Spectral fitting method and system for a spectrally tunable light source

Abstract

This invention relates to the field of spectral analysis technology, and particularly to a spectral fitting method and system for a spectrally tunable light source. The system includes a host computer and a slave computer. The slave computer includes a light source unit, a spectral sensing unit, an optical path guiding unit, a standard reference unit, a measurement platform unit, and a radiation dose monitoring unit. The host computer includes a communication control module, a data storage module, a spectral modeling module, and a fitting decision module. In this scheme, by employing a staged, physically isolated spectral fitting control, the time-consuming drive search and iterative optimization process is entirely completed on the standard reference unit, with only a very small number of verification irradiations performed on the photosensitive object under test, thereby technically minimizing the light radiation interference to the object under test.

Description

Technical Field

[0001] This invention relates to the field of spectral analysis technology, and in particular to a spectral fitting method and system for a spectrally tunable light source. Background Technology

[0002] In fields such as spectral analysis, digital archiving of cultural relics, and precision optical measurement, spectrally tunable light sources serve as core equipment. They can simulate various standard lighting conditions or generate specific spectral curves, providing a high-fidelity lighting environment for the color, material, and other characteristics of object surfaces. Existing technologies typically employ multi-channel LED combinations and independently drive the intensity of each channel to adjust the output target spectrum. Such systems are widely used in general industrial or laboratory settings, enabling flexible spectral control within a certain range and utilizing spectral sensors for real-time feedback and closed-loop correction to meet the needs of routine measurements and reproduction.

[0003] However, when the aforementioned technologies are applied to special objects that are extremely sensitive to light radiation and require strict control of total exposure dose, such as precious cultural relics, ancient books, biological samples, or photosensitive materials, existing methods exhibit significant shortcomings. Traditional spectral fitting processes often require multiple trial irradiations and measurement iterations on the surface of the object under test to gradually approximate the target spectrum. This not only subjects the object to unnecessary cumulative light radiation, increasing its risk of photochemical damage, but also makes it difficult to perform real-time, weighted safety assessment and control of the dose during irradiation. Furthermore, conventional systems lack a mechanism to physically isolate the calibration process from the final irradiation, making it impossible to complete the time-consuming spectral search and optimization while protecting the object under test. Therefore, there is an urgent need for a spectrally tunable light source system and method that can intelligently control and minimize intervention in the exposure dose of photosensitive objects while ensuring spectral fitting accuracy, to solve the problem of balancing safety and accuracy in such highly sensitive application scenarios. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a spectral fitting method and system for a spectrally tunable light source, thereby solving the technical problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A spectral fitting system for a spectrally tunable light source includes a host computer and a slave computer: The lower-level machine includes: The light source unit consists of multiple light-emitting diodes (LEDs) with different peak wavelengths and corresponding independent constant current driving circuits. The light emitted by all the LEDs is uniformly mixed by an optical integrating sphere before being output. The spectral sensing unit is a fiber optic spectrometer, whose detection range covers the emission band of the light source unit; The optical path guiding unit includes a dual-position mirror module driven by a stepper motor for selectively guiding a beam of light from an optical integrating sphere to one of two fixed physical positions. The standard reference unit is a diffuse whiteboard with calibrated and known spectral reflectance, which is fixedly placed at the first of the two fixed physical positions. The test platform unit is used to place the photosensitive object to be tested, which is fixedly placed at the second position of the two fixed physical positions; The radiation dose monitoring unit includes a photodetector with a flat spectral response and signal processing circuitry for real-time monitoring and accumulation of exposure dose values. The host computer includes: The communication control module is used to communicate with the lower-level units, send drive and switching commands, and receive measurement and dosage data. The data storage module is used to store the reflectivity data of the standard reference body, the light source calibration data, the spectral basis space data, the driving vector-spectral coordinate mapping table, the preset safe dose thresholds for different types of photosensitive objects under test, and their corresponding material photodamage sensitivity weighting functions. The spectral modeling module is used to calculate the spectral power distribution of the light source and construct the spectral basis space based on the calibration data; The fitting decision module is used to execute the spectral fitting control logic in two physically isolated spaces in stages, the characteristic band feedback operation based on the single-validation irradiation spectrum, and the weighted dose safety decision based on the cumulative dose and sensitivity weight.

