Continuous scanning microscopy methods, equipment, media, and products based on event stream time-series inversion

Abstract

This application discloses a continuous scanning microscopic imaging method, device, medium, and product based on event stream time-series inversion, relating to the fields of photoelectric detection and microscopic imaging. The method includes: acquiring continuous time-varying light signals and continuous relative motion states of a target sample; determining the original event stream by performing asynchronous detection simultaneously with continuous scanning based on the continuous relative motion state; determining the correspondence between motion parameters and event timestamps of discrete events based on the motion parameters corresponding to the original event stream; correcting the discrete events in the original event stream based on the continuous relative motion state and a spatiotemporal mapping model; dividing the time axis into integration windows according to the evolution law of the time-varying light signals; and performing integration operations on the corrected event stream within the corresponding integration window to generate a continuously scanned reconstructed image. This application can achieve high-fidelity microscopic imaging while maintaining continuous motion, improving the time duty cycle and reconstruction efficiency of scanning imaging.

Description

Technical Field

[0001] This application relates to the fields of photoelectric detection and microscopic imaging technology, and in particular to a continuous scanning microscopic imaging method, device, medium and product based on event flow time-series inversion. Background Technology

[0002] In current microscopic imaging, large-field-of-view scanning is typically achieved by capturing images from each field of view and then stitching them together. Specifically, the platform or sample first moves to a certain field of view, stops moving at that position, completes the exposure and acquisition, then moves to the next position and repeats the process. Finally, multiple independent field-of-view images are stitched together to obtain a large-area imaging result. However, this "start-stop" acquisition method has significant shortcomings in large-area, high-throughput imaging scenarios: First, the mechanical platform needs to be frequently started and stopped, resulting in significant time loss due to mechanical movement and low acquisition efficiency; second, in high-speed or large-area scanning scenarios, the effective acquisition time of the system is relatively low; third, each field of view is exposed independently at different physical moments before being stitched together, which easily leads to boundary discontinuities, inconsistent brightness, and uneven illumination artifacts. These problems are particularly prominent in high-throughput imaging in the biomedical field, such as continuous scanning of tissue sections, large-area screening of pathological samples, quantitative analysis of cell populations, and large-scale observation of subcellular structures, all of which require maintaining high acquisition efficiency and high structural fidelity within a large field of view.

[0003] On the other hand, time-delay integration imaging technology can achieve signal alignment and accumulation along the motion direction during continuous scanning, thereby improving signal utilization efficiency and imaging quality. However, traditional time-delay integration cameras typically rely on specialized charge transfer structures and specific readout methods. Their working principle involves step-by-step synchronous transfer and accumulation within the detector along the target's motion direction, thus exhibiting strong dependence on detector structure, motion direction, and operating mode. Furthermore, the integration level of traditional time-delay integration cameras is usually limited by the fixed number of charge transfer levels within the detector, commonly 96 or 128 levels. Consequently, their effective integration depth and subsequent reconstruction flexibility are limited by the device structure itself. When the sample undergoes dynamic changes, the fixed accumulation of traditional time-delay integration cameras easily generates motion artifacts.

[0004] In recent years, event-based detectors, with their asynchronous sampling and high-precision timestamp recording capabilities, have offered new technological possibilities for microscopic reconstruction under continuous scanning conditions. Unlike traditional frame detectors that output fixed-exposure images, event detectors output an event stream containing spatial coordinates, timestamps, and polarity, thus preserving the signal's temporal changes in greater detail. This data format means that imaging information at different physical moments during continuous scanning is no longer simply compressed into a single exposure image, but can be recorded as discrete events, providing a foundation for subsequent time-based reconstruction processing.

[0005] However, current event camera microscopy schemes are mostly used for local dynamic detection, brightness change recording, or motion trajectory analysis, and do not fully utilize the potential for spatiotemporal reconstruction of event streams in continuous scanning microscopy. Based on this background, there is a need to provide a new continuous scanning microscopy system and method that allows timestamp information in the event stream to go beyond being merely supplementary recording information, and instead serve as a core basis for continuous scanning reconstruction along with scanning motion parameters. Summary of the Invention

[0006] The purpose of this application is to provide a continuous scanning microscopy imaging method, device, medium and product based on event flow time-series inversion, which can complete high-fidelity microscopy imaging while maintaining continuous motion, and improve the time duty cycle and reconstruction efficiency of scanning imaging.

[0007] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a continuous scanning microscopic imaging method based on event flow temporal inversion, comprising: Acquire continuous time-varying optical signals and continuous relative motion states of the target sample; the continuous relative motion state includes the computational velocity vector acquired when the target sample is in a continuous scanning state; the computational velocity vector is used to characterize the spatiotemporal correlation with the time-varying optical signals; Based on the continuous relative motion state, asynchronous detection is performed while continuously scanning to determine the original event stream; the original event stream includes a series of discrete events that start independently over time. Based on the motion parameters corresponding to the original event stream, determine the correspondence between the motion parameters and the event timestamps of discrete events; Based on the continuous relative motion state and the spatiotemporal mapping model, the discrete events in the original event stream are corrected to obtain the corrected event stream; the spatiotemporal mapping model is determined based on the correspondence between motion parameters and the event timestamps of discrete events. The time axis is divided into integral windows according to the evolution law of the time-varying light signal to obtain the integral window; The corrected event stream is integrated within the corresponding integration window to generate a continuously scanned reconstructed image.

[0008] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described continuous scanning microscopic imaging method based on event flow time-series inversion.

[0009] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned continuous scanning microscopic imaging method based on event flow time-series inversion.

[0010] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned continuous scanning microscopic imaging method based on event flow time-series inversion.

