A fluorescence microscope calibration system based on machine learning

Abstract

The present invention relates to the technical field of optical instrument calibration, and particularly to a fluorescence microscope calibration system based on machine learning, including: an imaging data acquisition module for acquiring the original imaging data of a standard sample and the target imaging data of the standard sample, a device data acquisition module for acquiring the imaging environment data and the device data of the fluorescence microscope, a sample data processing module for processing the original imaging data of the standard sample and the target imaging data of the standard sample, a sample data analysis module for analyzing the imaging state of the target imaging data, an imaging quality analysis module for analyzing the imaging quality, a device anomaly analysis module for analyzing the device anomaly, and a calibration module for generating a calibration scheme for the fluorescence microscope. The present invention effectively improves the calibration efficiency and calibration accuracy of the fluorescence microscope.

Description

Technical Field

[0001] The present invention relates to the technical field of optical instrument calibration, and particularly to a fluorescence microscope calibration system based on machine learning. Background Art

[0002] Fluorescence microscopes are indispensable important tools in modern biomedical research. However, due to factors such as equipment aging and environmental changes, the imaging quality of fluorescence microscopes may deteriorate during long-term use, such as reduced resolution and color distortion. Traditional manual calibration methods are not only time-consuming and laborious, but also difficult to ensure the consistency and accuracy of each calibration. Therefore, developing a system that can achieve automated and intelligent calibration is of great significance for improving the working efficiency and imaging accuracy of fluorescence microscopes.

[0003] Chinese Patent Publication No. CN116957943A discloses a microscope stitching calibration method based on image fusion, including the following steps: collecting microscopic images between adjacent two frames at equal intervals; determining the change amount of blurriness of several consecutive adjacent microscopic images along the same axial direction; obtaining the position coordinates of key position points on the to-be-observed surface of the to-be-measured object in the coordinate system of the measurement platform; fitting the plane where the to-be-observed surface is located; analyzing the rotation angle and controlling the measurement platform to rotate; stitching and fusing the overlapping areas of adjacent two microscopic images; It can be seen that the calibration process of this invention for the microscope is a digital calibration of microscopic images, which does not analyze the microscope imaging process and does not calibrate the microscope imaging process, resulting in problems of low calibration efficiency and inaccurate calibration for the microscope. Summary of the Invention

[0004] The purpose of the present invention is to provide a fluorescence microscope calibration system based on machine learning to solve at least one of the problems existing in the prior art.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions:

[0006] A fluorescence microscope calibration system based on machine learning, characterized by including:

[0007] An imaging data acquisition module for acquiring the original imaging data and target imaging data of a standard sample;

[0008] A device data acquisition module for acquiring imaging environment data and device data of the fluorescence microscope;

[0009] A sample data processing module for processing the original imaging data and target imaging data of the standard sample to obtain original standard data, target image parameters, original processed images, and target processed images;

[0010] A sample data analysis module for analyzing the imaging state of target imaging data according to target image parameters;

[0011] An imaging quality analysis module for analyzing the clarity and light intensity abnormality of target imaging data according to the analysis result of the imaging state of target imaging data and the target processed image, and analyzing the imaging quality of target imaging data according to the analysis result;

[0012] A device abnormality analysis module for analyzing device abnormality according to the device data of the fluorescence microscope;

[0013] A calibration module for generating a calibration scheme for the fluorescence microscope according to the analysis result of the imaging quality of target imaging data and the analysis result of device abnormality of the fluorescence microscope.

[0014] Further, the sample data processing module includes an original data processing unit, which is used to identify the staining pixel position coordinates of the staining object in the original imaging data of the standard sample, and eliminate the background pixels of the original imaging data according to the staining pixel position coordinates of the staining object to obtain an original processed image;

[0015] The image data processing unit is also used to calculate the standard uniformity α1 according to the original processed image;

[0016] The image data processing unit is also used to obtain the standard contrast β1 of the original processed image, and use the standard contrast β1 and the standard uniformity α1 as the original standard data.

