Spectral segment computational reconstruction method, spectrometer, and device

Abstract

The present invention relates to the technical field of computational reconstructive spectrometers, and in particular, to a spectral segment computational reconstruction method, a spectrometer, and a device, which are used for performing spectral reconstruction on an object under test, of which a wavelength range falls within a set interval X. The spectral segment computational reconstruction method comprises: performing, in order, computational reconstruction on spectral segments of set intervals x1, x2, …, xk, to respectively acquire spectral feature information sets y1, y2, …, yk of the object under test in corresponding wavelength ranges; and performing splicing and reconstruction on the spectral feature information sets y1, y2, …, yk of the corresponding wavelength ranges, to obtain complete spectral feature information Y of the wavelength range of the object under test in the set interval X. The method can greatly reduce the computing power requirements on a computing device and reduce the bandwidth requirements on a single light source, so that a spectral sensing system can be further miniaturized and low in cost. In addition, customization and non-continuous spectrum analysis are facilitated, and product power consumption is reduced.

Description

Spectral segmentation calculation and reconstruction method, spectrometer and equipment Technical Field

[0001] The present invention relates to the field of spectrometers, and in particular to a spectrum segmentation calculation and reconstruction method, a spectrometer and a device. Background Art

[0002] The computational reconstruction spectrometer uses N different optical filters to sample the input spectrum, and then uses an ultra-wideband photodetector to obtain the photocurrent value after each spectral sampling. Finally, the input spectrum is reconstructed by solving the underdetermined equation group consisting of "input spectrum matrix, optical filter spectral response matrix, and photocurrent matrix".

[0003] The basic idea of ​​a computational reconstruction spectrometer is to obtain spectral information at M wavelength points (M>N) through N spectral samplings. This allows more spectral information to be obtained with fewer samplings, thereby reducing the requirements on the device, making the device smaller and more cost-effective, and thus having huge advantages in various portable, low-cost spectral analysis application scenarios.

[0004] In practical applications, it is often desirable to obtain spectral information with an ultra-wide spectral range and high wavelength resolution. In the spectral reconstruction calculations for such ultra-wide spectra, the value of M is often very large (e.g., over 1000). In some scenarios, the computing power resources of computing devices are limited, which severely restricts the application of the algorithm and hinders its integration into lower-cost, lower-power processors.

[0005] Furthermore, ultra-broadband light sources capable of simultaneously providing spectra spanning hundreds of nanometers are often bulky, making them difficult to integrate into small, low-cost devices. Furthermore, in most spectral analysis applications, only spectral data from certain wavelengths is required. Using a continuous spectrometer with such an ultra-broadband spectrum for detection would waste power in redundant wavelengths. This severely limits the battery life of small devices or equipment, hindering their integration.

[0006] Summary of the Invention

[0007] The present invention aims to address, to a certain extent, one of the technical problems in the related art. To this end, the present invention provides a spectral segmentation calculation and reconstruction method, a spectrometer, and a device. These methods enable segmented calculation and reconstruction, reducing the computing power requirements of the device and the power consumption of the light source in redundant bands. This facilitates the integration of the spectrometer into lower-cost, lower-computing-power devices, and facilitates the miniaturization of the device.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions in the first aspect:

[0009] A spectrum segmentation calculation and reconstruction method is used to reconstruct the spectrum of the detection object whose wavelength range falls within the set interval x, and the non-empty true subsets of the set interval x are the set intervals x1, x2, ..., x k And the set interval x=x1∪x2∪…∪x k ; where the set intervals x1, x2, ..., x k The corresponding wavelength ranges are all continuous;

[0010] The spectrum segmentation calculation and reconstruction method comprises:

[0011] Orderly the set intervals x1, x2, ..., x k The spectrum of the detection object is segmented and reconstructed to obtain the spectrum feature information set y1, y2, ..., y k ;

[0012] The spectral feature information set y1, y2, ..., y corresponding to the wavelength range k Perform splicing and reconstruction to obtain the complete spectral feature information Y of the wavelength range of the detection object in the set interval x.

