Fluorescence scanner light bleaching real-time compensation and excitation light intensity adaptive control method and device and medium

Abstract

The present application provides a fluorescence scanner light bleaching real-time compensation and excitation light intensity adaptive control method, device and medium, aiming at the problem of fluorescence signal attenuation and uneven brightness caused by light bleaching in full film scanning, the present application extracts fluorescence signal intensity in each block scanning, uses sliding window least square fitting to update the parameters of exponential decay model in real time; according to the updated model, the attenuation amount is predicted and the target excitation light intensity of the next block is calculated, and the output power of the light source is adjusted block by block; when the target excitation light intensity reaches the maximum safe power, the camera exposure time is increased to maintain the signal-to-noise ratio; when the exposure time reaches the upper limit and the signal-to-noise ratio continues to decrease, the light bleaching warning is triggered. The present application compensates signal attenuation from the physical source and is suitable for fluorescence slice panoramic high-resolution scanning imaging.

Description

Technical Field

[0001] This invention relates to the field of fluorescence scanning imaging technology, and in particular to a method, apparatus, and medium for real-time compensation of photobleaching and adaptive control of excitation light intensity in a fluorescence scanner. Background Technology

[0002] Fluorescent pathology slide scanners perform high-resolution excitation and image acquisition on fluorescently labeled biological tissue slides by stepping or moving continuously, and finally generate full-field fluorescent digital slides by image stitching. Full-slide scanning usually lasts for tens of minutes. Under continuous excitation light, fluorescent dye molecules undergo irreversible photochemical reactions, resulting in photobleaching. This causes the intensity of the emitted fluorescence signal to monotonically decay as the scanning progresses and the irradiation time accumulates.

[0003] Existing fluorescence scanning systems typically employ a constant, fixed output power of the excitation source for the entire scanning process. To mitigate image seams and brightness unevenness caused by photobleaching, conventional solutions primarily employ image post-processing algorithms, such as linear mapping of the grayscale mean of overlapping adjacent tiles or brightness normalization, to achieve visual smoothing after image acquisition.

[0004] However, photobleaching directly reduces the number of effective excitation photons in the sample area and causes physical signal loss. Post-processing algorithms can only achieve surface visual homogenization and cannot reversely recover the lost photon signals and the true dynamic range. At the same time, a fixed excitation light intensity is prone to overexposure of fluorescence and accelerated bleaching in the early stage of scanning due to excessive power, while the effective signal depletion in the later stage of scanning leads to a serious deterioration of the image signal-to-noise ratio. In addition, when faced with the situation where the bleaching rates of different dyes in multicolor fluorescent labeling are significantly different, uniform and fixed light intensity control cannot achieve differential compensation, resulting in some channels being overexposed and others being underexposed and distorted. Summary of the Invention

[0005] This invention provides a method, device, and medium for real-time compensation of photobleaching and adaptive control of excitation light intensity in a fluorescence scanner. Existing fluorescence scanning systems use a fixed excitation light power combined with post-processing brightness mapping, which cannot compensate for the attenuation of fluorescence signal caused by photobleaching at the physical source. In multicolor scanning, it is also impossible to provide differentiated compensation for dyes with different bleaching rates.

[0006] The core technology of this invention is to use a sliding window to online fit and update the parameters of the single exponential decay model based on the fluorescence signal intensity extracted watt by watt during the fluorescence scanning process, and to calculate the target excitation light intensity of the next watt in reverse to achieve watt-by-watt gradient adaptive incremental compensation of the output power of the excitation light source. After the excitation light intensity reaches the upper limit, the camera exposure time and scanning frame rate are linked in stages.

[0007] In a first aspect, the present invention provides a method for real-time compensation of photobleaching and adaptive control of excitation light intensity in a fluorescence scanner, the method comprising the following steps: During the full-slice fluorescence scanning process, the fluorescence image of the currently acquired tile is obtained tile by tile, and the fluorescence signal intensity of the current tile is extracted from the fluorescence image; The fluorescence signal intensity is input into a sliding window, and the decay time constant of the photobleaching exponential decay model is updated in real time using least squares fitting. The photobleaching exponential decay model is as follows: , For the first The fluorescence signal intensity of each tile, The initial fluorescence signal intensity, The scanned tile sequence number. The excitation time of a single watt is the time required for light to be emitted. The decay time constant; The signal attenuation of the next watt is predicted based on the updated photobleaching exponential attenuation model, and the excitation intensity compensation formula is applied: Calculate the target excitation light intensity of the next watt, where For the first The target excitation light intensity of each tile The initial excitation light intensity; The output power of the excitation light source is adjusted watt by watt according to the target excitation light intensity; when the target excitation light intensity reaches the maximum safe power of the excitation light source, the excitation light intensity is kept constant at the maximum safe power, and the camera's exposure time is increased in conjunction to maintain the image signal-to-noise ratio; When the exposure time reaches the preset upper limit and the signal-to-noise ratio continues to decrease, a photobleaching warning is triggered.

[0008] Further, the fluorescence signal intensity of the current tile is extracted from the fluorescence image, including: Adaptive threshold segmentation is performed on the fluorescence image to obtain the foreground region, and morphological opening is performed on the foreground region for noise reduction and edge artifact removal to obtain the region of interest of the fluorescence signal. Calculate the gray-level uniformity and signal-to-noise ratio within the region of interest; When the grayscale uniformity is not greater than the preset uniformity threshold and the signal-to-noise ratio is not less than the preset signal-to-noise ratio threshold, the average grayscale value within the region of interest is extracted as the fluorescence signal intensity of the current tile. When the grayscale uniformity is greater than the preset uniformity threshold or the signal-to-noise ratio is less than the preset signal-to-noise ratio threshold, the peak grayscale value at the preset high-resolution site in the region of interest is extracted as the fluorescence signal intensity of the current tile.