[0017] Further, the sample data analysis module is used to analyze the imaging state of the target imaging data according to the target image parameters, and send an imaging result signal to the user according to the analysis result: if α2 < α1×η1 and β1×η2 ≤ β2 < β1×η3, the sample data analysis module determines that the imaging state of the target imaging data is normal and sends an imaging success signal to the user; if α2 ≥ α1×η1 or β2 [β1×η2, β1×η3), the sample data analysis module determines that the imaging state of the target imaging data is abnormal and sends an imaging failure signal to the user;

[0018] Where η1 is a preset fluorescence imaging uniformity, η2 is a first preset contrast, η3 is a second preset contrast, and η2 < η3.

[0019] Further, the imaging quality analysis module includes an image light intensity analysis unit. The image light intensity analysis unit is used to grayscale the original processed image and the target processed image, and when the imaging state of the target imaging data is normal, calculate the light intensity difference index gz. At this time, if gz exceeds the preset light intensity difference index GZ, the image light intensity analysis unit determines that the light intensity of the target imaging data is abnormal; otherwise, it determines that the light intensity of the target imaging data is normal.

[0020] Further, the imaging quality analysis module further includes an ambient light intensity analysis unit. The ambient light intensity analysis unit is used to analyze the abnormality of ambient light according to the imaging environment data and the device data of the fluorescence microscope. The analysis result of the ambient light abnormality includes normal and abnormal, and when the ambient light is abnormal, the preset light intensity difference index is adjusted to GZ'.

[0021] Further, the imaging quality analysis module further includes an exposure analysis unit. The exposure analysis unit is used to analyze the exposure abnormality according to the exposure duration t of the target imaging data. The analysis result of the exposure abnormality includes normal, first abnormality, and second abnormality, and when there is a first abnormality, the preset interference light intensity is optimized to YQ'; when there is a second abnormality, the preset interference light intensity is optimized to YQ".

[0022] Further, the imaging quality analysis module is also provided with a clarity analysis unit, which is used to calculate the clarity index ▽T of the target processed image when the imaging state of the target imaging data is normal, and analyze the clarity of the target imaging data according to the clarity index ▽T of the target processed image. The analysis result of the clarity of the target imaging data includes abnormal clarity and normal clarity.

[0023] Further, the imaging quality analysis module is also provided with an imaging quality analysis unit. The imaging quality analysis unit is used to analyze the imaging quality of the target imaging data according to the clarity analysis result and the light intensity abnormality analysis result of the target imaging data. The analysis result of the imaging quality of the target imaging data includes good, unqualified, and qualified.

[0024] Further, the device abnormality analysis module is used to calculate the device abnormality index γ according to the device data of the fluorescence microscope, and analyze the device abnormality according to the device abnormality index. The analysis result of the device abnormality includes device normal and device abnormal.

[0025] Further, the calibration module includes a calibration unit. The calibration unit is used to generate a calibration scheme for the fluorescence microscope according to the imaging quality analysis result of the target imaging data and the device abnormality analysis result of the fluorescence microscope:

[0026] When the imaging quality of the target imaging data is good, the calibration unit does not calibrate the fluorescence microscope;

[0027] When the imaging quality is qualified and the equipment is normal, if a1×(▽T - YT) / YT + a2×(gz - GZ) / GZ < H, the calibration unit determines that the light source stability of the fluorescence microscope is abnormal and calibrates the excitation power of the excitation light source to P'; where H is the preset light source stability index, a1 is the light intensity weight, a2 is the clarity weight, and a1 + a2 = 1;

[0028] When the imaging quality is unqualified, the calibration unit determines that there is a malfunction in the fluorescence microscope equipment and recommends that the user replace the filter and the excitation light source;

[0029] When the imaging quality is qualified and the equipment is abnormal, the calibration unit determines the filter error of the fluorescence microscope and adjusts the excitation wavelength range to DG';