[0011] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a continuous scanning microscopic imaging method, device, medium, and product based on event stream time-series inversion. By acquiring continuous time-varying light signals and continuous relative motion states of the target sample, the time duty cycle can be improved, efficiency loss caused by mechanical start-stop processes can be reduced, and continuous scanning imaging efficiency can be improved. Based on the continuous relative motion state, asynchronous detection is performed during continuous scanning to determine the original event stream. Based on the motion parameters corresponding to the original event stream, the correspondence between the motion parameters and the event timestamps of discrete events is determined, which allows events triggered by the same object point at different physical times to be realigned and accumulated in a unified coordinate system. Based on the continuous relative motion state and spatiotemporal mapping model, discrete events in the original event stream are corrected. The time axis is divided into integration windows according to the evolution law of the time-varying light signals. The corrected event stream is integrated within the corresponding integration window to generate a continuous scanning reconstructed image. This can reduce boundary problems caused by independent exposure per field of view and subsequent stitching, improve the continuity and overall consistency of large field-of-view images, and is beneficial for dynamic process observation and analysis under continuous scanning conditions. Therefore, this application can achieve high-fidelity microscopic imaging while maintaining continuous motion, thereby improving the time duty cycle and reconstruction efficiency of scanning imaging. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of a continuous scanning microscopy imaging method based on event flow temporal inversion. Figure 2 This is a schematic diagram of the optical path of the continuous scanning microscopy imaging method based on event flow temporal inversion. Figure 3 This diagram illustrates the reconstruction steps of a continuous scanning microscopy imaging method based on event flow temporal inversion applied within a system. Figure 4 A schematic diagram of the reconstructed image imaging results; Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application.

[0014] Reference numerals: Illumination module-10, Structured light modulation and displacement module-20, Microscopic magnification module-30, Three-dimensional continuous scanning motion module-40, Event detection module-50, Motion parameter acquisition and synchronization control module-60, and Time series inversion and reconstruction module-70, First tube lens-31, Beam splitter element-32, Microscope objective lens-33, Second tube lens-34. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] The basic concept of this application is as follows: During continuous microscopic scanning, a mapping relationship between event time sequence and spatial displacement is constructed by utilizing the high-precision timestamps output by the event detector and the known or measurable scanning motion parameters of the system; through time-series inversion and redistribution, events recorded by different pixels at different physical times at the same object point are remapped to a unified spatial coordinate system for accumulation, thereby realizing continuous scanning microscopic imaging similar to time delay integration; and further, a posteriori time re-division binning is performed on the same continuous acquisition event stream to form time slice reconstruction results under different time windows.

[0017] Compared with traditional time-delay integration cameras, the technical approach of this application has the following significant characteristics: First, this application does not rely on a dedicated charge transfer detector structure, but instead utilizes the asynchronous event stream output by the event detector and the back-end time-series inversion redistribution algorithm to achieve a digital time-delay-like integration effect, thus making it easier to integrate with a continuous microscopic scanning system; Second, the effective integration depth of this application is not limited by the fixed integration stages of traditional time-delay integration cameras, but is determined by the number of available pixel rows in the motion direction of the event detector, which can fully utilize 720 pixel rows for event accumulation, significantly higher than the 96 or 128 integration stages commonly found in traditional time-delay integration cameras; Third, this application can not only output continuously scanned accumulated reconstructed images, but also perform posterior reconstruction through different time windows on the same continuously acquired event stream, thereby maintaining a large field of view for continuous scanning while outputting high temporal resolution time slice images with significantly reduced motion blur, a capability that traditional time-delay integration cameras typically do not possess.

[0018] Therefore, this application can solve the problems of low acquisition efficiency, obvious stitching artifacts, difficulty in high-fidelity reconstruction of continuous large field of view and difficulty in taking into account dynamic information in existing continuous scanning microscopy. It is especially suitable for imaging scenarios in the biomedical field that require large field of view, high throughput and high fidelity at the same time, such as continuous tissue scanning imaging, rapid screening of pathological samples, large-scale quantitative analysis of cells and subcellular structures, and continuous observation and a posteriori analysis of rapid biological dynamic processes.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] In one exemplary embodiment, such as Figure 1 As shown, a continuous scanning microscopic imaging method based on event flow temporal inversion is provided, including: Step 100: Acquire the continuous time-varying light signal and continuous relative motion state of the target sample. The continuous relative motion state includes the computational velocity vector acquired when the target sample is in a continuous scanning state; the computational velocity vector is used to characterize the spatiotemporal correlation with the time-varying light signal; the time-varying light signal is used to characterize the local brightness change information of the target sample after local brightness modulation under microscopic illumination; the time-varying light signal is a time-varying structured light signal containing spatial coordinates and time coordinates.

[0021] Step 200: Based on the continuous relative motion state, asynchronous detection is performed during continuous scanning to determine the original event stream. The original event stream consists of a series of discrete events that start independently over time.

[0022] Discrete events include: event space coordinate information Event timestamp Event polarity information and the time-varying optical signal; wherein, ; This represents a positive event triggered by a rise in light intensity threshold; This represents a negative event triggered by a decrease in light intensity threshold.

[0023] Step 300: Based on the motion parameters corresponding to the original event stream, determine the correspondence between the motion parameters and the event timestamps of the discrete events.

[0024] Step 400: Based on the continuous relative motion state and the spatiotemporal mapping model, the discrete events in the original event stream are corrected to obtain the corrected event stream. The spatiotemporal mapping model is determined based on the correspondence between motion parameters and the event timestamps of discrete events.

[0025] The expression corresponding to the corrected event flow is: .

[0026] in, The corrected event stream; These are discrete events; This represents the corrected position of the discrete event in the target coordinate system. The timestamp of the event; This refers to the polarity information of the event. It is a spatiotemporal mapping model; This refers to event space coordinate information; It is a state of continuous relative motion.

[0027] Spatiotemporal mapping model Used to characterize states based on continuous relative motion The event coordinate correction relationship is used to determine the event space coordinate information of discrete events. And the continuous relative motion state corresponding to the event timestamp, which maps events triggered at different times to the same target coordinate system.

[0028] Under the condition of uniform scanning in one direction, the spatiotemporal mapping model can degenerate into a linear reverse displacement compensation relationship, i.e. ; This is the initial timestamp.