[0006] In one possible implementation, the feature band feedback operation in the fitting decision module involves selecting a feature band based on the reflectance spectrum measured by a single verification irradiation of the photosensitive object under test, and then back-estimating the approximate reflectance information of the photosensitive object under test in the selected feature band.

[0007] In one possible implementation, the weighted dose safety decision in the fitting decision module is to calculate the equivalent weighted dose based on the accumulated historical dose, the estimated new verification irradiation dose, and the material photodamage sensitivity weighting function, and to determine whether to allow the final irradiation by comparing it with a safety threshold.

[0008] In one possible implementation, the spectral modeling module constructs a spectral basis space by controlling each LED channel of the light source unit to individually illuminate the standard reference unit under different driving currents, calculating the absolute spectral power distribution under each state using the measured reflection spectrum and the known reflectivity of the standard reference unit as basis functions, and performing principal component analysis on all basis functions to construct a low-dimensional spectral basis space.

[0009] In one possible implementation, a spectral fitting method for a spectrally tunable light source includes: S1: Characterization of the light source system and construction of the spectral basis space: On the standard reference unit, the output of each LED channel under different driving currents is calibrated to obtain the basis functions and construct the spectral basis space; S2: Import target spectrum and map initial driving values: Project the target illumination spectrum onto the spectral basis space and search for the driving candidate set based on the pre-stored mapping table; S3: Closed-loop iteration and driver pre-optimization: On the standard reference unit, starting from the driver candidate set, through iterative measurement and adjustment, the output spectrum is made to approximate the target, and a pre-optimized driving vector is obtained; S4: Single-time verification irradiation and feature information extraction of the photosensitive object under test: The optical path guiding unit is controlled to switch the beam to the second position, and the photosensitive object under test is irradiated once with the pre-optimized driving vector. The reflection spectrum and cumulative dose are collected, and feature band information is extracted from them to estimate the approximate reflectance of the object under test. S5: Feature-based feedback-based drive correction: The reflectivity curve of the test object is estimated using the feature band information, the target illumination spectrum is recalculated, and the optical path guiding unit is controlled to switch the beam back to the first position. The correction drive vector is calculated on the standard reference unit. S6: Safety Decision and Final Output: Calculate the equivalent weighted dose, compare it with the safety threshold, make a decision based on the comparison result, and output the final driving scheme or forced termination process.

[0010] In one possible implementation, in step S3, the iterative measurement and adjustment is based on the difference between the spectrum measured on the standard reference body under the current driving vector and the target theoretical reflectance spectrum, as well as the local response characteristic matrix of the system near the current operating point. The driving vector is iteratively adjusted by solving a least squares problem until the spectral difference is less than a preset optimization threshold.

[0011] In one possible implementation, in step S4, the extraction of characteristic band information involves selecting 3 to 5 characteristic bands containing peaks, valleys, or inflection points from the measured full spectrum, and using the known illumination spectrum to back-estimate the approximate reflectance of the object under test in these bands.

[0012] In one possible implementation, in step S6, the equivalent weighted dose The calculation formula is: in, The energy distribution of the historical irradiation spectrum. To predict and verify the spectral energy distribution of the irradiation, This is the weighting function for the material photodamage sensitivity of the photosensitive object under test.

[0013] In one possible implementation, the method is used for spectral illumination or measurement of a photosensitive object to which the total exposure dose needs to be strictly limited.