[0030] The calibration unit also takes the adjusted excitation wavelength range and the calibrated excitation power as the calibration scheme.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows: By comparing the standard imaging image of the fluorescence microscope with the currently acquired target imaging image, the imaging state of the current fluorescence microscope is obtained. Then, by analyzing the light source state of the target imaging, the state of the light source structure of the fluorescence microscope is analyzed. Combining with the analysis result of the equipment state of the fluorescence microscope, the fluorescence microscope is calibrated in terms of the light source and equipment, effectively improving the calibration efficiency and calibration accuracy of the fluorescence microscope. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0033] Figure 1 It is a schematic structural diagram of the fluorescence microscope calibration system based on machine learning in this embodiment.

[0034] Figure 2 It is a schematic structural diagram of the sample data processing module in this embodiment.

[0035] Figure 3 It is a schematic structural diagram of the imaging quality analysis module in this embodiment.

[0036] Figure 4 It is a schematic structural diagram of the calibration module in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] To illustrate the present invention more clearly, the present invention will be further described below in conjunction with preferred embodiments and the accompanying drawings. Similar components in the drawings are denoted by the same reference numerals. Those skilled in the art should understand that the specific content described below is illustrative rather than restrictive, and should not be used to limit the protection scope of the present invention.

[0038] It should be noted that although terms such as first, second, and third may be used in the embodiments of the present application for description, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0039] Specifically, a fluorescence microscope calibration system based on machine learning described in this embodiment is applied to the monochromatic imaging calibration of a fluorescence microscope; the fluorescence microscope described in this embodiment is specifically a wide-field fluorescence microscope used for cell staining; the system described in this embodiment acts on a connected computer of the fluorescence microscope, and it is data-connected to the fluorescence microscope through a circuit.

[0040] Please refer to Figure 1 as shown, which is a schematic structural diagram of the fluorescence microscope calibration system based on machine learning described in this embodiment, including:

[0041] An imaging data acquisition module, which is used to acquire the original imaging data of a standard sample and the target imaging data of the standard sample; the standard sample is a fixed detection object for fluorescence microscope calibration, which is specifically a biological cell staining sample, and the production method of the biological cell staining sample is not specifically limited in this embodiment; the original imaging data of the standard sample is the best historical imaging image of the standard sample under the fluorescence microscope, and its acquisition method is obtained through user interaction input; the target imaging data of the standard sample is the image obtained by currently imaging the standard sample through the fluorescence microscope; in this embodiment, the target imaging data of the standard sample is obtained by the fluorescence microscope shooting in the form of digital signals.

[0042] It can be understood that when the original imaging data and the target imaging data are collected by the fluorescence microscope in this embodiment, the imaging devices and operations such as the shooting angle, the objective lens height, the eyepiece and objective lens models and magnifications are the same, that is: the collection processes of the original imaging data and the target imaging data are the same.

[0043] Please continue to refer to Figure 1 as shown, the system further includes:

[0044] A device data acquisition module, which is used to acquire the imaging environment data and the device data of the fluorescence microscope; the imaging environment data includes the illumination intensity of the imaging environment,

[0045] The device data of the fluorescence microscope includes the excitation wavelength range, excitation light intensity, emission wavelength range, fluorescence intensity, and filter service life.

[0046] Specifically, in this embodiment, the excitation wavelength range, excitation light intensity, emission wavelength range, and fluorescence intensity in the device data of the fluorescence microscope are all collected by a spectrometer, and the filter service life is collected through user interaction input; the imaging environment data is collected by setting various intelligent sensors at the edge of the fluorescence microscope.

[0047] Please continue to refer to Figure 1 As shown, the system further includes:

[0048] A sample data processing module for processing the original imaging data of the standard sample and the target imaging data of the standard sample.