[0029] Under conditions of variable speed, variable direction, curve scanning, or closed-loop feedback scanning, the spatiotemporal mapping model can be represented as a generalized coordinate correction function determined based on motion trajectory integration, piecewise motion parameters, and position feedback interpolation.

[0030] Step 500: Divide the time axis into integration windows according to the evolution law of the time-varying light signal to obtain the integration window.

[0031] In one embodiment, the time axis is divided into integration windows according to the evolution law of the time-varying light signal to obtain an integration window, including: Based on the local brightness change information during continuous scanning, the periodic, phase or spatial features of the corresponding variable illumination mode are extracted, and the time dimension of the continuous event stream is divided into an integral window sequence corresponding to the preset illumination state to obtain the integral window; the local brightness change information is determined based on the preset known structured light modulation drive.

[0032] Step 600: Integrate the corrected event stream within the corresponding integration window to generate a continuously scanned reconstructed image.

[0033] The frame-level discrete integral mathematical expression for the reconstructed image from continuous scanning is: .

[0034] in, This is the frame-level discrete integral form corresponding to the continuously scanned reconstructed image; This is the starting window for the integration window; This is the closing window for the integration window; These are the spatial coordinates contained in the time-varying optical signal; It is the discrete Dirac function; It is a spatiotemporal mapping model; This refers to event space coordinate information; It is a state of continuous relative motion; The time coordinates are contained in the time-varying optical signal.

[0035] In a state of continuous motion, a mask is applied to discrete events and reconstructed to obtain the corresponding continuously scanned reconstructed image: .

[0036] .

[0037] in, A continuously scanned reconstructed image for applying a mask to discrete events and reconstructing them; For masking; This represents the polarity information of the event.

[0038] In one embodiment, this application employs a continuous scanning time-delay integration microscopy system based on an event detector to implement the aforementioned continuous scanning microscopy method based on event stream temporal inversion. This system is used to perform temporal inversion and spatial coordinate reallocation of events recorded at different physical moments during continuous scanning imaging, combining the asynchronous event stream output by the event detector and system motion parameters, thereby achieving continuous scanning microscopy reconstruction similar to time-delay integration.

[0039] like Figure 2 As shown, the system includes at least: an illumination module 10, a structured light modulation and displacement module 20, a microscopic magnification module 30, a three-dimensional continuous scanning motion module 40, an event detection module 50, a motion parameter acquisition and synchronization control module 60, and a time-series inversion and reconstruction module 70. These modules can be integrated into a single system or connected in a separate structure; the spatial arrangement of the modules does not constitute a limitation on the scope of protection of this application, as long as they can achieve their corresponding functions.

[0040] The illumination module 10 provides microscopic illumination to the target sample and generates a sample response light signal that can be sensed by the event detection module 50. The illumination module 10 includes an excitation source. The excitation source can be a single-mode laser, a multi-mode laser, a light-emitting diode, a supercontinuum source with a filter unit, or other light sources capable of providing stable illumination. The illumination module 10 can output single-wavelength illumination or multi-wavelength illumination; in multi-color imaging scenarios, multiple light sources can be directly combined, combined via optical fiber, or combined using a beam splitter. The specific device form of the illumination module 10 is not the only limitation of this application, as long as it can provide stable incident illumination to the target sample during continuous scanning.

[0041] The structured light modulation and displacement module 20 is used to form local brightness changes that are more suitable for event detection during continuous scanning, and to make the illumination pattern or sample form a continuous relative displacement with respect to the detection coordinate system, thereby providing input conditions for subsequent event stream acquisition, time series inversion and coordinate reassignment.

[0042] The structured light modulation and displacement module 20 includes a structured light modulation unit and a displacement driving unit. The structured light modulation unit is used to form multifocal patterns, discrete spot arrays, sparse structured light patterns, or other equivalent localized temporal modulation patterns. The structured light modulation unit can employ digital micromirror devices, spatial light modulators, microlens arrays, diffractive optical elements, speckle generators, or other devices capable of forming spatially discrete and temporally controllable illumination distributions. The aforementioned structured light modulation methods can generate periodic multifocal illumination, as well as quasi-random, pseudo-random, or other forms of localized modulation patterns.

[0043] The displacement driving unit is used to drive a continuous relative displacement between the modulation pattern, the sample, or the detection coordinate system. The displacement driving unit can preferably be a translation stage, a rotation platform, a galvanometer deflection assembly, a MEMS displacement device, an optical scanner, an acousto-optic deflector, an electro-optic deflector, or other driving mechanisms capable of forming continuous relative motion. In some embodiments, the modulation pattern can be fixed while the sample moves continuously; alternatively, the sample can be fixed while the modulation pattern moves continuously, or both can move together. Any method that establishes a known or measurable continuous relative displacement relationship during the imaging process is applicable.

[0044] The structured light modulation and displacement module 20 serves to enhance event detection sensitivity through localized illumination variations, providing a foundation for highly sensitive event generation under continuous scanning conditions. The specific device form of this module does not constitute a core limitation of this application.

[0045] The microscopic magnification module 30 is used to achieve microscopic imaging, magnification, and necessary beam splitting coupling between the sample and the detection plane. The microscopic magnification module 30 includes a first tube lens 31, a beam splitter 32, a microscope objective 33, and a second tube lens 34. The beam splitter 32 is used to separate or couple the illumination light and the sample response light; the microscope objective 33 is used to collect the sample response light signal and achieve primary microscopic magnification; the first tube lens 31 or the second tube lens 34 is used in conjunction with the microscope objective 33 to form a magnification of a specific magnification.

[0046] The microscopic magnification module 30 can be an inverted microscope, an upright microscope, a custom-designed microscopic optical path, or other optical platforms suitable for microscopic imaging. The beam-splitting element 32 can be a dichroic mirror, a beam-splitting prism, a filter group, or other elements capable of achieving optical path separation. The microscopic magnification module 30 may further include a relay lens, an aperture, a filter, or other auxiliary imaging units.

[0047] The three-dimensional continuous scanning motion module 40 is used to keep the sample in a continuous scanning state during acquisition, thereby getting rid of the position-by-position stop-and-go imaging mode and realizing a large field of view, high throughput and continuous sampling.