[0013] Compared to existing technologies, the present invention significantly reduces computing power and enables faster results by dividing the broad spectrum into multiple bands and performing computational reconstruction on each segment (band). After obtaining the spectral response of the test object in that band, all bands are combined to obtain the complete spectral signature information of the test object across the broad spectrum. Although multiple computational reconstructions are required after segmentation, the number of wavelength points required for a single computational reconstruction is reduced from thousands to a fraction or even one-tenth of the original number, significantly reducing the computing power required.

[0014] In the present invention, it is further preferred that the spectral width corresponding to the set interval x is greater than 200 nm.

[0015] In the present invention, it is further preferred that the spectral width corresponding to the set interval x is greater than 500 nm.

[0016] In the present invention, it is further preferred that the spectral width corresponding to the set interval x is greater than 1000 nm.

[0017] It is further preferred that in the present invention, at least two segments in the wavelength range of the set interval x are discontinuous.

[0018] The present invention is further preferably, in the set intervals x1, x2, ..., x k When calculating and reconstructing the spectrum on any interval segment, first calculate and reconstruct the spectrum of a band range covering the interval segment, and then cut and retain the spectrum corresponding to the interval segment.

[0019] The present invention is further preferably such that in the set intervals x1, x2, ..., x k Among the corresponding band lengths, the smallest band length is at least half of the largest band length.

[0020] The present invention is further preferably, in the set intervals x1, x2, ..., x k When the spectra are calculated and reconstructed separately, the wavelength resolution is below 0.5nm.

[0021] In a second aspect, the present invention further provides a spectrometer, comprising:

[0022] A broadband light source, comprising a plurality of light sources in different wavelength ranges and capable of operating independently of one another, wherein all light sources, when operating simultaneously, provide a spectrum having a wavelength range covering a set interval X;

[0023] Wide spectrum response filter, used for the set interval x1, x2, ..., x k When calculating and reconstructing the spectrum of the segmented spectrum, the corresponding set intervals x1, x2, ..., x are obtained. k The corresponding spectral response matrix;

[0024] a wide-spectrum light detector, receiving the spectral response matrix and converting it into a corresponding electrical signal; and

[0025] The data processing module is used to receive the electrical signal, perform calculation and reconstruction in sequence according to each segment, and splice the spectral feature information reconstructed in each segment after completing the segment reconstruction.

[0026] By setting up multiple light sources in different bands that can work independently of each other, the dependence on ultra-wide spectrum light sources is avoided. In actual production, the cost of producing ultra-wide spectrum light sources is very high and the power consumption is high; non-continuous band detection can be performed according to application requirements, avoiding the waste of power consumption of the light source in redundant bands.

[0027] It is further preferred that the wide spectrum response filter has multiple optical channels, and the spectral response matrix is ​​established by different transmission spectra output by the multiple optical channels.

[0028] It is further preferred that the wide spectrum response filter is a tunable filter, which can form multiple optical channels in time sequence through tuning to output different transmission spectra to establish the spectral response matrix.

[0029] The present invention is further preferably provided with a bandpass optical filter between the broadband light source and the broadband response filter; k When the segments are reconstructed by calculation, the bandpass optical filter is used to selectively transmit the spectrum corresponding to each segment.

[0030] It is further preferred that the present invention is provided with a plurality of band-pass optical filters, and the plurality of band-pass optical filters have different transmission bands and are respectively arranged on corresponding light sources.

[0031] In a third aspect, the present invention further provides an embedded computer device comprising a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program, implements the spectral segmentation calculation and reconstruction method described in the first aspect. The reasoning process for the beneficial effects of the embedded computer device provided by the present invention and the spectral segmentation calculation and reconstruction method described above is similar and will not be further described here.