[0009] Furthermore, the fluorescence signal intensity is input into a sliding window, and the decay time constant of the photobleaching exponential decay model is updated in real time using least-squares fitting, including: Before data acquisition at the beginning of the scan The fluorescence signal intensity of each effective tile is fitted to the initial decay time constant; Maintain length in subsequent scans A sliding window for each tile, for viewing the inside of the window. and Perform linear regression fitting to obtain the fitted decay time constant; According to the exponential smoothing formula Update the current decay time constant, where The updated decay time constant, To fit the decay time constant, The decay time constant of the previous update cycle. The preset smoothing coefficient; When the goodness-of-fit coefficient of the sliding window fitting reaches a preset frame number less than a preset residual threshold, the model is deemed to have failed and the decay time constant is reset to the initial decay time constant.

[0010] Furthermore, the output power of the excitation source is adjusted watt by watt according to the target excitation light intensity, including: For the first The target excitation light intensity of each tile is subject to a single-tile adjustment step size limit: ,in For the first The excitation light intensity of each tile The maximum adjustment step size for a single tile is preset; The excitation light intensity after applying a step size limit is sent to the excitation source driver for power adjustment.

[0011] Furthermore, the method includes three-stage coordinated control of excitation light intensity, exposure time, and scan frame rate: Phase 1: When the target excitation light intensity has not reached the maximum safe power, increase the output power of the excitation light source watt by watt while keeping the exposure time constant; Second stage: When the target excitation light intensity reaches the maximum safe power and the image signal-to-noise ratio meets the standard, maintain the excitation light intensity at the maximum safe power, and increase the exposure time watt by watt to the maximum allowable exposure time; Phase 3: When the exposure time reaches the maximum allowable exposure time and the signal-to-noise ratio continues to decrease, triggering a photobleaching warning, a frame rate reduction strategy is implemented. The scanning frame rate is reduced to extend the frame period and increase the corresponding maximum allowable exposure time. This is done to maintain the image signal-to-noise ratio and slow down the photobleaching rate by further increasing the exposure time. If the signal-to-noise ratio continues to fall below the minimum signal-to-noise ratio threshold after the frame rate reduction reaches the preset confirmation frame number, the corresponding area is marked as a photobleaching-restricted area and the scanning of the remaining areas continues.

[0012] Furthermore, in multi-dye multi-channel fluorescence scanning scenarios, the method includes: The sequential excitation mode is adopted, and images are acquired, decay time constants are updated, and target excitation light intensity of each channel is calculated independently. When the total available power of the excitation source is limited, the power allocation weight is calculated based on the reciprocal of the decay time constant of each fluorescence channel, according to the formula. Allocate an excitation power budget to each fluorescence channel, where The power budget allocated to the current channel. Total available power, This represents the decay time constant of the current channel. Let be the decay time constant of the j-th channel; The acquired multi-channel images are collaboratively processed to perform spectral crosstalk correction post-processing.

[0013] Furthermore, when a scanned area needs to be rescanned, the method also includes: Before rescanning, a preview frame of the rescanning area is acquired with a preview excitation power lower than the initial excitation light intensity to obtain the current measured signal intensity; Based on the ratio of preview excitation power to initial excitation light intensity, the current measured signal intensity is equivalently normalized to the reference signal intensity; The fluorescence recovery rate is calculated by comparing the reference signal intensity with the original signal intensity of the rescanned region during the original acquisition. When the fluorescence recovery rate is greater than the first recovery threshold, the initial scan parameters are used for rescanning; when the fluorescence recovery rate is less than the second recovery threshold, the compensation scan parameters recorded before rescanning are used for rescanning; when the fluorescence recovery rate is between the second recovery threshold and the first recovery threshold, linear interpolation is performed between the initial scan parameters and the compensation scan parameters based on the fluorescence recovery rate to obtain the rescanning parameters and perform rescanning.

[0014] Secondly, the present invention provides a real-time compensation and excitation light intensity adaptive control device for photobleaching in a fluorescence scanner, comprising: The photobleaching modeling module is used to acquire the fluorescence image of the current acquisition tile one by one during the full-slice fluorescence scanning process, extract the fluorescence signal intensity of the current tile from the fluorescence image, input the fluorescence signal intensity into the sliding window, and use least squares fitting to update the decay time constant of the photobleaching exponential decay model in real time. The compensation decision module is used to predict the signal attenuation of the next watt based on the updated photobleaching exponential attenuation model, and to calculate the target excitation intensity of the next watt according to the excitation intensity compensation formula. The excitation light source control module is used to adjust the output power of the excitation light source watt by watt according to the target excitation light intensity in order to compensate for the attenuation of fluorescence signal caused by photobleaching; The exposure linkage module is used to increase the camera's exposure time to maintain the image signal-to-noise ratio when the target excitation light intensity reaches the maximum safe power of the excitation light source, and to trigger a light bleaching warning when the exposure time reaches the preset upper limit and the signal-to-noise ratio continues to decrease. The multi-channel management module is used to independently maintain the photobleaching exponential decay model and compensation parameters for each fluorescence channel during multi-dye multi-channel fluorescence scanning, and to perform multi-channel power budget management.

[0015] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to execute the above-described method for real-time compensation and adaptive control of excitation light intensity for photobleaching of a fluorescence scanner.

[0016] Fourthly, the present invention provides a readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for real-time compensation of photobleaching and adaptive control of excitation light intensity in a fluorescence scanner.

[0017] The main contributions and innovations of this invention are as follows: 1. This invention compensates for signal attenuation from the physical source by fitting the photobleaching exponential decay model to each tile in real time during the scanning process and adaptively adjusting the excitation light intensity. This keeps the output signal intensity of each tile in the whole film constant, eliminates the brightness gradient after splicing, and retains all the information of the original fluorescence signal. This avoids the inherent problem that post-processing correction cannot recover the lost signal.