[0048] The three-dimensional continuous scanning motion module 40 is used to drive the sample to continuously translate along a preset direction to achieve continuous acquisition of a large field of view in a two-dimensional plane. Furthermore, the three-dimensional continuous scanning motion module 40 can also perform inter-layer displacement in a direction perpendicular to the plane scanning direction to achieve three-dimensional volume data acquisition.

[0049] The three-dimensional continuous scanning motion module 40 can employ an electric translation stage, a linear motor platform, a piezoelectric displacement stage, an encoder feedback platform, or other motion mechanisms capable of achieving continuous and measurable displacement output. The motion parameters output by the module may include, but are not limited to: preset speed, real-time speed, displacement, direction, acceleration, motion start time, reference time, and corresponding encoder feedback information.

[0050] The three-dimensional continuous scanning motion module 40 drives the sample itself, or it can drive the objective lens, detector, illumination pattern or its equivalent image plane, as long as the correspondence between continuous scanning motion and event flow can be established.

[0051] The event detection module 50 is used to asynchronously detect the sample response light signal during continuous scanning and output a raw event stream. The raw event stream includes at least one or more of the following information: event occurrence timestamp, event spatial coordinates, event polarity, and auxiliary information related to changes in event intensity. Among them, event polarity is used to characterize whether the brightness increases or decreases, timestamp is used to characterize the physical moment when the event occurs, and spatial coordinates are used to characterize the position of the event on the detector.

[0052] The event detection module 50 can employ an event camera, or it can use an asynchronous time-series image sensor, a dynamic vision sensor, a hybrid sensor with event output capability, or other detectors capable of outputting discrete spatiotemporal event data streams. The key is that the detector can output asynchronous event information corresponding to continuous scanning imaging. The event detection module 50 can operate independently or be combined with a conventional frame detector to form a composite detection system for system calibration, parameter verification, or reference image output.

[0053] The motion parameter acquisition and synchronization control module 60 is used to acquire motion parameters corresponding to the original event stream during continuous scanning and to perform unified or distributed synchronization control on various functional modules of the system.

[0054] The motion parameter acquisition and synchronization control module 60 includes at least one or more of the following functions: Obtain preset scan speed parameters; read encoder feedback displacement or motion state information; record scan start time and reference time; synchronize motion state information with event stream timestamps; output speed, displacement, direction and time parameters for subsequent time-series inversion.

[0055] Unified timing control is performed on the lighting module 10, the structured light modulation and displacement module 20, the three-dimensional continuous scanning motion module 40, and the event detection module 50.

[0056] The motion parameter acquisition and synchronization control module 60 can be implemented using a data acquisition card, controller, FPGA, microprocessor, industrial control unit, or a combination thereof. This module can output motion parameters in real time, or output motion information required for offline calibration after acquisition.

[0057] The function of the motion parameter acquisition and synchronization control module 60 is to establish a one-to-one correspondence between "event time - spatial position - motion parameters", so that the subsequent time series inversion and reconstruction module 70 can perform coordinate redistribution and cumulative reconstruction accordingly.

[0058] The temporal inversion and reconstruction module 70 is used to perform temporal inversion, coordinate reassignment and cumulative reconstruction on the original event stream based on the event timestamp and motion parameters, so as to output continuously scanned microscopic reconstructed images.

[0059] The time series inversion and reconstruction module 70 includes at least the following units: The time-series inversion unit is used to determine the unified reference time corresponding to different events based on the relationship between event timestamps and continuous scanning motion.

[0060] The coordinate reassignment unit is used to determine the physical scale transformation relationship between the target surface, sample surface and event coordinates of the event detection module 50 based on motion parameters and system magnification, and to calculate the corrected position of each event in the unified target coordinate system.

[0061] The time window reconstruction unit is used to perform posterior segmentation of the same continuous acquisition event stream according to different time windows, thereby decoupling the sampling rate from the equivalent exposure time.

[0062] The cumulative reconstruction unit is used to spatially accumulate the corrected events according to the segmented time window, thereby outputting a continuously scanned reconstructed image.

[0063] Optional filtering units are available for filtering event streams by polarity, spatial region, time period, or other rules to generate different reconstruction results.

[0064] The timing inversion and reconstruction module 70 can directly perform event domain remapping based on the event stream; alternatively, it can first divide the event stream into multiple event frames according to time, and then perform coordinate compensation and accumulation on the event frames. Both of these processing methods are acceptable.

[0065] The temporal inversion and reconstruction module 70 directly determines whether a reconstruction effect similar to time delay integration can be achieved under continuous scanning conditions. Through this module, events recorded by different pixels at different physical times for the same object point can be remapped to a unified spatial coordinate system for accumulation, thereby obtaining a large field of view, high fidelity, and continuously scanned reconstructed image.

[0066] During operation, the illumination module 10 illuminates the sample, the structured light modulation and displacement module 20 creates local brightness variations on the sample surface that are more suitable for event detection, the microscopic magnification module 30 is responsible for imaging the sample response light signal to the event detection module 50, the three-dimensional continuous scanning motion module 40 drives the sample or equivalent imaging object to maintain a continuous scanning state, the motion parameter acquisition and synchronization control module 60 synchronously records the motion parameters and coordinates the working sequence of each module, and the time series inversion and reconstruction module 70 completes coordinate redistribution and cumulative reconstruction based on the original event flow and motion parameters, and finally outputs a large field-of-view continuous scanning microscopic image similar to time delay integral.

[0067] In the process of continuous microscopic scanning, this application utilizes the asynchronous event stream output by the event detector and the known or measurable motion parameters of the system to perform time-series inversion and reconstruction of events recorded at different physical moments, thereby achieving high duty cycle, large field of view, and high-fidelity continuous scanning microscopic imaging similar to time delay integration.