[0032] These features and advantages of the present invention will be further disclosed in the following detailed description and accompanying drawings. The preferred embodiments and means of the present invention will be fully illustrated in conjunction with the accompanying drawings, but are not intended to limit the technical solutions of the present invention. Furthermore, although multiple features, elements, and components may be present in each of the following text and accompanying drawings, they may be labeled with different symbols or numbers for convenience, but all represent components with the same or similar structure or function. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly introduced below. Obviously, the drawings in the following description only relate to some embodiments of the present invention, but are not intended to limit the present invention.

[0034] FIG1 is a schematic structural diagram of the spectrometer of the present invention.

[0035] Figure 2(a)-Figure 2(e) are schematic diagrams of the continuous segmented spectrum reconstruction process based on the LED light source combination.

[0036] FIG3(a)-FIG3(d) are schematic diagrams of the discontinuous segmented spectrum reconstruction process based on the LED light source combination.

[0037] FIG4 is a comparison of the spectral characteristics of an LED light source before and after passing through a bandpass optical filter.

[0038] FIG5(a)-FIG5(e) are schematic diagrams of the continuous segmented spectrum reconstruction process after the LED light source is combined with the bandpass optical filter.

[0039] Wherein: 100, broadband light source; 200, broadband response filter; 300, broadband light detector; 400, sample to be tested. DETAILED DESCRIPTION

[0040] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described in the embodiments are intended to explain the present invention and are not to be construed as limiting the present invention.

[0041] Reference in this specification to "one embodiment," "an example," or "an example" means that a particular feature, structure, or characteristic described in connection with the embodiment itself can be included in at least one embodiment of the present patent disclosure. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

[0042] To facilitate understanding of the technical solution, the application scenarios of the computational reconstruction spectrometer (or computational reconstruction spectrometer) and its working principle are briefly explained first.

[0043] In applications such as biosensing and material analysis, accurate analysis of a substance's content requires spectral data spanning hundreds of nanometers. Analyzing the content of multiple substances often requires spectral information spanning over 1,000 nanometers. Spectral information spanning hundreds of nanometers or even over 1,000 nanometers falls within an ultra-wide spectral range.

[0044] Traditional spectral characteristic detection methods are usually achieved through swept lasers or spectrometers.

[0045] The working principle of the swept laser method is to output a single wavelength of the laser, obtain the scattering or absorption information of a single wavelength by irradiating the object to be measured and collecting the scattered or transmitted light, and then use the sweep function of the swept laser to scan each wavelength to obtain the characteristic information of the entire spectrum. This method requires scanning wavelength by wavelength. When the number of wavelength points required is large, the working time will be greatly increased. On the one hand, the time cost increases, and on the other hand, it is not suitable for certain samples that change rapidly. The price of swept lasers is also relatively expensive, and the size and mass are large. If the wavelength range of the required spectral information is large, a single swept laser may not be able to meet the application requirements, which will further increase the cost, size and mass.

[0046] The basic idea of ​​the computational reconstruction spectrometer is to obtain spectral information of M wavelength points through N spectral samplings (M>N). In this way, a larger amount of spectral information can be obtained through fewer samplings, which reduces the requirements for the device, making the device smaller and more cost-effective, and thus having huge advantages in various portable, low-cost spectral analysis application scenarios.

[0047] The specific operating principle of a computational reconstruction spectrometer is as follows: a light signal of unknown spectrum first passes through an optical filter with ultra-wide spectral response, then enters an optical detector with ultra-wide spectral response, generating a photocurrent value I1. This can be understood as the filter performing a full spectrum sampling of the input light signal, and then integrating and summing the optical power at each wavelength. Similarly, passing through N different optical filters with ultra-wide spectral response results in N full spectrum samplings. Therefore, a matrix equation can be obtained:

[0048] make

[0049] Here, the F matrix represents the spectral response of N optical filters (here, an optical filter with N single optical channels is used; a single filter with N optical channels could also be used). Each row represents an optical filter, and the number of columns, M, represents the number of wavelengths. The Φ matrix represents the spectral information of the input optical signal, and the I matrix represents the photocurrent data of the photodetector. For simplicity, the photoresponsivity of the photodetector is assumed to be 1 at all wavelengths, and the losses of other optical components in the optical link are not considered.

