Time-series in-situ zymography quantitative imaging method of soil enzyme activity based on multi-parameter screening

Abstract

The application discloses a kind of time sequence in situ zymogram soil enzyme activity quantitative imaging methods based on multi-parameter screening, it is related to soil enzyme activity detection technical field, the method collects time sequence fluorescence image of enzymatic reaction, constructs pixel-level time sequence gray matrix;Multi-stage screening is carried out to pixel, including initial screening based on gray value and standard deviation, slope and goodness of fit screening of sliding window fitting, and reaction time window final screening;Extract maximum local linear growth rate as uncorrected enzyme activity, introduce diffusion correction model correction, obtain corrected enzyme activity;Quantitative conversion and visual output are realized in combination with fluorescence standard curve.This application changes static measurement into dynamic process analysis, effectively suppresses background noise, significantly improves the sensitivity, spatial resolution and quantitative accuracy of enzyme activity detection.

Description

Technical Field

[0001] This invention relates to the field of soil enzyme activity detection technology, and in particular to a time-series in-situ enzyme spectrum quantitative imaging method for soil enzyme activity based on multi-parameter screening. Background Technology

[0002] Soil enzymes, as core catalytic substances in the material cycle and energy transformation of soil ecosystems, dominate key biochemical processes such as soil organic matter decomposition and carbon, nitrogen, and phosphorus nutrient cycling. Their activity and spatial distribution characteristics directly reflect soil microbial function, nutrient transformation rate, and soil health. Especially in soil micro-interface regions such as the rhizosphere and litter zone, enzyme activity exhibits significant spatial heterogeneity. Accurate characterization of the in-situ spatial distribution of soil enzyme activity is a key technological foundation for elucidating the nutrient cycling mechanism of soil micro-interfaces and enhancing the soil's ecological function regulation capacity.

[0003] Among existing soil enzyme activity detection technologies, two-dimensional in-situ zymography based on fluorescent substrates is the mainstream method for in-situ visualization of soil enzyme activity. This technology involves attaching a reaction membrane loaded with fluorescent substrate to the soil surface, incubating it, and then measuring the fluorescence intensity on the membrane surface. The intensity of the fluorescence signal characterizes the spatial distribution of enzyme activity, which can intuitively present the "hot spots" of soil enzyme activity. It has been widely used in research scenarios such as rhizosphere enzyme activity distribution and soil microbial processes.

[0004] Traditional in-situ enzyme profiling techniques rely on static measurements at a single time point. However, the results are easily affected by factors in the actual system, such as substrate and product diffusion, fluorescence signal saturation, and reaction nonlinearity. This results in insufficient quantitative accuracy and limited spatial resolution, making it difficult to accurately characterize the heterogeneity of enzyme activity at the soil microscale. Subsequent developments in time-lapsezymography (TLZ) have improved the reliability of quantitative detection by continuously acquiring dynamic images of enzyme-catalyzed reactions and identifying linear reaction stages. However, existing methods have relatively simplified mechanisms for processing and filtering pixel-level time-series data, making it difficult to simultaneously achieve noise suppression and effective signal preservation, thus still failing to meet the technical requirements for fine analysis of enzyme activity at the soil micro-interface.

[0005] In summary, existing in-situ imaging techniques for soil enzyme activity struggle to simultaneously achieve high sensitivity, high spatial resolution, and high quantitative accuracy, failing to accurately resolve the spatial heterogeneity of enzyme activity at micro-interfaces such as the rhizosphere. Therefore, developing a quantitative imaging method for soil enzyme activity that integrates multi-parameter screening, temporal dynamic analysis, and diffusion correction to overcome existing technological bottlenecks has become an urgent need in this field. Summary of the Invention

[0006] The purpose of this invention is to provide a time-series in-situ enzyme spectrum quantitative imaging method for soil enzyme activity based on multi-parameter screening, which aims to overcome the technical problems of low sensitivity, poor spatial resolution and large quantitative error in traditional soil enzyme spectrum methods, and realize the transformation from static measurement to dynamic process analysis.

[0007] To achieve the above objectives, this invention provides a time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening, comprising the following steps: S1. Construct an in-situ fluorescence imaging platform, attach the enzyme reaction membrane loaded with fluorescent substrate to the soil rhizosphere surface, and collect time-series fluorescence images of soil enzyme-catalyzed reactions in a dark room. S2. Perform background subtraction and noise reduction preprocessing on time-series fluorescence images; S3. Obtain the time-series image data and the pixel-level time-series grayscale data matrix G(t); S4. Calculate the maximum gray value G for each pixel in the data matrix G(t). max Initial screening is performed using the grayscale standard deviation σG, discarding those that do not meet G. max ≥G min And σG≥σ min The pixels, of which G min and σ min The values ​​range from 0.5 to 5; S5. For the pixels that pass the initial screening, apply a sliding window local linear fitting and calculate the gray-level change rate dG / dt and the goodness of fit R. 2 Only those satisfying the dG / dt slope threshold ≥ 0.05~1 and R 2 Fitting results with a value ≥0.6~0.95; S6. Extract the maximum local linear growth rate of each pixel as the uncorrected enzyme activity of that pixel, and introduce a diffusion correction model to correct it, so as to obtain the corrected enzyme activity. S7. The window center time t corresponding to the corrected enzyme activity center Perform final screening and retain only t center Pixels located in the 5-55 minute range are used to reconstruct the spatial distribution matrix of enzyme activity; S8. Prepare a fluorescence standard curve to obtain the grayscale-enzyme activity conversion coefficient, convert the corrected enzyme activity into quantitative enzyme activity data and output it visually.

[0008] Preferably, in step S1, the acquisition time of the time-series fluorescence image is 60 to 90 minutes, and the sampling time interval is 1 minute.

[0009] Preferably, in step S2, the background subtraction includes subtracting the background value using the first frame image; the noise reduction includes preliminary noise reduction using median filtering.

[0010] Preferably, in step S4, the G min and σ min The preferred value is 1.

[0011] Preferably, in step S5, the half-width of the sliding window is 2 to 10 minutes, and the number of window points is 5 to 21.