[0068] Based on the above system composition, starting from structured light modulation, the system sequentially undergoes three-dimensional continuous relative motion, event stream acquisition, motion parameter acquisition, temporal inversion, coordinate redistribution, time window reconstruction, and event accumulation, ultimately outputting a large field-of-view continuous scanning image and / or a high temporal resolution temporal image.

[0069] The core technical concept of this application does not lie in simply using an event detector for acquisition, nor in ordinary scanning stitching, but rather in: during continuous scanning imaging, utilizing the high-precision time information carried by the event stream and the known or measurable continuous scanning motion parameters of the system, remapping events recorded by different pixels at different physical times for the same object point onto a unified spatial coordinate system for accumulation, thereby achieving continuous scanning reconstruction similar to time delay integration; and further, performing posterior reconstruction of the same continuously acquired event stream at different time windows. Compared to the traditional intermittent scanning method of "move-stop-exposure," this application can achieve high-fidelity microscopic imaging while maintaining continuous motion, reducing the time loss caused by mechanical start-stop, improving the time duty cycle and reconstruction efficiency of continuous scanning imaging, and balancing large field-of-view image acquisition with high temporal resolution temporal reconstruction.

[0070] Without departing from the concept of this application, each of the above modules can be implemented using other equivalent devices, equivalent connection methods, or equivalent control methods. For example, the structured light modulation and displacement module 20 can be implemented using different combinations of modulation patterns and scanning mechanisms; the three-dimensional continuous scanning motion module 40 can be implemented using sample motion, optical path motion, or image plane motion; the event detection module 50 can use different types of asynchronous event detectors; and the time-series inversion and reconstruction module 70 can be implemented in real-time hardware or offline software. Any system configuration that can achieve time-series inversion and coordinate reassignment based on event flow and continuous scanning motion parameters, thereby completing a similar time-delay integral continuous scanning reconstruction, falls within the protection scope of the product solution described in this application.

[0071] like Figure 3 As shown, the specific operation process of image reconstruction using the above-mentioned system in the continuous scanning microscopic imaging method based on event flow temporal inversion is as follows: Step 1: Establish continuous scanning imaging input conditions.

[0072] The target sample is microscopically illuminated by an illumination module, and a structured light modulation and displacement module creates local brightness variations on the sample surface suitable for event detection. These local brightness variations can manifest as multifocal patterns, discrete spot arrays, sparse structured light patterns, or other equivalent local temporal modulation patterns, thereby generating time-varying light signals on the target sample that are more suitable for the event detector's response during continuous scanning. The role of structured light modulation is not solely to complete image restoration, but to overcome the physical limitation of event detectors in pure translational continuous scanning, which can only capture gradient information in the motion direction (i.e., only generate unidirectional gradient maps). This allows the morphological information of the target sample's surface in all directions to be converted into high-frequency time-varying light signals and trigger an event response when passing through the modulation field, providing a more stable data foundation for time-series inversion and continuous accumulation reconstruction. The continuously time-varying structured light signal is generated by local brightness modulation. .

[0073] Step 2: Establish continuous scanning imaging status.

[0074] A continuous scanning motion module drives the sample, structured light modulation pattern, detection field of view, or a combination thereof to generate continuous relative motion, ensuring the target sample is always in a continuous scanning state during acquisition, rather than the traditional stop-and-go acquisition state of field-by-field. The continuous scanning motion module drives the sample-carrying platform to continuously translate along a preset direction at an approximately constant speed, thus forming a continuous displacement of the sample relative to the detection coordinate system. Alternatively, continuous displacement can be achieved by driving the structured light modulation pattern, or by having multiple motion units work together to form an equivalent continuous relative motion between the sample and the detection system.

[0075] Establishing continuous sampling states without interrupting the scanning motion provides precise kinematic constraints for subsequent event flow analysis, thus forming a traceable and computable mapping foundation between event timestamps and spatial displacements. This step establishes continuous relative motion states with definite spatiotemporal correlations. That is, the operation speed vector.

[0076] Step 3: Acquire the raw event stream during continuous scanning.

[0077] Following the aforementioned continuous relative motion state, while continuous scanning is being performed, the event detection module asynchronously detects the sample response light signal and outputs the raw event stream in real time. The original event stream consists of a series of discrete events that start independently over time: .

[0078] Each discrete event At least include event space coordinates Event timestamp and event polarity information (+1 represents a positive event triggered by the light intensity rising threshold, and -1 represents a negative event triggered by the light intensity falling threshold.) If necessary, auxiliary information related to local brightness changes, i.e., continuously time-varying structured light signals, may also be included. Because the event detection module operates asynchronously, it does not need to synchronously expose the entire field of view at a fixed frame rate during continuous scanning. Instead, it outputs discrete events when local brightness changes meet the triggering conditions. Therefore, the system can maintain high temporal resolution data recording capabilities during continuous motion and record rich temporal information during a single continuous scan. The output of this step is not a complete frame image formed within a fixed exposure time, but a raw event stream containing high-precision temporal and spatial location information. This raw event stream provides the data foundation for subsequent temporal inversion, coordinate reassignment, and similar time-delay integral reconstruction. The output of this step is: Raw Event Stream. .

[0079] Step 4: Obtain the continuous scan motion parameters and their correspondence with event time.

[0080] The motion parameter acquisition and synchronization control module acquires motion parameters corresponding to the original event stream during continuous scanning and establishes a correspondence between these parameters and event timestamps. Motion parameters include, but are not limited to, sample stage speed parameters, pattern scanning speed parameters, real-time position feedback parameters, motion start time, reference time, motion direction parameters, and motion trajectory parameters. In one implementation (e.g., an open-loop control architecture), the system directly uses a pre-set continuous scanning speed as the motion parameter input. In another preferred implementation (e.g., a closed-loop feedback architecture), the system acquires position data in real-time during continuous scanning via an encoder, closed-loop position sensor, or other position feedback device, and synchronizes this data with the event stream timestamps. Furthermore, in some implementations, preset parameters and real-time feedback parameters can be combined to correct the spatiotemporal relationship during continuous scanning. The purpose of this step is to establish a parameterized correspondence between the event occurrence time and the sample spatial displacement, providing a basis for subsequent event time-series inversion and spatial coordinate correction. The output of this step is a spatiotemporal mapping model containing a high-precision timestamp and an absolute / relative spatial position coordinate correspondence. .