[0014] Beneficial effects compared to existing technologies: 1. In this scheme, by employing a phased, physically isolated spectral fitting control, the time-consuming driving search and iterative optimization process is entirely completed on a standard reference cell. Only a very small number of verification irradiations are applied to the photosensitive object under test, thus technically minimizing the light radiation interference on the object. This method not only significantly reduces the cumulative dose suffered by the photosensitive object due to repeated trial exposures, avoiding the risk of light damage that may occur due to multiple irradiations in traditional fitting processes, but also significantly improves the matching efficiency and first-pass success rate of the final irradiation spectrum through pre-completed optimization calculations on the standard reference cell. This makes the system more practical in scenarios that are extremely sensitive to light, such as the digitization of cultural relics and the detection of precious samples. 2. In this scheme, by introducing a characteristic band feedback and weighted dose safety decision-making mechanism, key spectral features are extracted and reflectivity is estimated through a single verification irradiation. Combined with the material photodamage sensitivity weight, the equivalent dose is calculated in real time, realizing intelligent safety management in the spectral fitting process, and improving the active protection capability of photosensitive objects and the controllability of the irradiation process. 3. In this scheme, an intelligent mapping system between the spectral basis space and the driving vector is constructed, and spectral approximation is performed in conjunction with a local response model. This significantly improves the system's adaptability and reproduction accuracy for various target spectra, while reducing dependence on hardware performance and ensuring the efficiency and stability of the fitting process. Attached Figure Description

[0015] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0016] Figure 1 This is a schematic diagram of the framework of the spectral fitting system of the present invention; Figure 2 This is a schematic diagram illustrating the process of the spectral fitting method of the present invention. Detailed Implementation

[0017] Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention can also be implemented in various different forms, and therefore the present invention is not limited to the embodiments described below. In addition, for the purpose of more clearly describing the present invention, parts not connected to the invention will be omitted from the drawings. The technical solutions in this application are designed to address the problems described in the background, and are generally as follows: Please refer to Figure 1 and Figure 2 This embodiment introduces a spectral fitting method and system for a spectrally tunable light source. This system and method are specifically designed for high-fidelity spectral illumination or measurement of objects that are extremely sensitive to light radiation and require strict control of total exposure dose, such as in the digital archiving of precious paintings, calligraphy, and rare books in museums. Specifically, the entire spectral fitting process is decoupled into two physically isolated stages: In the first stage, all time-consuming drive value searches and optimization calculations requiring numerous trial irradiations are completed on an independent and repeatedly irradiated standard reference body; in the second stage, only the object under test undergoes minimal (usually 1 to 2) verification irradiations, maximizing information acquisition, combined with real-time exposure dose monitoring and intelligent decision-making throughout the process. This minimizes light interference on the photosensitive object under test while ensuring the required spectral accuracy for the application.

[0018] To implement the above method, this embodiment constructs a system in which an upper-level control platform and a lower-level execution mechanism work together. The lower-level machine contains all physical execution and sensing units, and its specific structure is as follows: The light source unit consists of multiple light-emitting diodes (LEDs) with different peak wavelengths. For example, it uses six LED channels with center wavelengths of 450 nm, 520 nm, 590 nm, 620 nm, 660 nm, and 730 nm. Each channel is powered by an independent constant current drive circuit that can be precisely controlled by digital signals. The light emitted by all LEDs is thoroughly and uniformly mixed through an optical integrating sphere to form a beam of output light with a programmable adjustable spectral shape.

[0019] Spectral sensing unit: This is a fiber optic spectrometer whose effective detection range must cover all emission wavelengths of the light source unit, for example, 380 nm to 780 nm. This unit is used to accurately measure the spectral power distribution of light signals reflected or directly received after irradiating the target surface.

[0020] Optical Path Guiding Unit: The core of this unit is a dual-position reflector module driven by a high-precision stepper motor. Based on received control commands, it precisely reflects and guides the light beam from the integrating sphere's exit port to one of two fixed physical positions.

[0021] The standard reference unit is a calibrated, highly stable diffuse reflective whiteboard whose spectral reflectance characteristics are known and constant within the system's operating wavelength range. This whiteboard is rigidly fixed to a guiding terminal position (e.g., position A) of the optical path guiding unit, ensuring its surface can be uniformly and repeatedly illuminated. Its precise spectral reflectance data is stored in the host computer as reflectance values ​​corresponding to discrete wavelength points.

[0022] Test object platform unit: Used to place the artifact or sample to be measured, located at another guiding terminal position of the optical path guiding unit (e.g., position B). The platform has the necessary fixing and positioning functions.

[0023] Radiation dose monitoring unit: This unit includes a photodetector and its signal processing circuit. The photodetector has a flat spectral response, and its output current is proportional to the total luminous flux of the incident light. This unit can perform real-time integration of the illumination signal, outputting a dose value proportional to the exposure energy, and also has an accumulation function.