[0049] Please refer to Figure 2 As shown, the sample data processing module includes an original data processing unit, which is used to process the original imaging data of the standard sample to obtain the original standard data of the original imaging data, and use the original standard data to represent the best imaging parameters of the original imaging data;

[0050] The image data processing unit identifies the staining pixel position coordinates of the stained object in the original imaging data of the standard sample, and eliminates the background pixels of the original imaging data according to the staining pixel position coordinates of the stained object to obtain the original processed image;

[0051] The image data processing unit is also used to calculate the standard uniformity α1 according to the original processed image, and set ;

[0052] Among them, YN represents the number of pixel points in the original imaging data, Qyn represents the fluorescence intensity of the yn-th pixel point, and QJ represents the average fluorescence intensity of each pixel point in the original imaging data;

[0053] The image data processing unit is also used to obtain the standard contrast β1 of the original processed image, and use the standard contrast β1 and the standard uniformity α1 as the original standard data; by calculating the original standard data to extract the parameters representing the quality in the original imaging data, and using this original standard data as the digital basis for judging the imaging state, the accuracy of imaging state judgment is improved.

[0054] Specifically, in this embodiment, the process of staining objects in the original imaging data of the recognition sample standard is not specifically limited. In this embodiment, the recognition is performed through professional recognition software (BD FACSChorus™ software) connected to a computer in a fluorescence microscope. Those skilled in the art can use other methods for recognizing stained objects. At the same time, the process of obtaining the standard contrast ratio in this embodiment is a publicly known prior art known to those skilled in the art, and this embodiment does not specifically limit it. In this embodiment, the standard contrast ratio is represented by calculating the contrast ratio.

[0055] Please continue to refer to Figure 2 As shown, the system further includes a target data processing unit. The target data processing unit is connected to the original data processing unit. The target data processing unit is used to process the target imaging data of the standard sample to obtain the target image parameters of the target imaging data;

[0056] The target data processing unit obtains a target processed image based on the target imaging data, and calculates target image parameters based on the target processed image;

[0057] The target image data has the same format as the original standard data. The target image data includes a target uniformity α2 and a target contrast β2.

[0058] Specifically, the process of "obtaining a target processed image based on the target imaging data and calculating target image parameters" in this embodiment is the same as the analysis and calculation process of the original processed image and the original standard data, and this embodiment will not elaborate.

[0059] Please continue to refer to Figure 1 As shown, the system further includes a sample data analysis module. The sample data analysis module is connected to the sample data processing module. It is used to analyze the imaging state of the target imaging data based on the target image parameters, and send an imaging result signal to the user according to the analysis result: if α2 < α1 × η1 and β1 × η2 ≤ β2 < β1 × η3, the sample data analysis module determines that the imaging state of the target imaging data is normal and sends an imaging success signal to the user; if α2 ≥ α1 × η1 or β2 ∉[β1 × η2, β1 × η3), the sample data analysis module determines that the imaging state of the target imaging data is abnormal and sends an imaging failure signal to the user. By using the original standard data to judge the imaging state of the target image parameters of the target sample data, obvious target imaging data that does not meet the calibration standard is excluded, improving the calibration efficiency of the fluorescence microscope;

[0060] Among them, η1 is a preset fluorescence imaging uniformity, η2 is a first preset contrast ratio, η3 is a second preset contrast ratio, and η2 < η3.

[0061] Specifically, in this embodiment, no specific limitations are imposed on the values of the preset fluorescence imaging uniformity η1, the first preset contrast η2, and the second preset contrast η3. It only needs to meet the value requirements of the preset fluorescence imaging uniformity η1, the first preset contrast η2, and the second preset contrast η3. In this embodiment, the preset fluorescence imaging uniformity η1 is set to 0.5, the first preset contrast η2 is set to 0.8, and the second preset contrast η3 is set to 1.2.

[0062] Please continue to refer to Figure 1 As shown, the system further includes an imaging quality analysis module. The imaging quality analysis module is connected to the sample data analysis module and the device data acquisition module. The imaging quality analysis module is used to analyze the imaging quality based on the analysis result of the imaging state of the target imaging data and the target processed image.