[0012] Preferably, the half-width of the sliding window for local linear fitting in step S5 is 4 minutes, and the number of window points is 9.

[0013] Preferably, in step S6, the diffusion correction model is an empirical function model with a correction factor f. corr The calculation formula is:

[0014] in, p a , p b and p c The values ​​are 2.95, 26 and 0.84 respectively.

[0015] Preferably, in step S8, the fluorescence standard curve is prepared by immersing the enzyme reaction membrane in fluorescent standard solutions of different concentrations, performing fluorescence imaging under the same imaging conditions, establishing a linear relationship between grayscale value and fluorescence product concentration, and obtaining the conversion coefficient b; the conversion coefficient b includes: for acid phosphatase, b1 = 6.13 pmol MUF mm -2 Greyscale 1 For alkaline phosphatase, b2 = 5.33 pmol MUF mm -2 Greyscale -1 .

[0016] A system for implementing the above method includes: An imaging module is used to perform step S1 and acquire the time-series fluorescence image; The image preprocessing module is used to perform steps S2 and S3 to perform background subtraction and noise reduction on the image and obtain a temporal grayscale data matrix. The multi-parameter screening and fitting module is used to execute steps S4, S5 and S6 to perform initial pixel screening, sliding window fitting and diffusion correction. The post-processing and visualization module is used to execute steps S7 and S8, performing final data screening, spatial reconstruction, and visualization output.

[0017] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: 1. Significantly improves detection sensitivity and signal-to-noise ratio This invention effectively suppresses background fluorescence accumulation and random noise interference through a multi-level parameter screening strategy. Simultaneously, by extracting the rate of change of fluorescence signal rather than its cumulative intensity, it transforms the "cumulative signal" in traditional methods into a "transient reaction rate signal," allowing the true enzyme reaction signal to be fully released. This significantly enhances the ability to identify signals in low-activity regions and greatly improves detection sensitivity.

[0018] 2. Significantly improves spatial resolution. This invention effectively corrects the signal ambiguity caused by substrate and product diffusion through the synergistic effect of sliding window local linear fitting and diffusion correction model, resulting in a more concentrated enzyme activity distribution and clearer hotspot region boundaries. Compared with the discrete patchy distribution of enzyme activity signals in traditional methods, the spatial distribution of enzyme activity obtained by this invention presents a continuous strip-like or network-like structure, realizing the transformation from "ambiguous diffusion distribution" to "structurally clear distribution".

[0019] 3. Significantly improves quantitative accuracy This invention extracts the maximum local linear growth rate of each pixel as a parameter for characterizing enzyme activity, avoiding the underestimation of high-activity regions caused by fluorescence signal saturation or nonlinear reaction in traditional methods. Simultaneously, a diffusion correction model quantitatively compensates for diffusion effects, further reducing quantitative bias. This allows the invention to more realistically reflect the spatial heterogeneity of enzyme activity in the soil microenvironment, significantly improving the accuracy and reliability of quantification.

[0020] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the time-series in situ enzyme imaging device of the present invention; Figure 2 This is a time-series in situ enzyme spectrum standard curve of the present invention; Figure 3 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil A using traditional enzyme profiling and the time-series enzyme profiling method of this invention; Figure 4 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil B using traditional enzyme profiling and the time-series enzyme profiling method of this invention. Figure 5 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil C using traditional enzyme profiling and the time-series enzyme profiling method of this invention; Figure 6 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil D using traditional enzyme profiling and the time-series enzyme profiling method of this invention; Figure 7 The image shows the frequency distribution characteristics of pixel intensity of phosphatase activity determined by traditional enzyme chromatography and the time-series enzyme chromatography method of this invention. Figure 8 This is a curve showing the dynamic change of single-pixel grayscale value over time using the time-series enzyme spectroscopy method of this invention. Figure 9 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil A using the method of Guber et al. (2021) and the temporal enzyme spectrometry method of this invention; Figure 10 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil B using the method of Guber et al. (2021) and the temporal enzyme spectrometry method of this invention; Figure 11 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil C obtained by the method of Guber et al. (2021) and the temporal enzyme spectrometry method of this invention; Figure 12 A comparison diagram of the spatial distribution of rhizosphere phosphatase activity in soil D using the method of Guber et al. (2021) and the temporal enzyme spectrometry method of this invention; Figure 13 For different R in soil A 2 Comparison of time-series enzyme profiling results under the screening threshold; Figure 14 For different R in soil B 2 Comparison of time-series enzyme profiling results under the screening threshold; Figure 15 For different R in soil C 2 Comparison of time-series enzyme profiling results under the screening threshold; Figure 16 For different R in soil D 2 Comparison of time-series enzyme profiling results under the screening threshold; Figure 17 Comparison of temporal enzyme spectrum imaging results at different sliding window half-widths in soil A; Figure 18 Comparison of temporal enzyme spectrum imaging results at different sliding window half-widths in soil B; Figure 19 Comparison of temporal enzyme spectrum imaging results at different sliding window half-widths in soil C; Figure 20 This is a comparison of time-series enzyme profile imaging results at different sliding window half-widths in soil D. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] Example This embodiment provides a soil enzyme activity quantitative imaging method based on time-lapse zymography (TLZ) with multi-parameter screening. By acquiring time-series fluorescence images of soil enzyme-catalyzed reactions, pixel-level time-series data is constructed, and reaction rate characteristic parameters are extracted based on local linear fitting. At the same time, combined with a multi-parameter screening mechanism and diffusion correction model, the pixel-level data is constrained and optimized, thereby achieving high-resolution quantitative imaging of soil enzyme activity.

[0026] The method mainly includes the following steps: S1. Construct an in-situ fluorescence imaging platform, attach the enzyme reaction membrane loaded with fluorescent substrate to the soil rhizosphere surface, and collect time-series fluorescence images of soil enzyme-catalyzed reactions in a dark room. Experimental setup: planar ultraviolet excitation plate, Canon camera (EOS 200D) with filter (460 nm), camera bracket, computer connected to Canon EOS Utility remote control software, polyamide film, and plexiglass root box (dimensions 15cm×10cm×2cm, which are height, width and thickness, respectively).