[0018] 2. In multicolor fluorescence scanning, this invention maintains an independent bleaching model and compensation strategy for each fluorescence channel. Each channel independently calculates the excitation light intensity compensation amount according to its own bleaching rate, and allocates the power budget according to the severity of bleaching when the total power budget is limited. This achieves differentiated and precise compensation for multiple dyes and avoids the problems of overexposure in slow bleaching channels and underexposure in fast bleaching channels.

[0019] 3. This invention expands the available compensation range to the maximum extent and effectively delays the occurrence of scan termination by triggering a three-stage linkage control of light intensity and exposure time—first increasing light intensity compensation, then increasing exposure time compensation after the light intensity reaches the upper limit, and triggering an early warning when the exposure time also reaches the upper limit and the signal-to-noise ratio continues to decrease.

[0020] 4. After both light intensity and exposure time reach their upper limits, this invention extends the upper limit of exposure time and slows down the photobleaching rate by reducing the frame rate, thereby avoiding scanning interruption and data incompleteness due to excessive photobleaching and ensuring the integrity of the entire film's data.

[0021] Details of one or more embodiments of the present invention are set forth in the following drawings and description, so that other features, objects and advantages of the invention will be more readily understood. Attached Figure Description

[0022] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a diagram illustrating the excitation intensity-exposure time linkage compensation strategy of the fluorescence scanner photobleaching real-time compensation and excitation intensity adaptive control method according to an embodiment of the present invention. Figure 2 This is a multi-dye differential photobleaching compensation matrix diagram according to an embodiment of the present invention; Figure 3 This is a diagram illustrating the architecture and data path of a real-time compensation system for photobleaching according to an embodiment of the present invention. Figure 4 This is the main flowchart of real-time compensation for photobleaching according to an embodiment of the present invention. Detailed Implementation

[0023] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0024] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.

[0025] I. System Overall Architecture and Working Principle like Figure 3 As shown, the present invention provides a real-time compensation and adaptive excitation light intensity control system for photobleaching in a fluorescence scanner, comprising five proprietary functional modules: an excitation source control module, a photobleaching modeling module, a compensation decision module, an exposure linkage module, and a multi-channel management module. Based on these modules, the system can independently complete online photobleaching modeling and adaptive closed-loop parameter control during the full-scan fluorescence scan process.

[0026] The functions and data interaction relationships of each module are as follows: Photobleaching modeling module: Communicates with the image acquisition device (CMOS camera) to receive fluorescence images tile by tile during full-slice fluorescence scanning, and extracts the fluorescence signal intensity of the current tile from the image. The signal strength is then input into a sliding window to perform least-squares fitting, and the decay time constant in the photobleaching exponential decay model is updated in real time online. The updated version will follow. Output to the compensation decision module.

[0027] The compensation decision module communicates with the photobleaching modeling module, the excitation source control module, the exposure linkage module, and the image storage / stitching module. The compensation decision module determines the compensation based on the received decay time constant. Predict the signal attenuation of the next watt and calculate the target excitation intensity of the next watt using the excitation intensity compensation formula. .like The maximum safe power of the excitation source was not reached. If so, the excitation light intensity compensation command will be sent to the excitation light source control module; if achieve If the exposure time reaches the upper limit and the signal-to-noise ratio continues to decrease, an exposure compensation command will be sent to the exposure linkage module; if the exposure time reaches the upper limit and the signal-to-noise ratio continues to decrease, a photobleaching warning will be triggered.

[0028] Excitation source control module: This module is hardware-connected to the driver of the excitation source (LED or laser). It receives control commands from the compensation decision module, adjusts the output power of the excitation source watt by watt, and provides real-time feedback of the actual output power value to the compensation decision module for anomaly detection. When the system hardware is equipped with a power feedback sensor, a hardware fault alarm will be triggered if the deviation between the feedback value and the set value exceeds 10%. If the system is not equipped with a power feedback sensor, this anomaly detection function can be disabled or skipped.

[0029] Exposure linkage module: Connected to the camera's timing control terminal, it is used to adjust the camera's effective exposure time parameters according to the instructions issued by the compensation decision module.

[0030] Multi-channel management module: In multi-color fluorescence scanning scenarios, this module independently maintains the photobleaching exponential decay model parameters and compensation strategies for each fluorescence channel, independently controls the excitation light intensity of each channel, and allocates the power budget according to the bleaching severity of each channel when the total power budget for multiple channels is limited. In monochromatic fluorescence scanning scenarios, this module is in its default configuration state, maintaining only the model parameters of a single channel and not performing multi-channel collaborative functions.

[0031] like Figure 4 As shown, the present invention provides a method for real-time compensation and adaptive control of excitation light intensity for photobleaching in a fluorescence scanner, the main loop of which includes steps S101 to S108: S101: Acquire the fluorescence image of the current tile and record the current excitation light intensity. Compared to the current exposure time .

[0032] S102: Adaptively extract the fluorescence signal intensity of the current tile from the current tile fluorescence image. .

[0033] S103: Will Substitute into the sliding window and update the parameters of the photobleaching attenuation model. .

[0034] S104: Predict the signal attenuation of the next watt based on the updated model, and calculate the target excitation light intensity. .

[0035] S105: Adjust the excitation light intensity to... ;like achieve If so, exposure time compensation will be activated.

[0036] S106: Acquire the image of the next tile according to the adjusted parameters, and verify online whether the signal-to-noise ratio (SNR) of the acquired image meets the standard.

[0037] S107: If the SNR meets the standard, output the current tile image and pass it to the storage / stitching module to enter the next tile scanning schedule.

[0038] S108: If the SNR does not meet the standard and both the excitation light intensity and exposure time have reached the upper limit, trigger the light bleaching warning and notify to reduce the scanning frame rate.