[0050] Here, the optical filter's spectral response matrix, F, and the photocurrent matrix, I, are measurable. Since M > N, the spectral information of the input optical signal is reconstructed by solving this underdetermined system of M unknowns and N equations. Because input optical signals outside the wavelength window of the matrix calculation are not considered, but are often within the photodetector's response wavelength range, to minimize reconstruction errors, the energy of the input optical signals outside the calculation wavelength window must be sufficiently low.

[0051] The following describes the calculation steps of an existing computational reconstruction spectrometer in conjunction with a specific scenario: In the analysis of the content of a multi-component substance, spectral information with a spectral width of 1000nm and a wavelength resolution of 0.2nm is required. Using an ordinary computational reconstruction spectrometer, it is necessary to measure the light intensity information of 5000 different wavelength points. In a computational reconstruction spectrometer, the number of wavelength points is equal to the number of unknowns to be solved. To solve 5000 unknowns at one time, the number of equations required often reaches more than 100. Faced with such a large amount of calculation, even based on a personal computer platform, its calculation and solution time often reaches several minutes, not to mention the low-power processor of smart wearables. The inventors found that if the number of wavelength points to be solved can be controlled below 1000, it only takes a few seconds based on the same processor platform; when the number of wavelength points is controlled below 500, the calculation and reconstruction time is less than 1 second.

[0052] Based on the above situation, the present invention proposes a spectral segmentation calculation and reconstruction method for spectral reconstruction of detection objects whose wavelength range falls within the set interval X. The spectral width corresponding to the set interval X is above 500nm or 1000nm, and the wavelength resolution of the required spectral width is below 0.5nm. Therefore, it is necessary to measure the light intensity information of at least 1000 wavelength points.

[0053] The non-empty proper subsets of the set interval X are the set intervals x1, x2, ..., x k And the set interval X=x1∪x2∪…∪x k ; where the set intervals x1, x2, ..., x k The corresponding wavelength ranges are continuous.

[0054] The spectrum segmentation calculation and reconstruction method mainly includes two steps. The first step is to sequentially calculate the set intervals x1, x2, ..., x k The spectrum of the detection object is segmented and reconstructed to obtain the spectrum feature information set y1, y2, ..., y k It should be noted that the calculation and reconstruction method here is an existing technology, which can be referred to in the previous article. The only difference here is that the frequency band of the reconstructed light source is a small segment divided from the wide spectrum, rather than the entire wide spectrum, and multiple calculations and reconstructions are required to obtain the spectral feature information set on (different) segments.

[0055] The second step is to convert the spectral feature information set y1, y2, ..., y k Perform splicing and reconstruction to obtain the complete spectral feature information Y of the detection object wavelength range in the set interval X, Y = y1∪y2∪…∪y k .

[0056] It should be noted that, in an exemplary embodiment, the spectrum feature information sets y1, y2, ..., y k Before splicing and reconstruction, the spectral feature information sets y1, y2, ..., y k After optimization, the optimized spectral feature information set is y1', y2', ..., y k ', and finally the optimized spectral feature information set is y1', y2', ..., y k 'Perform splicing and reconstruction.

[0057] In the set interval x1, x2, ..., x kIn the above example, the length of a single interval determines the number of wavelength points that need to be measured when calculating and reconstructing the spectrum. As the number of points increases, the computational time increases exponentially. Therefore, it is best to divide the set interval X equally according to the computing power, spectral width, and resolution required by the computing device, and obtain a suitable number (i.e., the value of k) of non-empty true subsets as the set intervals x1, x2, ..., x k .

[0058] If the spectral width is required to be 1000nm and the resolution of the spectrum is required to be 0.2nm, then 5000 wavelength points need to be calculated. Assuming that 500 wavelength points is the optimal calculation time consumption of the computing device, it can be determined that the value of k is 10, and the set intervals x1, x2, ..., x 10 The length of each interval is 100 nm.