[0063] Please refer to Figure 3 As shown, the imaging quality analysis module includes an image light intensity analysis unit. It is used to analyze the light intensity abnormality of the target imaging data based on the analysis result of the imaging state of the target imaging data, the original processed image, and the target processed image, and use the situation of light intensity abnormality to represent the light leakage phenomenon of the background in the target imaging data.

[0064] The image light intensity analysis unit is used to grayscale the original processed image and the target processed image. When the imaging state of the target imaging data is normal, it calculates the light intensity difference index gz, and sets gz = Σgc(i), where gc(i) is the light difference of the i-th pixel point, and sets gc(i) = |ys(i) - mb(i)| / N. N is the number of pixel points of the original processed image, ys(i) is the gray value of the i-th pixel point in the original processed image, and mb(i) is the gray value of the i-th pixel point in the target processed image. At this time, if gz < GZ, the image light intensity analysis unit determines that the light intensity of the target imaging data is normal; if gz ≥ GZ, the image light intensity analysis unit determines that the light intensity of the target imaging data is abnormal; where GZ is the preset light intensity difference index; by calculating the light intensity difference of the target processed image and using this to represent the proportion of the background light intensity in the image brightness, and then analyzing the light intensity abnormality, the accurate judgment of the light leakage phenomenon of the target imaging data is realized.

[0065] Specifically, in this embodiment, no specific limitations are imposed on the value of the preset light intensity difference index GZ. Those skilled in the art can freely set it as long as it meets the value requirements of the preset light intensity difference index GZ. The best value of the preset light intensity difference index GZ in this embodiment is 10.

[0066] Please continue to refer to Figure 3As shown, the imaging quality analysis module further includes an ambient light intensity analysis unit. The ambient light intensity analysis unit is connected to the image light intensity analysis unit. The ambient light intensity analysis unit is used to analyze the abnormality of ambient light according to the imaging environment data and the device data of the fluorescence microscope, and adjust the light intensity abnormality analysis process of the target imaging data according to the analysis result of the ambient light abnormality: If QHG < YQ, the ambient light intensity analysis unit determines that the ambient light is normal and does not make any adjustment; if QHG ≥ YQ, the ambient light intensity analysis unit determines that the ambient light is abnormal and adjusts the preset light intensity difference index to GZ’, where GZ’ = GZ × [1 - (QHG - YQ) / QHG]; where QHG ∈ DG, QHG is the light intensity in the ambient light that belongs to the excitation wavelength range DG, DG is the excitation wavelength range and DG = {BC1 ≤ LX ≤ BC2}, BC1 and BC2 are the left value and the right value of the excitation wavelength range respectively, and YQ is the preset interference light intensity; By analyzing the abnormality of the ambient light intensity, the interference of the ambient natural light on the fluorescence imaging is excluded, and the analysis result of the light intensity abnormality is adjusted according to the analysis result, so that the analysis result of the light intensity abnormality is more accurate.

[0067] Specifically, in this embodiment, the value of the preset interference light intensity YQ is not specifically limited, and those skilled in the art can freely set it as long as it meets the value requirement of the preset interference light intensity YQ. The best value of the preset interference light intensity YQ in this embodiment is 50 lx.

[0068] Please continue to refer to Figure 3 As shown, the imaging quality analysis module further includes an exposure analysis unit. The exposure analysis unit is connected to the ambient light intensity analysis unit. The exposure analysis unit is used to analyze the exposure abnormality according to the exposure duration of the target imaging data, and optimize the light intensity abnormality adjustment process of the target imaging data according to the analysis result, where the exposure abnormality refers to the light leakage situation caused by the overlong exposure when the light intensity abnormality is represented:

[0069] When the light intensity of the target imaging data is normal, if t < T1, the exposure analysis unit determines that the exposure of the target imaging data is normal and does not perform any optimization; if t ≥ T1, the exposure analysis unit determines that the exposure of the target imaging data is the first abnormality and optimizes the preset interference light intensity to YQ’, where YQ’ = YQ × ln[e-(t - T1) / T1], and e is the natural logarithm;