[0027] (1) When constructing the in-situ fluorescence imaging platform, first fix the Canon EOS 200D camera on the camera bracket, ensuring that the lens is vertically aligned with the sample area, and adjust the height so that the shooting range completely covers the root box observation area. At the same time, install a 460 nm bandpass filter at the front of the lens to filter the fluorescence signal. Connect the camera to the computer using the Canon EOS Utility remote control software to observe the results. Then, place two planar ultraviolet excitation plates (main peak about 350 nm) on both sides of the root box, and adjust the distance between the light source and the root box (generally 10-20 cm) to make the illumination uniform and without obvious shadows or hot spots. Place the plexiglass root box containing soil and rhizosphere stably in the center of the imaging platform and keep the surface level. Tightly attach the prepared enzyme reaction membrane (polyamide membrane) to the surface of the soil rhizosphere, avoiding air bubbles and wrinkles, to ensure uniform signal transmission. Finally, turn off the ambient light source or build a light shield (dark box) around the device to reduce external light interference and improve the fluorescence signal-to-noise ratio. See the schematic diagram below. Figure 1 .

[0028] (2) Time-series in situ fluorescence zymography imaging: First, an enzyme reaction membrane containing fluorescently labeled substrate was prepared. The polyamide membrane (4 cm × 4 cm) was fully soaked with a 10 mM 4-methylumbelliferyl phosphate (MUF-P) substrate solution prepared with 10 mM MMES (morpholinoethanesulfonic acid) buffer. The pH of the buffer was selected according to the target enzyme type: pH=6.5 for acid phosphatase and pH=11.0 for alkaline phosphatase. Then, it was flatly attached to the surface of the rhizosphere soil for in situ reaction. Under ultraviolet light excitation (excitation wavelength of 350 nm), the membrane was imaged in a dark room using a camera (EOS 200D, Canon) and a bandpass filter (460 nm). The camera parameters were (shutter speed: 1 / 160, aperture: F5.6, ISO: 1600), the acquisition time was 60-90 min, and the sampling time interval was 1 min, thus obtaining a set of time-series fluorescence images.

[0029] S2. Perform background subtraction and noise reduction preprocessing on time-series fluorescence images; Background subtraction preprocessing: Import the captured footage into After Effects (AE, 2024) or other image processing software. Use the pen tool or mask to crop the desired portion. Extract the first frame image and convert it to a subtraction overlay mode, i.e., subtract the background value of the first frame. Then add the "Median" effect with a radius of 3 for initial noise reduction. Finally, export it as a series of images in PNG or TIFF format. Use ImageJ software to integrate the previously obtained series of images into a single 8-bit continuous TIFF image and export it. This yields the initial noise-reduced time-series imaging result after subtracting the background value of the first frame, denoted as 1.tif.

[0030] S3. Acquisition of temporal image data and pixel-level temporal grayscale data matrix G(t); The 1.tif file was analyzed using RStudio software and R language code to obtain soil in-situ enzyme spectrum time series images (8-bit grayscale images). Each image corresponds to a time point with a time interval of 1 min, and a data matrix G(t) of pixel grayscale values ​​changing over time was constructed.

[0031] S4. Calculate the maximum gray value G for each pixel in the data matrix G(t). max Initial screening is performed using the grayscale standard deviation σG, specifically including: Extract the time-series grayscale value of each pixel and calculate: the maximum grayscale value G. max =max(G(t)), grayscale standard deviation σG=std(G(t)), performs preliminary filtering of pixels, the filtering condition is: G max ≥G min (0.5~5, preferably 1), σG≥σ min (0.5 to 5, preferably 1), grayscale values ​​that do not meet the requirement are assigned to 0 to remove low signal or noise pixels, and only pixels with effective dynamic change characteristics are retained for subsequent analysis.

[0032] S5. For the pixels that pass the initial screening, apply a sliding window local linear fitting and calculate the gray-level change rate dG / dt and the goodness of fit R. 2 Only those satisfying the dG / dt slope threshold ≥ 0.05~1 and R 2 Fitting results with a value of ≥0.6 to 0.95 include, specifically: (1) Local linear fitting with sliding window: For pixels that pass the initial screening, a local linear fitting analysis is performed on their time series using a sliding window. The half-width of the sliding window is ±4 min, and the number of window points is 9. Linear regression calculation is performed with time as the independent variable and gray value as the dependent variable. Fitting calculations are performed sequentially for time points that meet the conditions. For example, the fitted value at the 10th minute is fitted with the gray values ​​at the 6th, 7th, 8th, 9th, 10th, 11th, 12th, 13th, and 14th minutes to obtain the gray value change rate (dG / dt) and the goodness of fit R within each window. 2 Only retain those satisfying a slope of not less than 0.1 and R 2 A fitting result of at least 0.8 is used, and the rest are assigned a value of 0, thereby ensuring that the extracted signal has high reliability and linear characteristics.

[0033] (2) Maximum reaction rate extraction: In all sliding windows that meet the screening criteria, the maximum local linear growth rate of each pixel is extracted as the uncorrected enzyme activity a_raw of that pixel, and the corresponding window center time t is recorded.center , where t center This indicates the time position at which the pixel reaches its maximum reaction rate, thus enabling an accurate characterization of the reaction kinetics.

[0034] (3) Diffusion Correction: To eliminate the influence of substrate or product diffusion on the measurement results, a diffusion correction model is introduced to correct the uncorrected enzyme activity. The diffusion correction factor f for MUF as the product is... corr The calculation formula is as follows:

[0035] Where parameters p a , p b and p c The values ​​of are 2.95, 26 and 0.84, respectively. This correction factor is used to quantitatively compensate for the diffusion effect of different time responses (the formula and parameters are from existing literature: Guber A, Blagodatskaya E, Juyal A, et al. Time-lapse approach to correct deficiencies of 2D soilzymography[J]. Soil Biology and Biochemistry, 2021, 157: 108225).

[0036] S6. The uncorrected enzyme activity a obtained above raw With diffusion correction factor f corr Pixel-by-pixel multiplication is performed to obtain the corrected enzyme activity a. corr This enables quantitative characterization of the true rate of enzyme reactions, improving the accuracy and comparability of results.