[0081] Step 5: Perform time-series inversion and coordinate reassignment based on event timestamps and motion parameters.

[0082] The temporal inversion and reconstruction module receives the original event stream and corresponding motion parameters, and performs temporal inversion and spatial coordinate reallocation on the original event stream based on the timestamps of each event and the relative motion relationships in continuous scanning. The system presets a reference time. (For example, the scan start time or a specific exposure reference time) and establish a unified target spatiotemporal coordinate system based on this, and combine the continuous relative motion state according to the time difference between the timestamp of each event and the reference time. and spatiotemporal mapping model For the original event stream The events in the target coordinate system are corrected, and the corrected positions of the events in the target coordinate system are calculated. The corrected event stream is obtained. : .

[0083] In the simplest case, where the mapping relationship between high-precision timestamps and absolute / relative spatial coordinates is linear and no additional correction is needed, the above changes can be simplified to: .

[0084] in, This is the start time of the scan.

[0085] Thus, events originally recorded at different physical times and from different detection locations are remapped back to a unified spatial coordinate system, resulting in a corrected event stream. In other implementations, time-series inversion can employ direct calculation based on a single velocity model, or it can use piecewise velocity models, trajectory compensation models, lookup table mapping, or other processing methods that can establish a correspondence between event time and spatial displacement. However, mathematically, all of these can be abstracted into... Although the same object point is recorded at different detection positions at different physical moments during continuous scanning, these events can be realigned in a unified spatial location through temporal inversion and coordinate correction using known or measurable motion relationships, providing a basis for subsequent continuous accumulation. The purpose of this step is to eliminate spatiotemporal misalignment caused by continuous scanning, enabling events generated by the same object point at different times to be superimposed in a unified coordinate system. The output of this step is: a corrected event stream that has completed spatiotemporal alignment and mapping to a unified target coordinate system. .

[0086] Step 6: Perform posterior reconstruction of the event stream based on different time windows.

[0087] After acquiring the correction event stream, the system precisely divides the time axis into integration windows based on the time-varying illumination pattern, i.e., the evolution law of the time-varying light signal, and completes the posterior reconstruction of the image within the corresponding integration window. Specifically, since the local brightness changes during continuous scanning are driven by known structured light modulation, the system extracts the periodic, phase, or spatial features of this time-varying illumination pattern, dividing the temporal dimension of the continuous event stream into an integration window sequence that strictly corresponds to a specific illumination state. - (From reference time) Starting with k integration windows, which can be closely spaced, spaced apart, or overlapping, the system independently accumulates or jointly calculates the correction events falling within each corresponding integration window to correctly demodulate the image component or the true grayscale image matching the specific lighting state. The output of this step is a sequence of integration windows corresponding to a specific time-varying lighting state. - .

[0088] Step 7: Perform integral reconstruction on the correction event stream to generate a continuously scanned reconstructed image. Figure 4 This is a schematic diagram of the image reconstruction results.

[0089] The temporal inversion and reconstruction module performs integration operations directly on the correction event stream obtained in step 5 within the corresponding integration window segmented in step 6, thereby outputting a final series of continuously scanned reconstructed images.

[0090] .

[0091] The above describes discrete event-level continuous integration. Coordinate reassignment directly affects the asynchronous original event stream. The system then directly performs continuous time integration and accumulation on the polarity or count value of the corrected events mapped to the same target spatial coordinates within the corresponding time window. For scenarios where high time precision is not required, a discrete version of the above formula can be used to achieve frame-level discrete integration. That is, coordinate reassignment is based on spatial compensation performed on event frames divided by time. The system then directly performs pixel-level discrete summation on multiple discrete event frame matrices that have completed spatial translation and alignment within the corresponding time window. .

[0092] .

[0093] Divide the time window into its length. To reconstruct images from continuous scans; In spatial coordinates Image The intensity value; That is ,at this time, It refers to coordinates ; This refers to the speed at which the translation stage moves. For summation indicator variables; This is the lower bound of the k-th time interval when summing; To meet the time range All events within; The coordinates are the coordinates of the event space coordinate information.

[0094] Since valid events triggered at the same point at different times have been realigned through temporal inversion and coordinate reallocation, valid signals corresponding to the same point can be superimposed at the target location during the accumulation process. Random noise without a consistent spatial correspondence is difficult to accumulate stably, thereby improving the signal-to-noise ratio and structural fidelity of the output image. Compared with the traditional field-by-field stop-and-go acquisition and stitching method, this application does not require separate stabilization, exposure, and readout processes at each location. Instead, it directly completes event stream acquisition and continuous reconstruction in a continuous motion state. Therefore, it can improve the imaging time duty cycle and reduce boundary discontinuities, brightness inconsistencies, and stitching artifacts caused by mechanical start-stop.

[0095] Furthermore, regarding the accumulation of integrals, events can be manually gated, meaning that a mask can be applied to the events before reconstruction. .

[0096] Regarding the specific form of the mask, taking the example of only superimposing positive events: .

[0097] Step 8: Output the combined result of the large field-of-view image and the time slice image.

[0098] Regarding the points window - It can simultaneously output high signal-to-noise ratio (SNR) large field-of-view continuous scan images accumulated based on all events or a longer time window, as well as high temporal resolution time slice images or dynamic image sequences obtained based on a shorter time window. Through this combined output method, this application can simultaneously achieve large field-of-view coverage, high-fidelity structure reconstruction, and rapid dynamic process resolution in a single continuous scan acquisition, and unify the high SNR requirement under a long time window with the high temporal resolution requirement under a short time window.

[0099] The benefits of this application are: First, achieve high duty cycle microscopic imaging under continuous scanning conditions.