[0059] If the spectral width is required to be 500nm and the resolution required for the spectrum is 0.5nm, then 1000 wavelength points need to be calculated. Assuming that 500 wavelength points is the optimal calculation time consumption of the computing device, it can be determined that the value of k is 2, and the length of each interval in the set intervals x1 and x2 is 250nm.

[0060] After reconstructing the spectrum of the sample to be tested by calculating it one by one according to the segmentation of different set intervals, the spectral feature information obtained in each segment is finally spliced ​​in the frequency domain in order to obtain the complete spectral feature information of the wavelength range of the detection object in the set interval X. The time consumption of the splicing process is basically negligible.

[0061] From the perspective of computing power requirements or computational time, the segmented computational reconstruction method helps to realize the application of computational reconstruction spectrometers in low-cost, low-power devices.

[0062] The following will describe in detail the specific implementation process of the spectrum segmentation calculation and reconstruction method on spectrometers with different structures in conjunction with specific embodiments.

[0063] Example 1:

[0064] As shown in FIG1 , a spectrometer is shown for obtaining spectral characteristic information of a sample 400 to be tested using a spectral segmentation computational reconstruction method. The spectrometer includes a broadband light source 100, a broadband response filter 200, a broadband light detector 300, and a data processing module. The spectrum generated by the broadband light source 100 is detected by the broadband response filter 200 and then received by the broadband light detector 300. The sample 400 to be tested can be placed between the broadband light source 100 and the broadband response filter 200, or between the broadband response filter 200 and the broadband light detector 300. Both are acceptable, and this embodiment is not limited thereto (the figures only illustrate one embodiment of the sample to be tested as an example for illustration).

[0065] The broadband light source is used to provide a spectrum with a wavelength range that can cover the set interval X. Currently, ultra-broad spectrum light sources that can simultaneously provide spectra above hundreds of nanometers often occupy a large volume, making it impossible to integrate them into small-sized, low-cost devices. Even if they are integrated into small-sized devices regardless of cost, they have to cover such a wide frequency band when working, so the power consumption is relatively large, which seriously affects the battery life of small-sized devices. Therefore, the broadband light source in this embodiment is composed of a plurality of light sources located in different wavelength ranges and that can work independently of each other. Light sources in different wavelength ranges only work when the segment is calculated and reconstructed, so the power consumption of the light source in the non-working band can be effectively reduced, while reducing the device temperature and heat dissipation requirements, thereby improving the battery life of small-sized devices.

[0066] The wide spectrum response filter is used to select the set interval x1, x2, ..., x k When calculating and reconstructing the spectrum of the segmented spectrum, the corresponding set intervals x1, x2, ..., x are obtained. k The corresponding spectral response matrix.

[0067] Specifically, the wide spectrum response filter has multiple optical channels in space, and the spectral response matrix is ​​established through different transmission spectra output by the multiple optical channels.

[0068] It should be noted that, in another feasible alternative, the wide spectrum response filter may also be a tunable filter, which can form multiple optical channels in time sequence through multiple tunings to output different transmission spectra to establish the spectral response matrix.

[0069] The broadband optical detector receives the spectral response matrix and converts it into a corresponding electrical signal. In this embodiment, a single broadband optical detector is used; a single optical detector can be shared by multiple filters (of a single optical channel), or a single optical detector can be shared by multiple optical channels in the time domain of a filter. It should be noted that, in addition to the above scheme, a broadband optical detector can be connected after each single-channel filter, or after each optical channel of a filter with multiple optical channels, to achieve the same effect.

[0070] The data processing module is used to receive the electrical signal, perform calculation and reconstruction according to each segment, and after completing the segment reconstruction, splice the spectral feature information reconstructed in each segment. The data processing module is a central processing unit, microcontroller, microprocessor chip or computing system with computing functions.