[0037] S7. The window center time t corresponding to the corrected enzyme activity center Final screening is performed, and the spatial distribution matrix of enzyme activity is reconstructed, specifically including: (1) Final screening within a time window: To further ensure the reliability of the results, the calculated t is subjected to a final screening. center Filter by time range, keeping only t center Pixels located between 5 and 55 min exhibit a stable linear increase in grayscale value. Figure 8 For pixels that do not meet this condition, their corresponding enzyme activity results are set to 0, thereby removing interference from abnormal reactions or unstable phase data.

[0038] (2) Result output and spatial reconstruction: The uncorrected enzyme activity a of each pixel obtained by the final screening is then used to output the results. raw Maximum reaction time tcenter diffusion correction factor f corr and enzyme activity a after final screening correction within the time window corr The data were reconstructed into two-dimensional spatial distribution matrices and output as data files, achieving high-resolution quantitative imaging and spatial visualization of soil enzyme activity at the microscale. The results were output as 2.csv, 3.csv, 4.csv, and 5.csv files, respectively.

[0039] S8. Construct a fluorescence standard curve to obtain the grayscale-enzyme activity conversion coefficient, convert the corrected enzyme activity into quantitative enzyme activity data, and visualize the output. Specifically, this includes: (1) Prepare a fluorescence standard curve: First, calculate the final enzyme activity according to the formula:

[0040] Where q(MUF) is the amount of MUF released per unit area, in pmol MUF / mm². -2 G represents the image grayscale value, and b is the conversion coefficient between the grayscale value and the amount of MUF released per unit area, in pmol MUF / mm². -2 Greyscale -1 A represents the final enzyme activity, characterized by the MUF release rate per unit area, in pmol MUF / mm². -2 min -1 .

[0041] Then, the polyamide membrane was immersed in MUF standard solutions of different concentrations (0.25–10 mM), and a 10 mM MUF-P substrate solution was prepared using 10 mM MME S buffer to completely wet the membrane (approximately 2 mL / 100 cm²). 2 The pH of the buffer system was selected according to the imaging target: pH=6.5 for acid phosphatase and pH=11.0 for alkaline phosphatase. Fluorescence imaging was performed on the standard membrane under the same imaging conditions, and the image after 1 h of reaction was selected as the fully developed state, from which the corresponding gray values ​​were extracted. By establishing a linear relationship between gray values ​​and MUF concentration, a standard curve was constructed to achieve the quantitative conversion of gray value signals into enzyme activity.

[0042] The final conversion coefficients were: acid phosphatase b1 = 6.13, alkaline phosphatase b2 = 5.33, in pmol MUF mm. - 2 Greyscale -1 ,like Figure 2 As shown, (a) represents acid phosphatase and (b) represents alkaline phosphatase, which are used to convert the gray values ​​of in situ enzyme spectrum images into the amount of MUF released per unit area, thereby realizing the quantitative analysis of enzyme activity.

[0043] (2) Grayscale Conversion and Visualization: The conversion coefficient b between grayscale and enzyme activity obtained through the datum lines was calculated using RStudio software and R language code on the 5.csv file obtained above. The grayscale change rate data was multiplied by the conversion coefficient b to obtain the final 32-bit precision enzyme activity data file, where acid phosphatase is 6.csv and alkaline phosphatase is 7.csv. The calculated enzyme activity data file was then imported into ImageJ software for visualization. The specific steps were as follows: Image, Lookup Tables, and 16 Colors were selected in sequence to apply pseudo-color mapping to the image, generating a spatial distribution image of enzyme activity. Then, Analyze, Tools, and Calibration Bar were selected to add a color gradient and a scale corresponding to the enzyme activity value, which was used as the final result display.

[0044] Comparative Example I. Comparison between traditional single-time-point in situ enzyme profiling and the method of this invention To verify the technical effectiveness of the method of this invention, a comparative analysis was conducted between the traditional single-time-point in-situ zymography method and the time-lapse zymography (TLZ) method based on multi-parameter screening proposed in this invention. Under the same soil sample and the same spatial scale conditions, enzyme activity distribution images were obtained using both the traditional zymography method based on single-time-point fluorescence intensity and the time-lapse zymography method of this invention, which is based on time-series information and combines a multi-parameter screening mechanism. A comprehensive evaluation was then performed based on spatial imaging characteristics, quantitative results, and intensity distribution characteristics.

[0045] 1. Comparison of spatial imaging features according to Figure 3 , Figure 4 , Figure 5 , Figure 6 In the image, (a) is the original image of the soil-root system; (b1) and (c1) are the spatial distributions of acidic and alkaline phosphatase activities measured by traditional enzyme chromatography, respectively; and (b2) and (c2) are the spatial distributions of acidic phosphatase (ACP) and alkaline phosphatase (AKP) activities measured by the time-series enzyme chromatography method of this invention, respectively. The units are pmol MUF mm. -2 min -1In terms of spatial imaging, traditional enzyme spectrum images often show enzyme activity distributed over large areas with strong background signals, blurred root outlines, and high-value areas typically appearing as diffuse patches, resulting in poor spatial continuity and difficulty in accurately reflecting the true distribution characteristics of rhizosphere enzyme activity. This is mainly because traditional methods analyze fluorescence intensity based on a single time point, which is easily affected by substrate diffusion, fluorescence accumulation, and signal saturation, thus introducing strong background interference and reducing spatial resolution. In contrast, the time-series enzyme spectrum method of this invention extracts the rate of change (dG / dt) of fluorescence grayscale signal through multi-time-point image acquisition and pixel-level time-series analysis, effectively suppressing background noise and random fluctuations, so that the enzyme activity signal is mainly concentrated around the root system and presents a clear strip-like or network-like hotspot distribution. This method significantly enhances the spatial consistency between root structure and enzyme activity distribution, realizing the transformation from "fuzzy diffuse distribution" to "structurally clear distribution," and significantly improving microscale spatial resolution.