[0039] II. Adaptive Extraction Method for Fluorescence Signal Intensity In step S102, to eliminate the interference of uneven fluorescence distribution, local overexposure, or noise on modeling and fitting, the following anti-interference adaptive extraction process is adopted: Region of Interest (ROI) Determination: The Otsu algorithm is used to perform adaptive thresholding segmentation on the current tile image. The set of foreground pixels above the threshold is determined as the fluorescence signal region, which is then used as the ROI. For tissue slices with uneven signal distribution, the ROI is the union of all segmented foreground regions.

[0040] Morphological denoising and artifact removal: employing The structuring element performs a morphological opening operation on the segmentation results to filter out isolated noise points and remove edge artifacts with a width of 2 pixels at the edge of the field of view. When the segmented foreground pixels account for less than 1% of the total tile area, the current tile is determined to be a region without fluorescence signal (such as the background area of ​​a blank slide), the signal intensity is recorded as 0, and the update of the attenuation model for the current tile is skipped to prevent idle data from polluting the model.

[0041] Dynamic switching of signal extraction strategy: calculating grayscale uniformity within the ROI (defined as the standard deviation of pixel grayscale within the ROI) With grayscale mean The ratio: and signal-to-noise ratio (defined as the average gray level within the ROI) Standard deviation of pixel grayscale from background area The ratio: ).

[0042] By default, the average grayscale value of all pixels within the ROI is extracted as the fluorescence signal intensity of the current tile. .

[0043] When grayscale uniformity is detected or signal-to-noise ratio When this occurs, it indicates that the average grayscale is affected by local overexposure or strong noise interference. The system automatically switches to extracting the grayscale values ​​at preset high quantiles within the ROI as the fluorescence signal intensity. .

[0044] In this embodiment, the grayscale value at the 98th percentile is preferably extracted (i.e., the grayscale value at the 98th percentile is taken after sorting the grayscale values ​​of all pixels within the ROI region in ascending order). This value eliminates the interference of thermal noise or abnormal bright spots in individual pixels, resulting in robust extraction results. In practical implementation, this preset high percentile can be pre-configured within the range of 95% to 99% according to the detector noise level. Different detector models and imaging conditions can correspond to different percentile values.

[0045] Background region selection: Standard deviation of background pixels in signal-to-noise ratio calculation It can be determined in any of the following ways: Method 1 (Preferred Implementation): The annular region formed between the ROI and its original boundary after expanding it by a preset pixel width (20 pixels in this example) is taken as the background region. This annular region is directly adjacent to the tissue signal region, which can effectively avoid statistical bias caused by excessive brightness in the blank area of ​​the slide or excessive darkness in the tissue gap.

[0046] Method 2: Extract preset-sized areas from the four corners of the image (e.g., each...) (Pixel blocks) are used as the background area.

[0047] Method 3: Take all pixel regions below the Otsu segmentation threshold as the background region.

[0048] III. Physical Model and Online Real-Time Modeling of Photobleaching In tile-scale fluorescence section scanning, the excitation light irradiation time of a single tile is on the order of milliseconds, and the total cumulative irradiation time of a single slide is typically 30–120 seconds. The photobleaching process is described using a single exponential decay physical model.

[0049] In the formula, For the first Fluorescence signal intensity during individual tile acquisition; The initial fluorescence signal intensity; The scanned tile number; The excitation time per watt (related to frame rate and exposure time); This is the photobleaching time constant of the fluorescent dye (related to the dye type, excitation light power, and ambient temperature).

[0050] The selection criteria for the single exponential model are as follows: within the full-film observation window of 30 to 120 seconds, the fast bleaching component (if present) in the double exponential decay tends to stabilize in the first few tiles, and the remaining observation interval is mainly dominated by the slow component. The fitting accuracy of the single exponential model within this window can fully meet the real-time compensation requirements.

[0051] The online fitting and real-time update mechanism of the decay model (step S103) is as follows: Initial parameter fitting: before acquisition at the start of the scan One effective tile (example taken) The fluorescence signal intensity of the excitation source is maintained at its initial excitation intensity during this initialization period. Output, without power compensation. This approach enhances statistical robustness while meeting the minimum sample requirement for univariate linear regression, and avoids initial fitting crossing the boundaries of different tissue regions. The initial decay time constant is obtained by fitting these five sample points. And compensation will begin from the 6th tile.

[0052] Sliding window least squares fitting: After entering subsequent continuous scans, maintain a length of... One tile (example taken) The sliding window allows you to view the individual tiles within the window. With tile serial number When performing a univariate linear regression, the regression slope is... This allows us to calculate the current fitted decay time constant. .Pick At 30fps, this corresponds to approximately 0.67 seconds, balancing noise resistance statistical stability and... Real-time tracking of actual changes.

[0053] Exponential smoothing filter update: To avoid dimming oscillations caused by single-batch fitting noise, the decay time constant is updated using an exponential smoothing formula:

[0054] In the formula, The updated decay time constant, The fitted value for the current sliding window. The decay time constant of the previous update cycle. The preset smoothing coefficient (in the example, is taken as follows) ).Pick This makes the equivalent response period of the time constant approximately 5 tiles, balancing tracking speed and smoothness.

[0055] Model Failure Detection and Reset: Calculating the Determination Coefficients of the Sliding Window Fit .when Continuous Frame (Example) When the single-exponential model is deemed to be mismatched and ineffective in the tissue region, the system automatically... Reset to the initial decay time constant It also records the failure log and restarts the data accumulation of the sliding window. The corresponding residual variance ratio is 15%. A value lower than this indicates model mismatch. Confirmation of 3 consecutive frames can effectively filter out false positives due to instantaneous noise in a single frame.

[0056] Alternative bi-exponential model: In alternative implementations where fluorescent dyes exhibit significant fast and slow two-phase bleaching characteristics, a switch to a bi-exponential decay model can also be triggered after model mismatch.

[0057] In the formula, , These represent the initial intensity and time constant of the rapid bleaching component, respectively; , The initial intensity and time constant of the slow bleaching component are respectively determined using the Levenberg-Marquardt nonlinear fitting algorithm.