[0100] Traditional microscopic scanning imaging typically employs a move-stationary-exposure-re-move working method, requiring repeated start-stop cycles between acquisition positions for the sample or imaging component. This results in significant mechanical motion time loss and a low effective acquisition time ratio. This application addresses this issue by using a three-dimensional continuous scanning motion module 40 to maintain continuous motion of the sample, illumination pattern, or its equivalent imaging object during acquisition. Combined with the asynchronous acquisition method of the event detection module 50, the imaging system eliminates the need to wait for stabilization at each acquisition position before exposure, thereby increasing the duty cycle, reducing efficiency losses caused by mechanical start-stop processes, and improving continuous scanning imaging efficiency.

[0101] Second, it enables the cumulative reconstruction of events, similar to time delay integration.

[0102] Traditional time-delay integration imaging typically relies on the synchronous transfer of charge within the detector along the direction of motion to achieve signal accumulation during continuous scanning. This application does not rely on traditional charge transfer structures. Instead, it utilizes the asynchronous event stream and its timestamp information output by the event detection module 50, combined with continuous scanning motion parameters, to perform temporal inversion and coordinate reassignment on the original event stream. This allows events triggered at the same point at different physical times to be realigned and accumulated in a unified coordinate system. Therefore, this application can achieve event accumulation and reconstruction effects similar to time-delay integration under continuous scanning conditions, realizing signal superposition and image enhancement during continuous scanning without employing a traditional TDI detector structure. Simultaneously, the corrected events can accumulate at a unified spatial location, enhancing effective events corresponding to the same point, while random noise is less likely to form stable superposition, thus further improving the signal-to-noise ratio and structural fidelity of the reconstruction results.

[0103] Third, reduce stitching boundary artifacts and improve the continuity of large field-of-view imaging.

[0104] Traditional stop-and-go wide-field-of-view imaging typically requires acquiring multiple independent fields of view separately and then stitching the images together. This can easily lead to inconsistent brightness, structural breaks, edge misalignment, or stitching artifacts at the field of view boundaries, thus affecting the overall imaging continuity and the quality of the wide-field-of-view image. This application employs a continuous scanning acquisition method and uses a temporal inversion and reconstruction module 70 to redistribute and cumulatively reconstruct events in a unified target coordinate system. This allows for the direct generation of wide-field-of-view reconstruction results from a continuous event stream, thereby reducing boundary problems caused by independent exposures for each field of view and subsequent stitching, and improving the continuity and overall consistency of the wide-field-of-view image.

[0105] Fourth, it balances wide field-of-view coverage with rapid dynamic analysis.

[0106] This application can not only output large field-of-view continuous scanning reconstructed images with high signal-to-noise ratio, but also construct high temporal resolution time slice images or dynamic image sequences around specific time points, thereby simultaneously achieving large field-of-view coverage, high-fidelity reconstruction, and rapid dynamic information acquisition in a single continuous scanning acquisition.

[0107] Compared with continuous scanning schemes that only output a single long exposure result, this application can reconstruct image results at different time scales based on the same original event stream according to different analysis needs, which is more conducive to the observation and analysis of dynamic processes under continuous scanning conditions.

[0108] Fifth, the system has a flexible configuration and is suitable for different continuous scanning microscopy scenarios.

[0109] The core of this application lies in utilizing the correspondence between continuous scanning motion parameters and event stream time information to perform time-series inversion, coordinate redistribution, and cumulative reconstruction of the original event stream. Therefore, it does not depend on a single detector structure or a single scanning implementation method.

[0110] Furthermore, the continuous scanning motion module 40 in the system mentioned in this application can be implemented using sample motion, illumination pattern motion, or equivalent relative motion; the event detection module 50 can be implemented using different types of event detectors; and the time-series inversion and reconstruction module 70 can be implemented either by direct remapping of the event domain or by first dividing the frame and then compensating. Therefore, this application has good system adaptability and scalability.

[0111] The relative motion during continuous scanning employs a regular, describable scanning trajectory, and directly uses the system's preset scanning speed or displacement parameters as the basis for temporal inversion and coordinate redistribution. Alternatively, the relative motion during continuous scanning can be uniform linear motion, segmented uniform motion, variable speed motion, curvilinear trajectory motion, or other continuous motions that can establish a spatiotemporal correspondence. Correspondingly, the motion parameters can be preset parameters, real-time measured motion parameters, segmented updated motion parameters, interpolated motion parameters, or equivalent motion parameters obtained from system control inversion. In some implementations, preset parameters can be combined with real-time feedback parameters to correct the spatiotemporal relationship during continuous scanning. Essentially, the purpose of this application can be achieved as long as a definite or estimable relationship can be established between the event occurrence time and the target spatial displacement during event acquisition.

[0112] Based on the event timestamps and continuous scanning motion parameters, direct temporal inversion is performed on the original events, and the corrected positions of the events in a unified coordinate system are further calculated. Alternatively, segmented temporal inversion, reference time backtracking, relative time difference inversion, temporal correction based on the scanning trajectory model, or other processing methods that can unify events from different acquisition times into the same reference time system can be used. Simultaneously, for spatial coordinate redistribution, mapping based on displacement lookup tables, coordinate compensation based on segmented trajectories, spatial resampling based on interpolation, coordinate correction based on local registration, or other equivalent coordinate transformation methods can be employed. In some implementations, temporal inversion and spatial coordinate redistribution can be directly applied to the original event stream; in other implementations, the event stream can first be divided into multiple event frames according to time, and then temporal correction and spatial compensation can be performed on the event frames. Essentially, as long as the temporal misalignment between events at different physical moments during continuous scanning can be eliminated and remapped to a unified target coordinate system, the purpose of this application can be achieved.

[0113] Alternatively, direct event counting accumulation can be performed on the corrected event stream. Other methods for forming an image based on the corrected event stream include weighted accumulation, polarity-separated accumulation, polarity-joint accumulation, region-selective accumulation, temporal-selective accumulation, or other methods. In some implementations, different accumulation weights can be set according to the polarity, temporal location, spatial location, or local statistical characteristics of the events to improve the reconstruction results. Essentially, the objective of this application can be achieved as long as a continuously scanned reconstructed image can be formed based on the temporally corrected event stream.