[0071] In this embodiment, multiple broadband light sources of different bands are turned on in chronological order. While a broadband light source of a certain band is turned on, the existing computational reconstruction method is used to computationally reconstruct the band to obtain the spectral feature information of the band. After the spectral feature information of all bands is obtained, it is spliced ​​together to obtain the complete spectral feature information of the detection object on the ultra-wide spectrum.

[0072] The spectrometer adopts a segmented computational reconstruction method, which can significantly reduce the computing power requirements of the ultra-wide spectrum computational reconstruction spectrometer, and also reduces the bandwidth requirements of a single light source, making the spectral sensing system more compact and low-cost.

[0073] In addition, since the spectral absorption characteristic bands of different substances are not continuous, and even the spectral bands required for the analysis of the same substance are not continuous, if a spectrometer using a segmented calculation and reconstruction method is used, it can save the device cost, device power consumption and computing power waste of redundant bands; it helps to realize customized and non-continuous spectral analysis and save product power consumption. For details, see Example 3.

[0074] Finally, portable devices lack ultra-broad spectrum light sources exceeding several hundred nm. For example, the bandwidth of near-infrared LED light sources is only around 100 nm, and that of near-infrared SLD light sources is only a few tens of nm. If spectral analysis of over several hundred nm is required for one-time reconstruction, this means that several, or even more than ten, light sources will need to operate simultaneously. This would require enormous instantaneous power consumption, generate significant heat, and thus degrade device performance, hindering the implementation of low-power devices. However, if a segmented computational reconstruction spectrometer is used, these light sources of different wavelengths can operate sequentially in chronological order, significantly reducing instantaneous power consumption. This would make the power management module circuits for low-power devices such as portable wearables easier to implement and miniaturize.

[0075] In summary, the idea of ​​using light sources of multiple different bands and segmented computational reconstruction can solve the problems of computing power and power consumption in portable spectral analysis application scenarios, and can also reduce the waste of resources in unnecessary spectral bands.

[0076] Example 2:

[0077] As shown in Figure 2 (a) to Figure 2 (e), another continuous segmented spectrum reconstruction process of a spectrometer is shown. The wide-spectrum light source of the spectrometer is based on a combination of LED light sources. In this embodiment, only four LEDs of different wavelengths are used as an example for illustration. The specific number of light sources can be determined according to actual conditions and is not limited. The wavelength range set interval [λ1, λ5] is the spectral range expected to be detected, and the four segments [λ1, λ2], [λ2, λ3], [λ3, λ4] and [λ4, λ5] are divided according to the required resolution of the spectrum, the computing power of the equipment and the wavelength range of the light source.

[0078] The dashed box in Figure 2(a) represents the desired spectral range for detection. The spectra of the four LEDs are shown as the four Gaussian-like solid lines in Figure 2(a), with some overlap. Each LED is illuminated one by one in a time-shaded manner. When LED1 is illuminated, the spectral range for this segment is reconstructed as shown in the shaded portion of Figure 2(b). The wavelength range corresponding to the shaded portion exceeds the dashed box. After reconstruction, only the data within the dashed box, i.e., the spectral range of the [λ1, λ2] segment, is retained. It is important to note that the reconstructed spectral range must be wide so that the power of the LEDs outside the reconstructed spectral range is sufficiently low to negligibly affect the reconstruction accuracy. The spectral range of the final retained data in the figure is smaller than the calculated spectral range because the error in the spectral reconstruction is larger at wavelengths with lower LED power. By illuminating LEDs 2-4 sequentially and repeating the above steps, the spectral response of the object to be detected in the remaining wavelengths is obtained. The remaining spectral ranges after four reconstructions are superimposed to form the desired continuous spectral response of the object to be detected from λ1 to λ5.

[0079] It should be noted that, for ease of understanding, in this embodiment, the spectral wavelength range provided by light source LED1 is exactly [λ1, λ2]. This range, divided according to factors such as computing power and resolution, is the wavelength range that needs to be calculated. However, in actual applications, due to factors such as manufacturing process and cost, the spectral wavelength provided by a single LED in a light source may be much smaller than [λ1, λ2] and insufficient to fully cover this range. Therefore, multiple light sources can also operate simultaneously to provide a spectrum that is just enough to cover [λ1, λ2]. Similarly, this situation may also exist in other wavelength ranges, and will not be discussed in detail here.