[0058] IV. Adaptive Gradient Control of Excitation Light Intensity per Tile According to the updated attenuation model, in order to keep the output signal strength constant (i.e. ), No. The target excitation light intensity of each tile is calculated using the following excitation light intensity compensation formula (step S104):

[0059] In the formula, For the first The target excitation light intensity of each tile; The initial excitation light intensity (in this example, it is taken as 50% of the maximum safe power of the light source, i.e.) ).

[0060] The physical nature of compensation and the analysis of boundedness of error: The first The signal attenuation factor of each tile is The excitation light intensity compensation factor is The product of the two is always equal to 1, which is the signal strength after compensation. This ensures that the effective excitation light dose received by each tile remains constant.

[0061] Increasing the excitation power will slightly accelerate photobleaching. Follow (increases and decreases slightly), but because To achieve real-time online fitting and updates, and subject to exponential smoothing constraints, the relationship between adjacent tiles... The upper limit of the relative rate of change satisfies In typical fluorescent dyes (such as FITC, Under the limitation of single-step dimming, The rate of change of a single tile is less than 2%. of The error causes the relative error of the compensated signal strength to not exceed Taking full-screen scanning as an example ( The entire film was powered by 2000 watts, and the total illumination time was approximately 67 seconds. This error is only about 1.7%, far less than the 30% to 70% attenuation without compensation; even Even when the value is increased to 1.5, the error is only about 3%, which proves that the error under closed-loop control is bounded and converges.

[0062] Dimming constraints: 1. Maximum safe power constraint: ,in To excite the maximum safe power calibration value of the light source, and to prevent burning of the sample or accelerating photobleaching.

[0063] 2. Single tile adjustment step size limit: ,in The maximum adjustment step size for a single tile is preset (in the example, it is taken as follows). This avoids abrupt changes in brightness between adjacent tiles.

[0064] 3. When achieve At that time, stop increasing the excitation light intensity and maintain... No change, switch to exposure time-linked compensation.

[0065] Table 1. Definitions and values ​​of excitation light intensity gradient control parameters in examples.

[0066] V. Excitation Light Intensity and Exposure Time Linked Compensation Strategy When the excitation light intensity reaches Then, the system initiates exposure time compensation. The imaging signal-to-noise ratio meets the requirements. ,exist Locked to At that time, increase the exposure time It can maintain the image SNR.

[0067] Maximum permissible exposure time Physical determination method: In continuous scanning imaging mode: the stage moves at a constant speed. To suppress motion blur, the maximum allowable number of blurred pixels is set to... (Example example) If the exposure time is 100%, then the maximum exposure time is:

[0068] In the formula, The scanning speed of the stage ( ); This refers to the magnification of the objective lens. For the physical size of the sensor pixel ( For example, 20× objective lens, scanning speed. Pixel hour, .

[0069] In step-acquisition imaging mode: the stage remains stationary during exposure, there are no motion blur limitations, and the maximum permissible exposure time is affected by the frame period and readout time. (Example uses 3ms) Constraints:

[0070] At 30fps At 15fps .

[0071] Three-stage linkage control mechanism: like Figure 1 As shown in Table 2, the linkage compensation is divided into three stages: Table 2

[0072] The specific execution strategy and fault-tolerance logic for frame rate reduction in Phase 3 are as follows: Stepped frame rate reduction: Halves the current frame rate, i.e. And not lower than the minimum stable frame rate (The example is set at 10fps).

[0073] Effect quantification: Halving the frame rate doubles the frame period (e.g., from 33ms to 66ms), increasing the upper limit of exposure time. The corresponding increase from 30ms to 60ms resulted in a recovery of approximately [missing information - likely a missing word or phrase]. This is approximately +1.5 dB; simultaneously, the cumulative radiation dose to the sample per unit time is halved, and the equivalent bleaching time constant increases to [a certain value]. This significantly slows down the bleaching rate of subsequent tiles.

[0074] Scanning time assessment: Before frame reduction, assess the estimated time increment for the remaining tiles. If the increment exceeds the preset upper limit (50% in this example), mark the subsequent area as a low-quality acquisition area instead of reducing the frame.

[0075] Control linkage: The frame rate reduction request is sent to the acquisition frame rate control module of the scanner firmware layer through a shared memory message queue. The instruction includes the target frame rate, the expected number of continuous tiles, and the reason for the frame rate reduction.

[0076] Irreversible bleaching ultimate protection: If the signal-to-noise ratio remains below the minimum signal-to-noise ratio threshold after frame rate reduction. (Example: 25dB) Achieved Frame (Example) The system determines that the signal in the area has been irreversibly depleted, marks the area as a "photobleaching-restricted area" and records its coordinates, while scanning of the remaining areas continues.

[0077] VI. Differentiated Photobleaching Compensation and Power Budget Allocation for Multiple Dyes like Figure 2 As shown, in multicolor fluorescent section scanning scenarios (e.g., DAPI labeling of cell nuclei, FITC labeling of immunofluorescence, TRITC labeling of channels), the photobleaching rates of each dye differ significantly (e.g., DAPI's...). FITC's TRITC ).

[0078] The multi-channel management module adopts a sequential activation mode, and the control logic is as follows: Independent modeling and compensation: Images are acquired independently for each fluorescence channel, and the time constant is independently fitted and updated. and according to the formula Calculate the target excitation light intensity for each channel independently.

[0079] Dynamic weighted allocation of total power budget: When multiple channels are concurrent or the total output power of the driver is capped. When constraining, the relative magnitudes of the bleaching rates of each channel (i.e., the reciprocal of the time constant) are used. Perform weighted allocation:

[0080] In the formula, The power budget allocated to the current channel for the current watt. For the first The decay time constant of each fluorescence channel. The faster the bleaching rate ( The smaller the channel allocation, the more compensation power it receives, and the power allocated to each channel is limited by the maximum safe power of a single channel. limit.