[0046] 2. Comparison of quantitative results Regarding quantitative results, according to Table 1, the overall activities of acidic and alkaline phosphatases measured by traditional enzyme spectrometry were low (approximately 0.3–0.6 pmol MUF mm). -2 min -1 Furthermore, the differences between different soil types were not significant, indicating a certain degree of signal compression; while the enzyme activity measured by time-series enzyme spectroscopy was significantly increased (approximately 3–20 pmol MUF mm). -2 min -1 The overall improvement ranged from 8.9 to 36.1 times, with clearer differences between different soil types and enzyme types. Although the standard deviation of the time-series method is relatively large, this reflects its ability to more realistically reveal the spatial heterogeneity of enzyme activity in the soil microenvironment, rather than the "spurious stability" caused by background interference and signal averaging in traditional methods. Furthermore, the absolute difference in the average activities of the two phosphatases was more pronounced in the time-series enzyme profile results, further demonstrating the significant advantages of this method in enhancing differential biological expression and improving detection sensitivity.

[0047] Table 1. Comparison of phosphatase activities in different soils determined by traditional enzyme chromatography and time-series enzyme chromatography.

[0048] 3. Comparison of pixel intensity distribution characteristics From the pixel distribution characteristics of enzyme activity intensity, according to Figure 7The intensity distribution characteristics of acid and alkaline phosphatase activities determined by traditional and time-series enzyme chromatography methods are shown in (a) and (b), where (c) and (d) represent the pixel intensity distribution of acid and alkaline phosphatase activities determined by traditional enzyme chromatography, and (c) and (d) represent the pixel intensity distribution of acid and alkaline phosphatase activities determined by the time-series enzyme chromatography method of this invention. Different symbols represent different soil types (soil A, soil B, soil C, and soil D). In traditional enzyme chromatography images, pixel intensity is mainly concentrated in the low-value range, showing a single-peak or slightly skewed distribution overall, with a low proportion of high-intensity pixels, indicating that its signal dynamic range is compressed and it is difficult to effectively distinguish highly active micro-regions. In contrast, the pixel intensity distribution of time-series enzyme chromatography is significantly broadened, extending into the high-value region, exhibiting long-tail or even multi-peak characteristics, with a significant increase in highly active pixels. This indicates that this method can more fully release and amplify the real enzyme reaction signal, thereby significantly improving signal contrast and expanding the detection dynamic range.

[0049] Further analysis shows that the method of this invention significantly enhances the recognition ability of effective signals at the same unit scale by fitting the rate of change of fluorescence signals through a sliding window. Simultaneously, by introducing a diffusion correction model, the signal diffusion caused by substrate and product diffusion during the enzyme reaction is corrected, resulting in a more concentrated enzyme activity distribution and clearer boundaries of hotspot regions. This further improves spatial resolution and quantitative accuracy, leading to a significant increase in the final enzyme activity after correction (8.9–36.1 times). In contrast, traditional methods do not consider diffusion effects, resulting in a widespread "fuzzy diffusion" phenomenon in highly active regions, which easily leads to misjudgments of hotspot locations and ranges.

[0050] II. Comparison with the time-series zymography method of Guber et al. (2021) To compare with existing similar time-series in situ enzyme profiling methods, a control method based on linear phase identification was constructed, referencing the enzyme activity estimation method published by Guber et al. (2021). The method was as follows: 5–55 min was uniformly selected as the linear phase, and overall linear fitting was performed on pixel-level time-series data to obtain enzyme activity parameters; simultaneously, only G… max / σG is used as the initial screening criterion to remove low-signal pixels, without setting the slope or goodness of fit (R). 2 The selection criteria are used to form a relatively simplified control scheme.

[0051] 1. Qualitative comparison of imaging effects Its imaging results are as follows Figures 9-12As shown, (a) is the original rhizosphere image, (b1) and (c1) are the acid and alkaline phosphatase activity distributions measured by the method of Guber et al. (2021), respectively, and (b2) and (c2) are the acid and alkaline phosphatase activity distributions measured by the time-series enzyme spectrometry method of this invention, respectively. Under different soil conditions, the two methods showed significant differences in the spatial distribution of rhizosphere phosphatase activity. Overall, the results obtained by referring to the control method of Guber et al. (2021) showed a relatively continuous distribution characteristic, but the background signal was high, the overall contrast was low, the boundary between hotspot areas and the background was unclear, and there was obvious signal diffusion in some areas, making the hotspot structure blurred and difficult to accurately reflect the spatial heterogeneity of rhizosphere microregions. In contrast, the time-series enzyme spectrometry method proposed in this study extracts reaction rates using time-series information and combines a multi-level screening strategy (including slope threshold and R...) 2 The method effectively suppressed background noise, significantly reducing low-intensity background and concentrating enzyme activity signals in the rhizosphere region, forming a clear and continuous hotspot structure. Its spatial distribution better conforms to root morphology, resulting in more accurate localization. This difference remained consistent across different soil types and acidic and alkaline phosphatase systems: the control method generally exhibited a "high background + weak contrast" characteristic, while this method showed a "low background + high contrast + concentrated hotspots" characteristic, indicating its good stability and applicability.

[0052] 2. Comparison of quantitative indicators To further quantify the differences between the two methods, three indicators were introduced for evaluation: non-zero pixel proportion, hotspot area proportion, and enrichment factor (EF). The non-zero pixel proportion is defined as the percentage of pixels with values ​​greater than zero in the final image, representing the degree of effective signal preservation. The hotspot area proportion is defined as the percentage of pixels with values ​​higher than the overall mean plus two standard deviations, reflecting the spatial distribution of highly active regions. The enrichment factor is defined as the ratio of the average enzyme activity of pixels in all hotspot regions to the average enzyme activity of pixels in non-hotspot regions, measuring the signal enhancement of hotspot regions relative to the background region. The results (Table 2) show that the two methods exhibit significant differences in all three indicators, following a consistent pattern.

[0053] Table 2. Comparison of spatial characteristic parameters of phosphatase activity in different soils using the linear stage identification method (Guber et al., 2021) and the time-series enzyme profiling method.