[0070] When the light intensity of the target imaging data is abnormal, if t < T2, the exposure analysis unit determines that the exposure of the target imaging data is normal and does not perform optimization; if t ≥ T2, the exposure analysis unit determines that the exposure of the target imaging data is the second anomaly, and optimizes the preset interference light intensity to YQ”, where YQ” = YQ × [1 - (t - T2) / T2]; through the comparison and analysis of the exposure duration, the accurate analysis of the light leakage caused by exposure can be realized, and then the adjustment process of strong anomaly can be optimized, making the analysis result of light intensity anomaly more accurate;

[0071] Among them, t is the exposure duration of the target imaging data, T1 is the first preset exposure duration, T2 is the second preset exposure duration, and T1 > T2.

[0072] Specifically, in this embodiment, the values of the first preset exposure duration T1 and the second preset exposure duration T2 are not specifically limited, and those skilled in the art can freely set them as long as the value requirements of the first preset exposure duration T1 and the second preset exposure duration T2 are met. In this embodiment, the best value of the first preset exposure duration T1 is 100 ms, and the best value of the second preset exposure duration T2 is 60 ms.

[0073] It can be understood that the exposure duration in this embodiment is an imaging parameter controlled by the fluorescence microscope itself.

[0074] Please continue to refer to Figure 3 As shown, the imaging quality analysis module is also provided with a clarity analysis unit, which is used to analyze the clarity of the target imaging data according to the target processed image;

[0075] When the imaging state of the target imaging data is normal, the clarity analysis unit calculates the clarity index ▽T of the target processed image, and sets , where D represents the pixel area of the target processed image, and D = {(x, y)|0 ≤ x ≤ X, 0 ≤ y ≤ Y}, X is the maximum number of horizontal pixel points of the target processed image, Y is the maximum number of vertical pixel points of the target processed image, Gx represents the gray gradient of the pixel point with coordinates (x, y) on the horizontal coordinate axis, Gy represents the gray gradient of the pixel point with coordinates (x, y) on the vertical coordinate axis, and ZG represents the total gray value of the target processed image; the clarity analysis unit quantifies the clarity state presented by the target imaging data with the clarity index ▽T;

[0076] The clarity analysis unit analyzes the clarity of the target imaging data according to the clarity index ▽T of the target processed image: if ▽T < YT, the clarity analysis unit determines that the clarity of the target imaging data is abnormal; if ▽T ≥ YT, the clarity analysis unit determines that the clarity of the target imaging data is normal; where YT is the preset image clarity; by judging the clarity of the target imaging data and reflecting the stability of the fluorescence microscope light source thereby (when the light source is unstable, the wavelength of the excitation light emitted by the light source will also be unstable, resulting in blurred imaging), an accurate judgment of the clarity is achieved.

[0077] Specifically, in this embodiment, no specific limitation is imposed on the value of the preset image clarity YT, and those skilled in the art can freely set it as long as the value requirement of the preset image clarity YT is satisfied. The optimal value of the preset image clarity YT in this embodiment is 0.2.

[0078] It can be understood that in this embodiment, no specific analysis is made on the blurred imaging phenomenon caused by focusing. It is defaulted in this embodiment that the focusing of the target imaging data is accurate; at the same time, in this embodiment, the calculation processes of Gx and Gy are Gx = gray(x + 1, y) - gray(x - 1, y), Gy = gray(x, y + 1) - gray(x, y - 1), where gray(x + 1, y) represents the gray value of the pixel point with coordinates (x + 1, y), and the meanings of gray(x, y + 1), gray(x, y - 1) and gray(x - 1, y) are similar to that of gray(x + 1, y), which are the gray values of pixel points with different coordinates respectively, and are not elaborated in this embodiment.