[0081] Taking DAPI, FITC, and TRITC three-channel as an example ( (300s, 80s, and 150s respectively): Normalized weights are When the total available power for At that time, the three channels were allocated approximately , and The FITC channel, which bleachs the fastest, receives the maximum compensation power.

[0082] Spectral crosstalk collaborative correction: Increasing the excitation light intensity of a specific channel may cause spectral leakage to adjacent emission bands. The system combines a pre-calibrated crosstalk coefficient matrix to perform crosstalk cancellation post-processing on multi-channel images, eliminating inter-channel leakage and ensuring the accuracy of multi-channel quantitative analysis.

[0083] VII. Monitoring and Adaptive Interpolation of Bleaching Recovery in Rescanned Areas When a scanned tile needs to be rescanned due to quality defects such as focusing failure, the following adaptive reconstruction process with rescan parameters is adopted because fluorescent dyes have dark recovery (reversible bleaching) physical properties: Low-power preview acquisition: Before rescanning, a preset low-power excitation light is used (example example). Acquire a single-frame preview image and measure the current signal strength. .

[0084] Excitation intensity equivalent normalization: Within the fluorescence linear response range, low power measurements are equivalently normalized to the reference initial power. The equivalent strength below:

[0085] In the high-power nonlinear response region, the light intensity-emission intensity response curve is obtained by interpolation. .

[0086] Recovery rate calculation and segmentation decision: The normalized intensity is compared with the original signal intensity of the region during the initial scan. Compare and calculate fluorescence recovery rate. Based on recovery rate Perform piecewise adaptive parameter reconstruction: when When the first recovery threshold (0.9 in this example) is reached, it is determined that the dye has been nearly fully recovered, and the initial scan parameters are directly used. (and initial exposure time) to perform a rescan.

[0087] when When the second recovery threshold (0.5 in this example) is reached, it is determined that the dye has not been basically recovered, and the current compensated parameters recorded last before rescanning are directly used. (and corresponding exposure time) to perform a rescan.

[0088] When the second recovery threshold First recovery threshold (in the example, it is taken as follows) When the dye is partially recovered, interpolation is performed based on the recovery rate within the linear interval, with the interpolation coefficient α = (R... (Second recovery threshold) / (First recovery threshold) Second recovery threshold (in the example) ), calculate the excitation intensity of the rescan:

[0089] according to Perform a rescan and update the photobleaching model parameters for that region after the rescan. The first and second recovery thresholds mentioned above can be pre-calibrated and configured according to the dark recovery kinetics of specific fluorescent dye molecules. The values ​​of 0.9 and 0.5 in the examples are only examples. The dark recovery ratios and recovery time constants of different fluorescent dyes (such as FITC, Cy3, Cy5, etc.) vary. In practical applications, the corresponding thresholds should be determined through pre-experimental calibration.

[0090] VIII. Anomaly Detection and Handling Mechanism In the signal extraction (S102) and signal-to-noise ratio verification (S106) steps, the system performs multi-dimensional anomaly monitoring in parallel, as shown in Table 3.

[0091] Table 3. Anomaly Types and Handling Strategies IX. Specific Implementation Methods Example 1: Real-time compensation for full-slice scanning with single-dye FITC This embodiment is applied to whole-slice scanning of monochromatic FITC fluorescently labeled slides.

[0093] Scanning configuration: 20× objective lens, slide size single-watt resolution Pixel size The nominal frame rate is 30fps, and the total number of tiles planned for the entire film is 2000. The initial excitation light intensity is set. Initial exposure time Maximum permissible exposure time .

[0094] Operation process: After the scan begins, the first 5 tiles are used Collect data, extract the average grayscale value of the ROI, and fit the data to obtain the initial decay time constant. Closed-loop compensation is initiated starting from the 6th tile. In step-acquisition mode, the excitation light irradiation time per tile... This is equal to the exposure time (20ms); however, in the full-film scanning sequence, the total time interval between tiles is determined by the frame period (approximately 33ms), and the cumulative illumination time is calculated based on the frame period. When the scan reached the 800th tile, the fluorescence attenuation factor reached [value missing]. The system automatically increases the excitation light intensity according to the compensation formula. (approximately) The output image signal strength is maintained at When scanning to the 1600th tile, the excitation light intensity smoothly increases to... When the 1664th tile was scanned, the target excitation light intensity reached [value missing]. The system automatically and seamlessly transitions to stage 2 exposure compensation, maintaining a constant excitation light intensity. The exposure time was increased from 20ms to 30ms per tile. When the 1900th tile was scanned, the exposure time reached the 30ms limit and the SNR began to decrease, triggering a stage 3 warning.

[0095] Implementation results: Among the 2000 tiles in the entire chip, the signal strength of the first 1600 tiles remained consistent. Within the initial range (in the uncompensated comparative experiment, the signal attenuated to 0.44 times the initial value at the end, indicating a severe brightness gradient); the latter 400 tiles maintained the acceptable imaging SNR through exposure compensation and warning frame reduction. Actual measurement... As power increases, the signal duration gradually decreases from 80s to approximately 76s (a change of about 5%), and the signal consistency remains consistent after compensation. Within.

[0096] Example 2: Multicolor fluorescence DAPI+FITC dual-channel differential compensation This embodiment is applied to sequential excitation scanning of DAPI and FITC dual-color fluorescently labeled sections.

[0097] Scanning configuration: 40× objective lens, total 2000 tiles for the entire film, initial light intensity for both channels set to [value missing]. Actual measurements show that the DAPI channel decays slowly. FITC channel decays rapidly ( ).