[0054] First, regarding the proportion of non-zero pixels, the control method used by Guber et al. (2021) generally showed higher proportions, for example, acid phosphatase and alkaline phosphatase reached 83.2% and 94.1% in soil A, and 87.6% and 97.8% in soil B, respectively. Our method significantly reduced these proportions, decreasing them to 18.0% and 37.3% in soil A, and 18.2% and 10.6% in soil B, respectively. This result indicates that our method can effectively remove low-signal and noisy pixels, making the effective information more concentrated, thereby improving data quality.

[0055] Secondly, in terms of the proportion of hotspot areas, this method is generally higher than or close to the control method. For example, in soil B, acid phosphatase increased from 4.02% to 8.20%, and alkaline phosphatase increased from 4.40% to 7.89%; in soil C, acid phosphatase increased from 5.49% to 8.62%. This indicates that while reducing the background, it did not compress the hotspot areas, but instead enhanced the ability to identify highly active areas.

[0056] Further analysis of the hotspot enrichment factor (EF) shows that this method significantly outperforms the reference method. For example, acid phosphatase in soil A increased from 5.18 to 16.38, in soil B from 3.25 to 14.61, and in soil C from 4.67 to 34.73. The increase in alkaline phosphatase is even more significant, with soil B increasing from 2.81 to 62.96 and soil D from 3.53 to 541.90. This indicates that this method can significantly enhance the contrast between hotspot signals and the background, thereby improving spatial resolution. It should be noted that in some cases (such as when the acid phosphatase EF in soil D reached 1482.36), the EF value was extremely high, mainly due to the significant suppression of the background signal, further reflecting the strong background suppression ability of this method. However, excessively high background suppression needs to be considered on a case-by-case basis.

[0057] Based on the combined image performance and quantitative indicators, it can be seen that the control method of Guber et al. (2021) exhibits the characteristics of "high non-zero pixel ratio + low hot spot contrast". Although it retains a large number of pixels, the noise interference is obvious. In contrast, our method exhibits "low non-zero pixel ratio + higher hot spot area ratio + significantly improved EF". While effectively suppressing background noise, it can retain and enhance the real enzyme activity signal, achieving a better balance between signal extraction and spatial structure recognition. This significantly improves the clarity, contrast and spatial resolution of the enzyme activity imaging results.

[0058] III. Sensitivity and Optimization Verification of Key Parameters To verify the rationality and advantages of the selection of screening parameters in this method, we examined the two most important parameters R, which have universality. 2 The parameters of the sliding window half-width were checked.

[0059] 1. Goodness of fit R2 Threshold optimization verification To verify the rationality and superiority of the parameter selection method proposed in this paper, the key parameter R was analyzed. 2 The screening thresholds were systematically analyzed and combined with imaging results ( Figures 13-16 A comprehensive evaluation was conducted using quantitative indicators (Tables 3 and 4). Figures 13-16 In the diagram, (a1) to (a4) represent R. 2 The acid phosphatase activity distribution corresponding to thresholds of 0.6, 0.7, 0.8, and 0.9, (b1) to (b4) are R 2 Alkaline phosphatase activity distribution corresponding to thresholds of 0.6, 0.7, 0.8, and 0.9. The results indicate that different R values... 2 Thresholds significantly affected the temporal zymography of both acidic and alkaline phosphatases, exhibiting a consistent pattern across different soil types. With the increase in R... 2 As the threshold is gradually increased from 0.6 to 0.9, the overall image quality gradually shifts from "high noise, low contrast" to "low noise, high contrast." However, when the threshold is too high, significant signal loss occurs. When R... 2 When R = 0.6, the image contains a large amount of scattered signal, has high background noise, exhibits random and discrete enzyme activity distribution, and has low contrast between hotspot areas and the background; when R 2 When R is increased to 0.7, the background noise decreases somewhat, and the signal begins to concentrate in local areas, but there are still many discrete pixels, and the spatial structure is not clear enough; when R 2 When R = 0.8, the image quality is significantly improved, background noise is significantly reduced, and the enzyme activity signal is mainly concentrated in the rhizosphere region. The hotspot structure is clear and has good spatial continuity, which can more realistically reflect the spatial distribution characteristics of enzyme activity; while when R = 0.8, the image quality is significantly improved, the background noise is significantly reduced, the enzyme activity signal is mainly concentrated in the rhizosphere region, the hotspot structure is clear and has good spatial continuity, and it can more realistically reflect the spatial distribution characteristics of 2 When the concentration was further increased to 0.9, although the background was further suppressed, the effective signal was significantly reduced, and some hotspot regions showed breakage or even disappearance, exhibiting obvious over-selection characteristics. The above trends were consistent in both acidic and alkaline phosphatases.

[0060] Table 3. Comparison of spatial characteristic parameters of acid phosphatase activity obtained by time-series enzyme profiling under different R² screening thresholds. (Where Inf indicates that the background average value approaches zero in the hotspot enrichment factor calculation, causing the result to approach infinity.)

[0061] Table 4. Comparison of spatial characteristic parameters of alkaline phosphatase activity obtained by time-series enzyme profiling under different R² screening thresholds.

