1.

Determination of Isombres: then a measurement is made to determine the intensity of darkness / ambient intensity. This ensures that other influencing factors can be measured outside of the subject matter.

Determination of the sample: then according to the method of the invention, the measurements are carried out on the sample.

It may also be advantageous to measure certain simple wavelengths without a filter element (see the processing of the results by calculation.

At the end, the results are processed with each other by calculation.

Figure 4 shows the succession of process steps according to an exemplary embodiment of the invention.

According to the method for managing a spectrometer device, a spectrometer according to the invention is supplied (SI); the wavelengths of the incident light are separated in space and / or in time (S2) using the separation installation; the light separated in space or in time is filtered (S3) with the filter installation and a measurement quantity of the light filtered by the filter installation is detected (S4) using the installation of detectors and the integration value of the measurement variable is generated (S5) by the spectral integration element.

In summary, the invention relates to a spectrometer device comprising a separation installation for the spatial and / or temporal separation of wavelengths of an incident light, a filter installation whose filtering effect is variable, a light separated by the separation installation with wavelengths separated in space and / or in time arrives in the filter installation, a detection installation for detecting a measured quantity of light of the filter installation, the detection installation comprising an element of spectral integration variable in time and / or in space for the measurement quantity generating an integration value of the measurement quantity and an installation of control, which controls the separation installation, the filter device as well as the detection installation.

This spectrometer device includes a signal processing installation which processes the integration value.

The separation installation comprises a diffraction grid, a network of statistical Fabry-Pérot filters or a linear variation filter element.

The time separation installation (2) is a Fabry-Perot filter with variable wavelength, a diffraction grating rotating with a slot.

The filter device is a gray filter element with a spatially variable filter effect comprising an array of diaphragms of different dimensions.

The filter device is a gray filter element with a time-varying filter effect using an electrically controlled liquid crystal cell, a network of mirror projectors, a network of MEMS shutters or an electrically controlled LCD network.

The filter device is variable in space and / or in time.

The spectrometer device also includes an optical focusing element (7) installed between the filter installation (3) and the detection installation (4).

The method for managing this spectrometer device 10 consists in providing (SI) a spectrometer device (10), separating (S2) in space and / or in time, wavelengths of light incident (L) by the separation installation (1,2), filter (S3) the light separated in space or in time by the filter installation (3), detect (S4) a measurement quantity of the light filtered by the filter installation (3) with the detection installation (4) and, generate (S5) the integration value of the measurement quantity by the spectral integration element (5).

In this method, the integration element (5) for the detector installation detector channels (4) applies different or equal weights to the measured variables to generate the integration value.

The measurement variable comprising the intensity of the detected wavelength light.

According to another characteristic, after the supply (SI) a reference measurement is made with the incident light (L) of known intensity and a light-shade measurement, the spectrometer device (10) comprising a reference light source .

Finally, according to the method, the integration value is processed by a signal processing installation (6) and a machine learning method is calculated and applied.

[0105] NOMENCLATURE OF MAIN ELEMENTS [0106] 1 Separation installation [0107] [0108] [0109] [ΟΠΟ] [YES] [0112] [0113] [0114]

Separation installation

Filter installation

Detection facility

Spectral integration element

Signal processing facility

Optical focusing element

Spectrometer device

S1-S5 Process steps

Claims (1)

Claims [Claim 1] Spectrometer device (10) comprising: a separation installation (1,2) for the spatial and / or temporal separation of wavelengths (λΐ,.,., λη) of an incident light (L), a filter installation (3) including filtering effect is variable, the filter installation (3) receiving, a light separated by the separation installation (1, 2) with wavelengths (λΐ,.,., λη) separated in space and / or over time, a detection installation (4) for detecting a measurement quantity of the light of the filter installation (3), the detection installation (4) comprising a spectral integration element (5) variable in time and / or in space for the measurement quantity generating an integration value of the measurement quantity, and a control installation (SE), which controls the separation installation (1,2), the filter installation (3) as well as the detection installation (4). [Claim 2] Spectrometer device (10) according to claim 1, comprising a signal processing installation (6) which processes the integration value. [Claim 3] Spectrometer device (10) according to claim 1 or 2, comprising the separation installation (1, 2), a diffraction grid, a network of static Fabry-Perot filters or a filter element with linear variation. [Claim 4] Spectrometer device (10) according to one of claims 1 to 3, according to which the time separation installation (2) is a Fabry-Perot filter with variable wavelength, or a rotating diffraction grating with a slot. [Claim 5] Spectrometer device (10) according to one of claims 1 to 4, according to which the filter installation (3) is a gray filter element with a spatially variable filter effect comprising an array of diaphragms of different dimensions . [Claim 6] Spectrometer device (10) according to one of claims 1 to 5, according to which the filter installation (3) is a gray filter element with a time-varying filter effect with electrically controlled liquid crystal cells, a network of mirror projectors, a network of MEMS shutters or an electrically controlled LCD network. [Claim 7] Spectrometer device (10) according to one of claims 1 to 6, according to which the filter installation (3) is variable in space and / or in time. [Claim 8] Spectrometer device (10) according to one of claims 1 to 7, comprising an optical focusing element (7) between the filter installation (3) and the detector installation (4). [Claim 9] Method for managing a spectrometer device (10) comprising the following steps consisting in: supply (SI) a spectrometer device (10) according to one of claims 1 to 8, to separate (S2) in space and / or in time wavelengths of the incident light (L) by the separation installation (1, 2), to filter (S3) the light separated in space or over time by installing a filter (3), detecting (S4) a measurement quantity of the light filtered by the filter installation (3) with the detection installation (4), and generating (S5) the integration value of the measurement quantity by the element spectral integration (5). [Claim 10] Method according to claim 9, according to which the integration element (5) for the detector channels of the detection installation (4) applies different or equal weights to the measured variables to generate the integration value. [Claim 11] Method according to claim 9 or 10, that the measured quantity includes the intensity of the detected wavelength light. [Claim 12] Method according to one of claims 9 to 11, according to which after the supply (SI) a reference measurement is made with the incident light (L) of known intensity and a light-shade measurement, the spectrometer device (10) having a reference light source. [Claim 13] Method according to one of Claims 9 to 12, in which the integration value is processed by a signal processing installation (6) and a machine learning method is calculated and applied. 1/2