[0098] Operation process: The multi-channel management module independently builds models for the two channels and calculates compensation amounts. When scanning the 2000th tile, the DAPI channel has accumulated approximately 66.7 seconds of irradiation, and the signal attenuates to... Its excitation light intensity is determined by Gradually increase to (Compensation amount 1.25 times), overall signal consistency reaches... The FITC channel reached its maximum intensity in the 1664th tile. (Compensation amount 2.0x), exposure compensation is activated and an alert is triggered at the 1900th tile. Based on the channel status, the system implements a frame rate reduction strategy of 15fps for the FITC channel, while maintaining a high-speed scan of 30fps for the DAPI channel.

[0099] Implementation Results: DAPI channel full-chip consistency achieved ; FITC channel front 1664 blocks consistency achieved The 336-watt module maintains the SNR within a usable range by reducing the frame rate, completely overcoming the severe underexposure defect caused by the signal attenuation to 0.43 times at the end of FITC under fixed excitation power.

[0100] Example 3: Linked Control of Optical Bleaching Warning and Adaptive Frame Rate Reduction This embodiment follows the response process of the FITC channel after the warning is triggered on the 1900th tile in Embodiment 2.

[0101] Triggering state: When scanning to the 1900th tile, the excitation light intensity is... Exposure time reaches the current frame rate limit The image signal-to-noise ratio was detected to drop from an initial 35dB to 28dB over 5 consecutive frames (below the high quality threshold).

[0102] Control Response: Step S108 triggers a light bleaching warning. The compensation decision module sends a frame rate reduction command to the acquisition control module, lowering the FITC channel acquisition frame rate from 30fps to 15fps (higher than 15fps). The frame period doubled from 33.3ms to 66.7ms. After deducting the 3ms readout time, the maximum permissible exposure time... Extended to 60ms. The system further increased the exposure time to 60ms, resulting in a recovery of the sampled SNR to approximately 29.5dB (an improvement of approximately 1.41 times); simultaneously, due to the halving of the light dose per unit time, the equivalent bleaching constant... The time was increased from 80 seconds to 160 seconds, significantly slowing down the bleaching speed of the remaining 100 tiles. The time taken to bleach the remaining 100 tiles only increased by about 3.3 seconds, satisfying the constraint that the time increment was less than 50%.

[0103] Implementation results: Scanning tasks were prevented from being interrupted due to severe localized bleaching; all data acquisition was complete; and signal consistency was maintained. Within.

[0104] Example 4: Adaptive Reconstruction of Bleached Recovery During Rescanning of Scanned Areas This embodiment follows the scenario in Embodiment 1 where the 1600th tile needs to be rescanned 8 minutes after data acquisition because the focusing mechanism loses focus due to micro-motion.

[0105] Rescan detection: Before rescanning, the system uses... A single-frame preview was captured, and the current signal strength was measured. .

[0106] Calculation and Interpolation: Calculation of Equivalent Reference Strength via Linear Normalization The fluorescence recovery rate was obtained. The recovery rate is at Calculate interpolation weights within the partial recovery interval. The historical compensation power during the initial scan of the 1600th tile was... Substitute the values ​​into the interpolation formula to calculate the rescan excitation power:

[0107] Set exposure time and with Power is used for rescan data acquisition.

[0108] Implementation results: The rescanned images had appropriate grayscale and the signal-to-noise ratio fully met the requirements; this avoided the need to directly reuse existing data. The resulting severe overexposure and saturation also prevented the direct use of the initial exposure. This results in an insufficient signal-to-noise ratio.

[0109] 10. Electronic devices and computer-readable storage media The present invention also provides an electronic device, including a processor and a memory. The memory stores a computer program, and the processor is configured to run the computer program to execute the various steps of the above-described method for real-time compensation and adaptive control of excitation light intensity in a fluorescence scanner.

[0110] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method. The storage medium includes, but is not limited to, non-transient computer-readable storage media such as read-only memory (ROM), random access memory (RAM), magnetic disk, optical disk, or flash memory.

[0111] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0112] The above embodiments are merely illustrative of several implementations of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the appended claims.

Claims

1. A method for real-time compensation and adaptive control of excitation light intensity in a fluorescence scanner for photobleaching, characterized in that, Includes the following steps: During the full-slice fluorescence scanning process, the fluorescence image of the currently acquired tile is obtained tile by tile, and the fluorescence signal intensity of the current tile is extracted from the fluorescence image; The fluorescence signal intensity is input into a sliding window, and the decay time constant of the photobleaching exponential decay model is updated in real time using least squares fitting, including: Before data acquisition at the beginning of the scan The fluorescence signal intensity of each effective tile is fitted to the initial decay time constant; Maintain length in subsequent scans A sliding window for each tile, for viewing the inside of the window. and Perform linear regression fitting to obtain the fitted decay time constant; According to the exponential smoothing formula Update the current decay time constant, where The updated decay time constant, The fitted decay time constant is, The decay time constant of the previous update cycle. The preset smoothing coefficient; When the goodness-of-fit coefficient of the sliding window fitting continuously reaches a preset frame number less than a preset residual threshold, the model is deemed to have failed and the decay time constant is reset to the initial decay time constant; wherein the photobleaching exponential decay model is: , For the first The fluorescence signal intensity of each tile, The initial fluorescence signal intensity, The scanned tile sequence number. The excitation time of a single watt is the time required for light to be emitted. The decay time constant is the value of the time. The signal attenuation of the next watt is predicted based on the updated photobleaching exponential attenuation model, and the excitation intensity compensation formula is applied: Calculate the target excitation light intensity of the next watt, where For the first The target excitation light intensity of each tile The initial excitation light intensity; The output power of the excitation light source is adjusted watt by watt according to the target excitation light intensity; when the target excitation light intensity reaches the maximum safe power of the excitation light source, the excitation light intensity is kept constant at the maximum safe power, and the exposure time of the camera is increased in conjunction to maintain the image signal-to-noise ratio; When the exposure time reaches the preset upper limit and the signal-to-noise ratio continues to decrease, a photobleaching warning is triggered. The method also includes three-stage linkage control of excitation light intensity, exposure time, and scanning frame rate: First stage: When the target excitation light intensity does not reach the maximum safe power, increase the output power of the excitation light source watt by watt while keeping the exposure time unchanged; Second stage: When the target excitation light intensity reaches the maximum safe power and the image signal-to-noise ratio meets the standard, maintain the excitation light intensity at the maximum safe power, and increase the exposure time watt by watt to the maximum allowable exposure time; The third stage: When the exposure time reaches the maximum allowable exposure time and the signal-to-noise ratio continues to decrease, triggering the photobleaching warning, a frame rate reduction strategy is executed to reduce the scanning frame rate to extend the frame period and increase the corresponding maximum allowable exposure time, so as to maintain the image signal-to-noise ratio and slow down the photobleaching rate by further increasing the exposure time; if the signal-to-noise ratio continues to be lower than the minimum signal-to-noise ratio threshold after the frame rate reduction reaches the preset confirmation frame number, the corresponding area is marked as a photobleaching restricted area and the scanning of the remaining areas continues.