[0080] Example 3:

[0081] As shown in Figures 3(a) to 3(d), another discontinuous segmented spectrum reconstruction process of a spectrometer is shown. The wide spectrum light source of the spectrometer is based on a combination of LED light sources. In this embodiment, only three LEDs with different wavelengths are used as an example for illustration. The specific number of light sources can be determined according to actual conditions and is not limited. The wavelength range [λ x1,λ x3 ]∪[λ y1 ,λ y2 ] is the spectral range expected to be detected, [λ x1 ,λ x2 ]、[λ x2 ,λ x3 ] and [λ y1 ,λ y2 ] These three segments are divided according to the required spectral resolution, device computing power and light source wavelength range. x1 ,λ x2 ] and [λ x2 ,λ x3 ] is divided from a continuous band, [λ y1 ,λ y2 ] is another separate band, [λ x1 ,λ x3 ] and [λ y1 ,λ y2 ]The two bands are discontinuous.

[0082] The dotted box in Figure 3(a) shows the spectral range that is expected to be detected. The spectra of the three LEDs are shown as the three Gaussian solid lines in Figure 3(a). The combination of LED1 and LED2 achieves λ x1 to λ xi Spectral detection of the band range, LED3 completes λ y1 to λ yj Spectral detection of wavelength range. The wavelength bands of each LED when being detected are shown in Figure 3(b) to Figure 3(d). The meaning of the diagrams is similar to that of Figure 2 and will not be repeated here.

[0083] The spectral characteristics of LED light sources are similar to Gaussian, and the sideband roll-off is very slow. In single spectrum reconstruction, the wavelength range required to be calculated is much larger than the half-maximum spectral width of the LED light source, which increases the computing power requirements of the device.

[0084] To reduce the computational wavelength window, a bandpass optical filter can be placed on the LED light source to perform a bandpass filtering on the LED spectrum. This allows the spectrum corresponding to each segment to be selectively transmitted through the bandpass optical filter during computational reconstruction of each segment, thus reducing the computational wavelength window. This bandpass optical filter can be implemented using a coated optical filter. The coated optical filter can be placed on the LED light source or behind the LED. Coated optical filters with different transmission bands are placed on LEDs of different wavelength bands. The spectral characteristics before and after filtering are shown in Figure 4. Continuous segmented spectral reconstruction based on the bandpass filtered LED light source is shown in Figures 5(a)-5(e), which qualitatively illustrate that the computational wavelength window for each spectral reconstruction can be significantly reduced.

[0085] This embodiment also provides an embedded computer device, including a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor performs the above-described spectrum segmentation calculation and reconstruction method. The embedded computer device can be a smart wearable device such as a smart watch or smart bracelet, which integrates the spectrometer of the present invention. The embedded computer device can also be a portable computing and reconstruction spectrum detection instrument.

[0086] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. Accordingly, the computer program can be stored in a non-volatile computer-readable storage medium, and when the computer program is executed, it can implement the method of any of the above-mentioned embodiments. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0087] The above are only specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Those skilled in the art should understand that the present invention includes but is not limited to the contents described in the drawings and the above specific embodiments. Any modifications that do not deviate from the functional and structural principles of the present invention are intended to be included within the scope of the claims.