[0062] From a quantitative perspective, R 2The threshold systematically affects the proportion of non-zero pixels, the proportion of hotspot area, and the hotspot enrichment factor (EF). Firstly, the proportion of non-zero pixels increases with R... 2 Significant decreases were observed, for example, in acid phosphatase, soil A decreased from 61.5% to 3.0%, soil B from 65.0% to 2.0%, soil C from 51.2% to 1.9%, and soil D from 31.3% to 0.5%; in alkaline phosphatase, soil A decreased from 90.4% to 7.3%, and soil C from 79.4% to 6.8%, indicating that increasing R... 2 Thresholding can effectively filter out low-quality signals, but it also leads to a significant reduction in effective information. Secondly, the proportion of hotspot area increases with R. 2 It shows a trend of "increasing first and then decreasing" in R. 2 When R = 0.6 and 0.7, the hotspot area is smaller, while R... 2 When R = 0.8, the enzyme reaches its maximum or near maximum value. For example, acid phosphatase in soils A, B, and C reaches 8.89%, 8.20%, and 8.62%, respectively, while alkaline phosphatase in soils B, C, and D reaches 7.89%, 5.93%, and 5.49%, respectively. This indicates that under this condition, the true hotspot areas can be sufficiently identified. However, when R... 2 When the threshold is 0.9, the hotspot area decreases significantly. For example, in acid phosphatase, soil A decreases to 2.94%, soil B to 2.04%, and soil C to 1.88%, indicating that excessively high thresholds lead to over-screening of hotspot areas. Further analysis of the hotspot enrichment factor (EF) reveals that it... 2 Within the range of 0.6 to 0.8, the concentrations showed a steady upward trend. For example, in acid phosphatase, soil A increased from 3.12 to 16.38, soil B from 3.11 to 14.61, and soil C from 3.81 to 34.73; in alkaline phosphatase, soil B increased from 3.58 to 62.96, and soil D from 5.16 to 541.90, indicating that the hotspot signal gradually strengthened relative to the background. However, in R... 2 When = 0.9, EF shows an abnormal increase or even an infinite value (Inf). For example, soil A in acid phosphatase reaches 24668.57 and soil C reaches 57031.64, while soil B and soil D in alkaline phosphatase show Inf. The reason is that the background pixels are excessively removed and approach zero, causing the ratio to lose its actual physical meaning.

[0063] Combining image features and quantitative indicators, it can be seen that R 2 =0.6 and R 2 While a value of 0.7 retains a relatively large number of pixels, it suffers from high background noise and limited hotspot recognition capabilities; R 2 While a ratio of 0.9 can significantly suppress background, it results in a significant loss of effective signal and causes abnormal amplification of the EF signal; in contrast, R... 2=0.8 exhibits optimal characteristics in different soils and acidic and alkaline phosphatase systems. It not only maintains a moderate proportion of non-zero pixels (approximately 13%–18%), but also achieves the largest hotspot area proportion (approximately 5%–9%). Meanwhile, EF is in a range that is significantly improved but still practically meaningful (approximately 14–60), indicating that this parameter condition achieves the best balance between noise suppression and effective information preservation.

[0064] 2. Verification of sliding window half-width parameter optimization To further verify the rationality and superiority of the sliding window half-width parameter selection in this method, a systematic test and comparative analysis were conducted on the imaging results under different window scale conditions. Figures 17-20 As shown in Tables 5 and 6, the half-width of the sliding window has a significant impact on the temporal enzyme spectrum imaging results, and this effect is consistent across different soil types and acidic and alkaline phosphatase systems. Figures 17-20 In the table, (a1) to (a4) represent the acid phosphatase activity distribution corresponding to the sliding window half-width ±2, ±3, ±4, and ±6 min, and (b1) to (b4) represent the alkaline phosphatase activity distribution corresponding to the sliding window half-width ±2, ±3, ±4, and ±6 min.

[0065] Table 5. Comparison of spatial characteristic parameters of acid phosphatase activity obtained by time-series enzyme profiling under different sliding window half-widths.

[0066] Table 6. Comparison of spatial characteristic parameters of alkaline phosphatase activity obtained by time-series enzyme profiling under different sliding window half-widths.

[0067] Overall, as the window half-width gradually increases from ±2 to ±6, the imaging results show a clear trade-off between noise suppression and effective information preservation.

[0068] First, regarding the proportion of non-zero pixels, it shows a continuous decreasing trend as the window size increases. For example, in acid phosphatase, soil A decreased from 61.7% to 9.4%, soil C from 54.3% to 8.3%, and soil D from 40.7% to 1.8%; in alkaline phosphatase, soil B decreased from 63.8% to 3.0%. This indicates that increasing the window size helps smooth the time series and suppress random noise, but at the same time weakens the effective signal, leading to a gradual loss of information.

[0069] Secondly, regarding the proportion of hotspot area, the changes show a trend of "increasing first, then stabilizing or decreasing." Under the ±2 window condition, the hotspot area is relatively small; for example, in acid phosphatase, it is only 1.95% in soil A and 2.04% in soil B. As the window increases to ±4, the hotspot area significantly increases, reaching 8.89%, 8.20%, and 8.62% in soils A, B, and C, respectively, which are the maximum or close to the maximum, indicating that this scale can adequately identify truly high-activity areas. However, when the window is further increased to ±6, the hotspot area in some soils decreases or fluctuates; for example, soil C decreases from 8.62% to 7.21%, and soil D decreases from 5.13% to 1.81%, indicating that an excessively large window weakens the local spatial structure, leading to the smoothing or even loss of hotspot information.

[0070] Further analysis of the hotspot enrichment factor (EF) reveals a significant increase with increasing window size. In acid phosphatase, the EF for soil A increased from 3.25 (±2) to 16.38 (±4), further increasing to 157.85 at ±6; for soil C, it increased from 3.78 to 34.73 (±4), reaching 167.27 at ±6. This trend is even more pronounced in alkaline phosphatase, as seen in soil B where the EF was 62.96 at ±4, but reached a high of 13551.75 at ±6. However, under ±6 conditions, EF exhibits an abnormal increase, even reaching infinity (e.g., Inf for soil D). The fundamental reason for this is that the background signal is excessively smoothed, approaching zero, causing the ratio to lose its actual physical meaning, reflecting a significant over-smoothing problem.

[0071] Further analysis of image data can verify the above conclusions: Under ±2 and ±3 conditions, numerous discrete noise points remain in the images, resulting in high background and unclear hotspot structures. Under ±4 conditions, background noise is significantly reduced, enzyme activity signals are mainly concentrated in the rhizosphere region, and hotspot structures are continuous and have clear boundaries, accurately reflecting spatial distribution characteristics. However, under ±6 conditions, although the background is further reduced, detailed structures are excessively smoothed, and hotspot areas shrink or even break, consistent with the abnormally increased EF. Combining the non-zero pixel ratio, hotspot area ratio, and EF indicators, it can be seen that ±2 to ±3 conditions result in high noise and limited structure recognition ability, while ±6 conditions suffer from signal loss and result distortion. In contrast, ±4 exhibits optimal characteristics in different soils and for both types of phosphatases, effectively reducing background noise while maintaining a high hotspot area (approximately 5%–9%) and a reasonable EF level, achieving the best balance between noise suppression and spatial information preservation. Therefore, ±4 can be considered the preferred sliding window half-width parameter for this method.