2. The method for real-time compensation and adaptive control of excitation light intensity for photobleaching in a fluorescence scanner as described in claim 1, characterized in that, Extracting the fluorescence signal intensity of the current tile from the fluorescence image includes: The fluorescence image is subjected to adaptive threshold segmentation to obtain the foreground region, and morphological opening is performed on the foreground region for noise reduction and edge artifact removal to obtain the region of interest of the fluorescence signal. Calculate the grayscale uniformity and signal-to-noise ratio within the region of interest; When the grayscale uniformity is not greater than a preset uniformity threshold and the signal-to-noise ratio is not less than a preset signal-to-noise ratio threshold, the average grayscale value within the region of interest is extracted as the fluorescence signal intensity of the current tile. When the grayscale uniformity is greater than the preset uniformity threshold or the signal-to-noise ratio is less than the preset signal-to-noise ratio threshold, the peak grayscale value at the preset high-resolution point in the region of interest is extracted as the fluorescence signal intensity of the current tile.

3. The method for real-time compensation and adaptive control of excitation light intensity for photobleaching in a fluorescence scanner as described in claim 1, characterized in that, Adjusting the output power of the excitation source watt by watt according to the target excitation light intensity includes: For the first The target excitation light intensity of each tile is subject to a single-tile adjustment step size limit: ,in For the first The excitation light intensity of each tile The maximum adjustment step size for a single tile is preset; The excitation light intensity after applying a step size limit is sent to the excitation source driver for power adjustment.

4. The method for real-time compensation and adaptive control of excitation light intensity for photobleaching in a fluorescence scanner as described in claim 1, characterized in that, In a multi-dye, multi-channel fluorescence scanning scenario, the method includes: The sequential excitation mode is adopted, and images are acquired, decay time constants are updated, and target excitation light intensity of each channel is calculated independently. When the total available power of the excitation source is limited, the power allocation weight is calculated based on the reciprocal of the decay time constant of each fluorescence channel, according to the formula. Allocate an excitation power budget to each fluorescence channel, where The power budget allocated to the current channel. Total available power, This represents the decay time constant of the current channel. Let be the decay time constant of the j-th channel; The acquired multi-channel images are collaboratively processed to perform spectral crosstalk correction post-processing.

5. The method for real-time compensation and adaptive control of excitation light intensity for photobleaching in a fluorescence scanner as described in claim 1, characterized in that, When a scanned area needs to be rescanned, the method further includes: Before rescanning, a preview frame of the rescanning area is acquired with a preview excitation power lower than the initial excitation light intensity to obtain the current measured signal intensity; Based on the ratio of the preview excitation power to the initial excitation light intensity, the current measured signal intensity is equivalently normalized to the reference signal intensity; The fluorescence recovery rate is calculated by comparing the reference signal intensity with the original signal intensity of the rescanned region during the original acquisition. When the fluorescence recovery rate is greater than the first recovery threshold, the initial scan parameters are used for rescanning; when the fluorescence recovery rate is less than the second recovery threshold, the compensation scan parameters recorded before rescanning are used for rescanning; when the fluorescence recovery rate is between the second recovery threshold and the first recovery threshold, linear interpolation is performed between the initial scan parameters and the compensation scan parameters based on the fluorescence recovery rate to obtain the rescanning parameters and perform rescanning.

6. An apparatus for implementing the real-time compensation and adaptive control method for photobleaching of a fluorescence scanner according to any one of claims 1 to 5, characterized in that, include: The photobleaching modeling module is used to acquire the fluorescence image of the current acquisition tile one by one during the full-slice fluorescence scanning process, extract the fluorescence signal intensity of the current tile from the fluorescence image, input the fluorescence signal intensity into the sliding window, and use least squares fitting to update the decay time constant of the photobleaching exponential decay model in real time. The compensation decision module is used to predict the signal attenuation of the next watt based on the updated photobleaching exponential attenuation model, and to calculate the target excitation intensity of the next watt according to the excitation intensity compensation formula. The excitation light source control module is used to adjust the output power of the excitation light source watt by watt according to the target excitation light intensity in order to compensate for the attenuation of fluorescence signal caused by photobleaching; The exposure linkage module is used to increase the camera's exposure time to maintain the image signal-to-noise ratio when the target excitation light intensity reaches the maximum safe power of the excitation light source, and to trigger a photobleaching warning when the exposure time reaches a preset upper limit and the signal-to-noise ratio continues to decrease. The multi-channel management module is used to independently maintain the photobleaching exponential decay model and compensation parameters for each fluorescence channel during multi-dye multi-channel fluorescence scanning, and to perform multi-channel power budget management.

7. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to execute the fluorescence scanner photobleaching real-time compensation and excitation light intensity adaptive control method according to any one of claims 1 to 5.

8. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements the real-time compensation and adaptive control method for photobleaching of a fluorescence scanner as described in any one of claims 1 to 5.

Images