Claims

1. A spectrum segmentation calculation and reconstruction method, characterized in that: It is used to calculate the spectrum reconstruction of the detection object whose wavelength range falls within the set interval X in the reconstruction spectrometer. The non-empty true subsets of the set interval X are the set intervals x1, x2, ..., x k And the set interval X=x1∪x2∪…∪x k ; where the set intervals are x1, x2, …, x k The corresponding wavelength ranges are all continuous; the total number of wavelength points on the collection interval X is determined according to the wavelength resolution; the collection intervals x1, x2, ..., x k The length of a single interval is determined by the total number of wavelength points; The spectrum segmentation calculation and reconstruction method comprises: Orderly set intervals x1, x2, ..., x k The spectrum is segmented and reconstructed to obtain the spectral feature information sets y1, y2, ..., y1 of the detection object in the corresponding wavelength range. k ; The spectral data used in calculating and reconstructing a single set interval is related to the number of wavelength points corresponding to the set interval; the spectral feature information set y1, y2, ..., y k is the light intensity information related to the number of wavelength points in the corresponding interval; The spectral feature information set y1, y2, ..., y corresponding to the wavelength range k Perform splicing and reconstruction to obtain the complete spectral feature information Y of the wavelength range of the detection object in the set interval X.

2. The spectrum segmentation calculation and reconstruction method according to claim 1, characterized in that: The spectral width corresponding to the set interval X is above 200 nm, 500 nm or 1000 nm.

3. The spectrum segmentation calculation and reconstruction method according to claim 1 is characterized in that: At least two segments in the wavelength range of the set interval X are discontinuous.

4. The spectrum segmentation calculation and reconstruction method according to claim 1, characterized in that: In the set interval x1, x2, ..., x k When calculating and reconstructing the spectrum on any interval segment, first calculate and reconstruct the spectrum of a band range covering the interval segment, and then cut and retain the spectrum corresponding to the interval segment.

5. The spectrum segmentation calculation and reconstruction method according to claim 1, characterized in that: In the set interval x1, x2, ..., x k When the spectra are calculated and reconstructed separately, the wavelength resolution is below 0.5nm.

6. A computational reconstruction spectrometer for obtaining spectral characteristic information of a sample to be tested, characterized in that: include: A broadband light source, wherein the broadband light source comprises a plurality of light sources in different wavelength ranges and capable of operating independently of each other, and when all the light sources operate simultaneously, they provide a spectrum with a wavelength range covering a set interval X; Wide spectrum response filter, used for the set interval x1, x2, ..., x k When calculating and reconstructing the spectrum segment by segment, we obtain the corresponding set intervals x1, x2, …, x k The corresponding spectral response matrix; A wide-spectrum optical detector receives the spectral response matrix and converts it into a corresponding electrical signal; as well as, A data processing module, used for receiving the electrical signal, performing calculation and reconstruction in sequence according to each segment, and splicing the spectral feature information reconstructed in each segment after completing the segment reconstruction; Among them, the set intervals x1, x2, ..., x k is a non-empty proper subset of the set interval X and x1∪x2∪…∪x k =X; set interval x1, x2, ..., x k The corresponding wavelength ranges are all continuous; the total number of wavelength points on the collection interval X is determined according to the wavelength resolution; the collection intervals x1, x2, ..., x k The length of a single interval is determined by the total number of wavelength points; The spectral data used in calculating and reconstructing a single set interval is related to the number of wavelength points corresponding to the set interval; the spectral feature information set y1, y2, ..., y k It is the light intensity information related to the number of wavelength points in the corresponding interval.

7. The spectrometer according to claim 6 is characterized in that The wide spectrum response filter has multiple optical channels, and the spectrum response matrix is ​​established through different transmission spectra output by the multiple optical channels.

8. The spectrometer according to claim 6 is characterized in that The wide spectrum response filter is a tunable filter, which can form multiple optical channels in time sequence through multiple tunings to output different transmission spectra to establish the spectral response matrix.

9. The spectrometer according to any one of claims 6 to 8, characterized in that: A bandpass optical filter is also provided between the broadband light source and the broadband response filter; k When the segments are computationally reconstructed, the bandpass optical filter is used to selectively transmit the spectrum corresponding to each segment.

10. The spectrometer according to claim 9, characterized in that: There are a plurality of bandpass optical filters, each of which has different transmission bands and is respectively arranged on a corresponding light source.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the spectral segmentation calculation and reconstruction method according to any one of claims 1 to 5 is implemented.

Images