[0024] The host computer serves as the control and computing center of the system, typically implemented by an industrial computer or a high-performance embedded controller. Its internal software is divided into the following functional modules: Spectral Modeling Module: Used to execute algorithms for analyzing and modeling the output characteristics of light source units.

[0025] Fitting Decision Module: This module contains the main control logic for running spectral fitting, optimization algorithms, and decision-making processes based on dose safety strategies.

[0026] Communication control module: responsible for managing the communication protocol between the host computer and all units of the slave computer, sending specific drive current commands and optical path switching commands, and receiving data from the spectral sensing unit and radiation dose monitoring unit.

[0027] Data storage module: Used to store all static and dynamic data required for system operation, including the reflectivity database of the standard reference body, light source calibration data, spectral feature vectors, preset safety dose thresholds for different types of cultural relics and corresponding material light damage sensitivity weighting functions, historical operation records, etc.

[0028] The host computer and the slave computer communicate with each other through a communication control module, and the host computer schedules the slave computer's various units through its various modules, thus forming the operation flow of the method, as follows: S1: Characterization of the light source system and construction of the spectral basis space This step is the system initialization and calibration process, using only the standard reference element. The communication control module first instructs the optical path guiding unit to stably point the beam at the standard reference element.

[0029] For each LED channel in the light source unit, the fitting decision module plans a series of drive current values ​​covering low, medium, and high intensities within its rated operating current range. For example, for an LED with a maximum current of 300 mA, seven calibration points can be selected: 0 mA, 50 mA, 100 mA, 150 mA, 200 mA, 250 mA, and 300 mA.

[0030] Subsequently, the system enters an automatic calibration loop: the communication control module controls the light source unit to sequentially illuminate only the currently calibrated individual LED channels to their planned current values. At each current point, after the light source output stabilizes, the communication control module triggers the spectral sensing unit to measure the reflection spectral signal on the standard reference unit. Upon receiving this signal, the spectral modeling module immediately retrieves the reflectivity data of the standard reference from the data storage module and, based on the principle of light reflection, calculates and converts the reflection spectral signal into the absolute spectral power distribution emitted by the LED under this specific driving current.

[0031] Specifically, let the measured reflectance spectral power distribution vector be... The spectral reflectance vector of the standard reference body at the corresponding wavelength point is: Then the absolute spectral power distribution vector emitted by the LED under this driving current The calculation formula is as follows: In this case, division involves dividing the vectors by their corresponding elements, and the result is... This refers to the spectral power distribution of the LED channel emitted independently under the current driving current, without reflection attenuation, and is also called a basis function.

[0032] Once all calibration points for one LED channel have been measured, the system automatically switches to the next LED channel and repeats the process until all calibration current points for all LED channels have been measured. Ultimately, the spectral modeling module obtains a raw database containing a large number of basis functions.

[0033] Next, the spectral modeling module performs principal component analysis on the original database of basis functions. Extraction The orthogonal principal component vector that best represents the spectral variation characteristics of the database is denoted as . Each of them Each is a principal component vector of the same dimension as the original spectral data. These principal component vectors span a low-dimensional spectral basis space. The output spectrum under any combination of driving forces generated by the system... (Here, the spectrum is considered as a vector), and can be linearly approximated by this set of orthogonal basis vectors. The mathematical relationship is as follows: Among them, coefficient The vector formed That is, the output spectrum The coordinates in the spectral basis space. The spectral modeling module will calculate all principal component vectors. And from raw spectrum to coordinates The projection matrix is ​​stored in the data storage module to complete the construction of the spectral basis space.

[0034] S2: Import the target spectrum and map the initial values ​​for the driving force. When it is necessary to photograph a photosensitive object to be tested, the operator selects or directly inputs the target spectrum from the host computer interface according to the shooting requirements (such as standard lighting conditions like D55 daylight or A source, or inputting a custom spectral curve). This target spectrum The ideal illumination spectrum expected to reach the surface of the photosensitive object under test is used, rather than the reflection spectrum of the object itself. The fitting decision module loads this spectrum. Then, the spectral modeling module is first invoked, and the projection matrix stored in S1 is used to project the target spectrum. Projecting onto the established spectral basis space yields its corresponding coordinate vector. .