[0072] Finally, it should be noted that parameters G (gray intensity), σG (gray standard deviation), and the fitting slope are all indicators sensitive to the intensity of the imaging signal. Their values ​​are greatly affected by imaging conditions (such as camera exposure parameters, sensor response characteristics, and ultraviolet light source intensity). Therefore, their universality is relatively lower than that of the sliding window and R... 2 These parameters are relatively weak and may require appropriate adjustments under different experimental conditions. Furthermore, the final screening time window (e.g., 5–55 min) is primarily used to constrain the time position of the maximum slope, preventing some pixels from exhibiting abnormally large fitting slopes in the early or late stages of the reaction (i.e., before entering or after deviating from the linear growth phase). Such anomalies usually originate from noise fluctuations or background signal changes, rather than the actual enzymatic reaction process. Therefore, this step is essentially a preventative or constraining screening measure to ensure that the extraction reaction rate has a clear kinetic significance, while its contribution to pixel selection is relatively weaker than the aforementioned parameters (e.g., G). max / σG Initial screening and slope / R 2 (Screening). It should be further noted that the optimal values ​​of each parameter are also influenced by soil type and its physicochemical properties (such as pH, organic matter content, and microbial activity). Enzyme-catalyzed reaction kinetics differ in different soil systems, thus potentially corresponding to different optimal parameter combinations. The parameters given in this study (such as sliding window half-width ±4 min, R...) 2 (≥0.8, etc.) is the relatively optimal solution obtained by comprehensive comparison under various soil conditions. It has good universality and stability, but it still needs to be appropriately optimized and adjusted in combination with actual experimental conditions in specific applications.

[0073] In summary, compared with existing traditional enzyme profiling methods, the time-series in-situ enzyme profiling method based on multi-parameter screening proposed in this invention effectively solves the problems of low signal-to-noise ratio, poor spatial resolution, and large quantitative error in traditional methods through the synergistic effect of time-series imaging, local linear fitting, and multi-level screening. It realizes the transformation from "single-time static characterization" to "dynamic process quantitative analysis", and shows significant improvements in enzyme activity detection accuracy, spatial recognition ability, and result stability, and has good application prospects.

[0074] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, 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 the present invention.

Claims

1. A time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening, characterized in that, Includes the following steps: S1. Construct an in-situ fluorescence imaging platform, attach the enzyme reaction membrane loaded with fluorescent substrate to the soil rhizosphere surface, and collect time-series fluorescence images of soil enzyme-catalyzed reactions in a dark room. S2. Perform background subtraction and noise reduction preprocessing on time-series fluorescence images; S3. Obtain the time-series image data and the pixel-level time-series grayscale data matrix G(t); S4. Calculate the maximum gray value G for each pixel in the data matrix G(t). max Initial screening is performed using the grayscale standard deviation σG, discarding those that do not meet G. max ≥G min And σG≥σ min The pixels, of which G min and σ min The values ​​range from 0.5 to 5; S5. For the pixels that pass the initial screening, apply a sliding window local linear fitting and calculate the gray-level change rate dG / dt and the goodness of fit R. 2 Only those satisfying the dG / dt slope threshold ≥ 0.05~1 and R 2 Fitting results with a value ≥0.6~0.95; S6. Extract the maximum local linear growth rate of each pixel as the uncorrected enzyme activity of that pixel, and introduce a diffusion correction model to correct it, so as to obtain the corrected enzyme activity. S7. The window center time t corresponding to the corrected enzyme activity center Perform final screening and retain only t center Pixels located in the 5-55 minute range are used to reconstruct the spatial distribution matrix of enzyme activity; S8. Prepare a fluorescence standard curve to obtain the grayscale-enzyme activity conversion coefficient, convert the corrected enzyme activity into quantitative enzyme activity data and output it visually.

2. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S1, the acquisition time of the time-series fluorescence image is 60 to 90 minutes, and the sampling time interval is 1 minute.

3. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S2, the background subtraction includes subtracting the background value using the first frame image; the noise reduction includes preliminary noise reduction using median filtering.

4. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S4, the G min and σ min The preferred value is 1.

5. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S5, the half-width of the sliding window is 2 to 10 minutes, and the number of window points is 5 to 21.

6. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 4, characterized in that: In step S5, the half-width of the sliding window for local linear fitting is preferably 4 minutes, and the number of window points is 9.

7. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S5, the preferred value of the dG / dt slope threshold is 0.1, and the preferred value of the R² threshold is 0.

8.

8. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S6, the diffusion correction model is an empirical function model, and its correction factor is... f corr The calculation formula is: in, p a , p b and p c The values ​​are 2.95, 26 and 0.84 respectively.

9. The time-series in-situ enzyme spectrum soil enzyme activity quantitative imaging method based on multi-parameter screening according to claim 1, characterized in that: In step S8, the fluorescence standard curve is prepared by immersing the enzyme reaction membrane in fluorescent standard solutions of different concentrations, performing fluorescence imaging under the same imaging conditions, establishing a linear relationship between grayscale value and fluorescence product concentration, and obtaining the conversion coefficient b; the conversion coefficient b includes: for acid phosphatase, b1 = 6.13 pmol MUF mm -2 Greyscale -1 For alkaline phosphatase, b2 = 5.33 pmol MUF mm -2 Greyscale -1 .

10. A system for implementing the method according to any one of claims 1 to 9, characterized in that, include: An imaging module is used to perform step S1 and acquire the time-series fluorescence image; The image preprocessing module is used to perform steps S2 and S3 to perform background subtraction and noise reduction on the image and obtain a temporal grayscale data matrix. The multi-parameter screening and fitting module is used to execute steps S4, S5 and S6 to perform initial pixel screening, sliding window fitting and diffusion correction. The post-processing and visualization module is used to execute steps S7 and S8, performing final data screening, spatial reconstruction, and visualization output.

Images