[0114] Reconstructing the original event stream or correction event stream using time windows of varying fixed lengths is an alternative approach. This can be achieved using adaptive time windows, dynamic time windows based on event density variations, time windows associated with the scan cycle, local time windows expanding around a specified time, or multi-level time windows. In some implementations, the time window can apply to the original event stream; in others, it can apply to the correction event stream after coordinate reassignment. Essentially, the objective of this application is achieved as long as reconstruction results with different temporal resolutions or equivalent exposure times can be output based on the same continuously acquired event stream.

[0115] This application can be used in combination with other imaging methods that enhance event triggering capabilities, such as introducing local brightness enhancement, structured illumination, or other active modulation mechanisms during continuous scanning, to make the event flow more conducive to subsequent reconstruction. However, it should be noted that the focus of this application is not on the event triggering method itself, but on the mechanism for temporal inversion, coordinate reassignment, and continuous cumulative reconstruction based on event time information and motion parameters under continuous scanning conditions. Therefore, regardless of whether the event triggering originates from brightness changes formed by natural continuous scanning or from brightness changes formed by additional modulation, as long as the continuous scanning reconstruction mechanism described in this application is subsequently used, it can be considered an implementation method of this application.

[0116] The time series inversion, coordinate reassignment, cumulative reconstruction, and time window posterior reconstruction are implemented by a unified processing module. As an alternative, the above processing flow can also be completed step by step by multiple independent processing units, or by a combination of some front-end hardware and some back-end software. It can be processed in real time or offline.

[0117] As long as the correlation processing between event time information and motion parameters can be completed and the continuous scanning reconstruction results can be output, the technical concept of this application will not be affected.

[0118] In summary, the alternative in this application does not lie in the change of specific device names, but in the equivalent change of the relevant implementation path in continuous scanning event imaging. The core of this application is to establish a mapping relationship between event timing and spatial displacement during continuous scanning by utilizing the temporal information in the event stream and its corresponding motion parameters. Through temporal inversion, coordinate reallocation, continuous cumulative reconstruction, and a posteriori reconstruction of the time window, continuous scanning microscopy imaging similar to time-delay integration is achieved. Any equivalent alternative that adopts the above technical concept and achieves the same or similar technical effects should be considered an alternative to this application.

[0119] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores continuous scanning microscopy imaging data based on event-stream time-series inversion. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a continuous scanning microscopy imaging method based on event-stream time-series inversion.

[0120] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0121] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0122] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0123] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0124] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0125] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0126] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logic devices, etc., and are not limited to these.

[0127] 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 are 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.

[0128] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method of continuous scanning microscopic imaging based on event stream time-of-flight inversion, characterized by, include: Acquire continuous time-varying optical signals and continuous relative motion states of the target sample; The continuous relative motion state includes the computational velocity vector acquired when the target sample is in a continuous scanning state; the computational velocity vector is used to characterize the spatiotemporal correlation with the time-varying light signal; Based on the continuous relative motion state, asynchronous detection is performed while continuously scanning to determine the original event stream; the original event stream includes a series of discrete events that start independently over time. Based on the motion parameters corresponding to the original event stream, determine the correspondence between the motion parameters and the event timestamps of discrete events; Based on the continuous relative motion state and the spatiotemporal mapping model, the discrete events in the original event stream are corrected to obtain the corrected event stream; the spatiotemporal mapping model is determined based on the correspondence between motion parameters and the event timestamps of discrete events. The time axis is divided into integral windows according to the evolution law of the time-varying light signal to obtain the integral window; The corrected event stream is integrated within the corresponding integration window to generate a continuously scanned reconstructed image.

2. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 1, characterized in that, The discrete events include: event space coordinate information. Event timestamp Event polarity information And the time-varying optical signal; in, ; This represents a positive event triggered by a rise in light intensity threshold; This represents a negative event triggered by a decrease in light intensity threshold.

3. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 1, characterized in that, The expression corresponding to the corrected event flow is: ； in, The corrected event stream; These are discrete events; This represents the corrected position of the discrete event in the target coordinate system. The timestamp of the event; This refers to the polarity information of the event. It is a spatiotemporal mapping model; This refers to event space coordinate information; It is a state of continuous relative motion.

4. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 1, characterized in that, The time axis is divided into integration windows according to the evolution law of the time-varying light signal, resulting in an integration window including: Based on the local brightness change information during continuous scanning, the periodic, phase or spatial features of the corresponding variable illumination mode are extracted, and the time dimension of the continuous event stream is divided into an integral window sequence corresponding to the preset illumination state to obtain the integral window; the local brightness change information is determined based on the preset known structured light modulation drive.

5. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 1, characterized in that, The frame-level discrete integral mathematical expression corresponding to the continuously scanned reconstructed image is: ； in, This is the frame-level discrete integral form corresponding to the continuously scanned reconstructed image; This is the starting window for the integration window; This is the closing window for the integration window; These are the spatial coordinates contained in the time-varying optical signal; It is the discrete Dirac function; It is a spatiotemporal mapping model; This refers to event space coordinate information; It is a state of continuous relative motion; The time coordinates are contained in the time-varying optical signal.

6. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 5, characterized in that, In a state of continuous motion, a mask is applied to discrete events and reconstructed to obtain the corresponding continuously scanned reconstructed image: ； ； in, A continuously scanned reconstructed image for applying a mask to discrete events and reconstructing them; For masking; This represents the polarity information of the event.

7. The continuous scanning microscopic imaging method based on event flow temporal inversion according to claim 1, characterized in that, The continuous scanning microscopy imaging method based on event stream temporal inversion further includes: Simultaneously output continuous scan images, as well as slice images or dynamic image sequences based on the integral window.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the continuous scanning microscopy imaging method based on event flow temporal inversion as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the continuous scanning microscopy imaging method based on event flow temporal inversion as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the continuous scanning microscopy imaging method based on event flow temporal inversion as described in any one of claims 1-7.

Images