[0035] The data storage module pre-stores a mapping table of driving vectors and feature coordinates. The table is generated as follows: after the system completes the S1 calibration, the host computer controls the light source unit to randomly generate thousands of physically realizable driving current combinations (vectors) on the standard reference body, measures the output spectrum under each combination, calculates its base space coordinates, and then stores the correspondence between the driving vector and the base space coordinates in the table.

[0036] In this step, the fitting decision module uses the target spectrum. coordinate vector Using the query criteria, a fast nearest neighbor search is performed on this mapping table. The algorithm will find those coordinates in the mapping table that are adjacent to the nearest neighbor. Minimum Euclidean distance This record, and this The driving vectors corresponding to each record are extracted to form a candidate driver set.

[0037] S3: Closed-loop iteration and driving pre-optimization This step is the main part of the spectral fitting calculation, and all physical iterations are performed on the standard reference volume element.

[0038] The fitting decision module selects the candidate driver with the best overall evaluation from the candidate driver set as the initial driver vector. The optimal criterion for comprehensive evaluation is: prioritizing selection based on the target spectral coordinates. The driving vector with the smallest Euclidean distance is selected. If there are multiple candidates with similar distances, the magnitude of their total driving current (i.e., the sum of the currents of each channel) is further evaluated, and the driving vector with the smaller total current is selected first to reduce the heat load and energy consumption of the light source unit.

[0039] The communication control module instructs the light source unit to... Output and control the spectral sensing unit to measure the reflectance spectrum on the current standard reference. .

[0040] Fitting decision module calculation The theoretical reflectance spectrum that the target spectrum should have on the standard reference body (Depend on The difference vector between (calculated from standard reflectance data) and .

[0041] The system then enters a fast local optimization loop. The fitting decision module is based on the current spectral differences. And estimate the system at the current operating point. Local response characteristic matrix of the vicinity (That is, the influence matrix of a small change in drive current on the output spectrum, which can be obtained through calibration data or online identification), calculate a set of drive current adjustment amounts. This is usually achieved through the least squares problem. ,in, For iteration The adjustment amount at this time. The communication control module will send new drive instructions. The signal is sent to the light source unit, and the spectrum is measured again. Evaluation of the fitted decision module and If the difference does not meet the preset convergence condition, then continue based on... And the new difference calculation adjustment amount This process is repeated iteratively. The preset convergence condition is defined as: the current iteration... The spectrum measured at this time With the target theoretical reflectance spectrum The root mean square error (RMSE) between them is less than a preset optimization threshold. ,Right now .

[0042] This iterative process continues, with each iteration updating the understanding of the local response model using the latest measurement data, making the adjustments more precise. After several iterations (usually less than 10), the spectrum measured on the standard reference is obtained. and The difference will be less than a preset optimization threshold. At this point, the fitting decision module obtains a pre-optimized driving vector. It is important to note that all the exploratory illumination generated during the entire iterative search process in this step is borne by the standard reference element; therefore, the photosensitive object under test has not been exposed to any light.

[0043] S4: Single-shot verification irradiation and feature information extraction of the photosensitive object under test This step is the first illumination of the photosensitive object to be tested, with the goal of obtaining real feedback information.

[0044] Before starting, the fitting decision module commands the radiation dose monitoring unit to clear its cumulative dose value to zero and enter the real-time monitoring state through the communication control module.

[0045] Subsequently, the communication control module instructs the optical path guiding unit to switch the beam to the test platform unit. The light source unit immediately uses the pre-optimized driving vector obtained in S3. Output. During illumination, the spectral sensing unit simultaneously acquires the reflectance spectrum of the surface of the photosensitive object under test. Simultaneously, the radiation dose monitoring unit integrates the light radiation from this irradiation throughout the entire process to obtain the dose increment. And update the cumulative dose to .

[0046] After irradiation, the fitting decision module performs a test on the measured spectrum. The analysis will be conducted. Specifically, based on the target spectrum... Based on the spectral characteristics of common photosensitive objects of this type, 3 to 5 characteristic bands are intelligently selected. These bands typically contain spectral peaks, valleys, or inflection points. Module extraction. The intensity values ​​in these characteristic bands, combined with known values, are derived from... The generated illumination spectrum data is used to estimate the approximate reflectance of the photosensitive object under test in these characteristic wavelength bands using the physical principle of light reflection (incident light spectrum × surface reflectance = reflected light spectrum).

[0047] S5: Feature-based feedback-driven correction The fitting decision module judges the full spectrum directly measured in S4. Matching accuracy with the target reflectance spectrum. If the application requirements (e.g., chromatic aberration) have been met. If <1.5), proceed directly to step S6.

[0048] If the requirements are not met, the calibration process is initiated. The fitting decision module utilizes the characteristic band reflectance information obtained in S4, and estimates the reflectance curve of the photosensitive object under test across the entire spectral range using piecewise linear interpolation or a model matching method based on a library of typical reflectance curves, generating an estimated reflectance curve for the photosensitive object under test. .

[0049] Next, the module recalculates the theoretically ideal illumination spectrum. The spectrum needs to meet the following requirements. (This is a vector dot product). Subsequently, the system switches the optical path back to the standard reference volume element. As described above... For the new target output spectrum, on the standard reference, the fitting decision module uses the locally optimized model that has converged at the end of S3 to perform a fast driving value calculation (usually only 1-2 calculations are needed) to obtain the correction driving vector. This process is completed entirely on a standard reference body, and the photosensitive object under test is not subject to additional illumination.

[0050] S6: Security decision-making and final output The fitting decision module first reads the preset safe exposure dose threshold for this type of photosensitive object from the data storage module. .

[0051] The decision is as follows: Case a: If the spectral matching accuracy has reached the target after S4 or S5, the fitting decision module immediately terminates the process. It will then determine the current optimal driving vector (which may be...). or As a final solution, the signal is output to the light source unit via the communication control module, and the photosensitive object under test can be selectively irradiated again for final imaging. The system records the total dose. The final spectral fitting process is now complete.

[0052] Scenario b: If the accuracy does not meet the standard, the fitting decision module will perform a safety check. It reads the current cumulative dose from the radiation dose monitoring unit. And predict The additional dose that may be generated during a single verification irradiation. The system employs a weighted evaluation model to more scientifically measure risk. First, based on the type of the photosensitive object under test, it retrieves a preset weighted function for the material photodamage sensitivity of the photosensitive object from the data storage module. The final equivalent weighted dose used for determination. The calculation formula is as follows: in, The energy distribution representing the historical irradiation spectrum, This represents the spectral energy distribution of the predicted verification irradiation. The summation operation iterates through all wavelengths. If the calculated equivalent weighted dose... Greater than or equal to the safe exposure dose threshold If the system fails to irradiate the artifact, it will be forcibly terminated, ceasing any further irradiation that could endanger the artifact. The host computer will output an alarm and correct the drive vector. The theoretically optimal solution is provided to the operator, along with a detailed dosage report, thus concluding the decision-making process.

[0053] Case c: If the accuracy does not meet the standard and the equivalent weighted dose Less than the safe exposure dose threshold If the system allows the final verification irradiation, the communication control module instructs the optical path to switch to the photosensitive object under test. Drive the light source to illuminate the area and measure the final light intensity. Simultaneously update the cumulative dose. The fitted decision module is evaluated. The accuracy, and The final driving scheme output is determined. The system completes the final spectral fitting and output operations.

[0054] This completes the entire process of high-precision spectral fitting illumination of the photosensitive object under strict dose control.

[0055] Finally, it should be noted that the above embodiments are merely examples for clearly illustrating the present invention and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method of spectral fitting of a spectrally tunable light source, characterized by, The method employs a spectral fitting system comprising a host computer and a slave computer. The slave computer includes a light source unit, a spectral sensing unit, an optical path guiding unit, a standard reference unit, a test object platform unit, and a radiation dose monitoring unit. The host computer includes a communication control module, a data storage module, a spectral modeling module, and a fitting decision module. The light source unit consists of multiple light-emitting diodes (LEDs) with different peak wavelengths and corresponding independent constant current driving circuits. The light emitted by all LEDs is uniformly mixed by an optical integrating sphere before being output. The spectral sensing unit is a fiber optic spectrometer, whose detection range covers the emission band of the light source unit. The optical path guiding unit includes a dual-position mirror module driven by a stepper motor, used to selectively guide the light beam from the optical integrating sphere to a first fixed physical position or a second fixed physical position. The standard reference unit is a diffuse reflective white plate with calibrated and known spectral reflectivity, which is fixedly placed at the first fixed physical position. The test object platform unit is used to place the photosensitive object to be tested, which is fixedly placed at the second fixed physical position. The radiation dose monitoring unit includes a photodetector with a flat spectral response and a signal processing circuit, used to monitor and accumulate exposure dose values ​​in real time. The method includes: S1: Characterization of the light source system and construction of the spectral basis space: On the standard reference unit, the output of each LED channel under different driving currents is calibrated to obtain the basis functions and construct the spectral basis space; S2: Import the target spectrum and map the initial driving values: Project the target illumination spectrum onto the spectral basis space, and search for the driving candidate set based on the pre-stored driving vector-spectral coordinate mapping table; S3: Closed-loop iteration and driver pre-optimization: On the standard reference unit, starting from the driver candidate set, through iterative measurement and adjustment, the output spectrum is made to approximate the target, and a pre-optimized driving vector is obtained; S4: Single-time verification irradiation and feature information extraction of the photosensitive object under test: The optical path guiding unit is controlled to switch the beam to the position of the platform unit of the object under test, and the photosensitive object under test is irradiated once with the pre-optimized driving vector. The reflection spectrum is collected and the dose is accumulated, and the feature band information is extracted from it to estimate the approximate reflectivity of the object under test. S5: Feature-based feedback-based drive correction: The reflectivity curve of the test object is estimated using the feature band information, the target illumination spectrum is recalculated, and the optical path guiding unit is controlled to switch the beam back to the position of the standard reference unit. The correction drive vector is calculated on the standard reference unit. S6: Safety decision and final output: Calculate the equivalent weighted dose, compare it with the safety threshold, make a decision based on the comparison result and output the final driving scheme or forced termination process; The equivalent weighted dose The calculation formula is: in, The energy distribution of the historical irradiation spectrum. To predict and verify the spectral energy distribution of the irradiation, The weighting function for the photodamage sensitivity of the material of the photosensitive object under test is used, and the summation operation is performed across all wavelengths.

2. The spectral fitting method for a spectrally tunable light source as described in claim 1, characterized in that, The data storage module is also used to store spectral basis space data, driving vector-spectral coordinate mapping table, and preset safe dose thresholds for different types of photosensitive objects under test and their corresponding material photodamage sensitivity weight functions. The spectral modeling module constructs a spectral basis space by controlling each LED channel of the light source unit to individually illuminate the standard reference unit under different driving currents. The absolute spectral power distribution under each state is calculated using the measured reflection spectrum and the known reflectivity of the standard reference unit as basis functions, and principal component analysis is performed on all basis functions to construct a low-dimensional spectral basis space. The fitting decision module executes spectral fitting control logic in two physically isolated spaces in stages, characteristic band feedback operation based on a single verification irradiation spectrum, and weighted dose safety decision based on cumulative dose and sensitivity weight.

3. The spectral fitting method for a spectrally tunable light source as described in claim 1, characterized in that, In step S3, the iterative measurement and adjustment is based on the difference between the spectrum measured on the standard reference body under the current driving vector and the target theoretical reflectance spectrum, as well as the local response characteristic matrix of the system near the current operating point. The driving vector is iteratively adjusted by solving the least squares problem until the spectral difference is less than the preset optimization threshold.

4. The spectral fitting method for a spectrally tunable light source as described in claim 1, characterized in that, In step S4, the extraction of characteristic band information involves selecting 3 to 5 characteristic bands containing peaks, valleys, or inflection points from the measured full spectrum, and using the known illumination spectrum to estimate the approximate reflectance of the object under test in these bands.

5. The spectral fitting method for a spectrally tunable light source as described in claim 1, characterized in that, The method is used for spectral illumination or measurement of photosensitive objects where the total exposure dose needs to be strictly limited.

Images