[0079] Please continue to refer to Figure 3 As shown, the imaging quality analysis module is also provided with an imaging quality analysis unit. The imaging quality analysis unit is connected to the clarity analysis unit and the image light intensity analysis unit. The imaging quality analysis unit is used to analyze the imaging quality of the target imaging data according to the clarity analysis result and the light intensity abnormality analysis result of the target imaging data: if the light intensity of the target imaging data is normal and the clarity is normal, the imaging quality analysis unit determines that the imaging quality of the target imaging data is good; if the clarity of the target imaging data is abnormal, the imaging quality analysis unit determines that the imaging quality of the target imaging data is unqualified; if the light intensity of the target imaging data is abnormal and the clarity is normal, the imaging quality analysis unit determines that the imaging quality of the target imaging data is qualified.

[0080] Specifically, in this embodiment, the "good" is a level of imaging quality above the "qualified", and the "qualified" in this embodiment does not include the "good".

[0081] Please continue to refer to Figure 1As shown, the system further includes a device anomaly analysis module, which is connected to the device data acquisition module and is used to analyze the device anomaly based on the device data of the fluorescence microscope;

[0082] The device anomaly analysis module calculates a device anomaly index γ based on the device data of the fluorescence microscope, and sets γ = (W - w) / W × (m - M) / M; where W is the set excitation wavelength of the fluorescent material, w is the actual excitation wavelength of the target imaging data, m is the remaining service life of the filter of the fluorescence microscope, and M is the total service life of the fluorescence microscope;

[0083] The device anomaly analysis module analyzes the device anomaly based on the device anomaly index: if γ < K, the device anomaly analysis module determines that the device of the fluorescence microscope is normal; if γ ≥ K, the device anomaly analysis module determines that the device of the fluorescence microscope is abnormal; where K is a preset device anomaly index; by indirectly analyzing the aging degree of the consumable parts of the fluorescence microscope and directly analyzing the wavelength direct error through the device anomaly analysis module, an accurate judgment of the device state is achieved.

[0084] Specifically, in this embodiment, no specific limitation is made on the value of the preset device anomaly index K, and those skilled in the art can freely set it as long as it meets the value requirement of the preset device anomaly index K. The best value of the preset device anomaly index K in this embodiment is 0.5.

[0085] Please continue to refer to Figure 1 As shown, the system further includes a calibration module, which is connected to the device anomaly analysis module and the data status analysis module, and is used to generate a calibration plan for the fluorescence microscope according to the data status analysis result of the target imaging data and the device anomaly analysis result of the fluorescence microscope, and output the calibration plan to the user.

[0086] Please refer to Figure 4 As shown, the calibration module includes a calibration unit, which is used to generate a calibration plan for the fluorescence microscope according to the imaging quality analysis result of the target imaging data and the device anomaly analysis result of the fluorescence microscope:

[0087] When the imaging quality of the target imaging data is good, the calibration unit does not calibrate the fluorescence microscope;

[0088] When the imaging quality is qualified and the equipment is normal, if a1×(▽T - YT) / YT + a2×(gz - GZ) / GZ < H, the calibration unit determines that the light source stability of the fluorescence microscope is abnormal, and calibrates the excitation power of the excitation light source to P’, where P’ = P×ln{e - [a1×(▽T - YT) / YT + a2×(gz - GZ) / GZ - K] / K}; where H is the preset light source stability index, P is the actual excitation power of the excitation light source, a1 is the light intensity weight, a2 is the clarity weight, and a1 + a2 = 1;

[0089] When the imaging quality is unqualified, the calibration unit determines that the fluorescence microscope equipment is faulty and recommends that the user replace the filter and the excitation light source;

[0090] When the imaging quality is qualified and the equipment is abnormal, the calibration unit determines the filter error of the fluorescence microscope and adjusts the excitation wavelength range to DG’, where DG’ = {BC1’ ≤ LX ≤ BC2’}, and BC1’ = BC1×exp[(γ - K) / K], BC2’ = BC2×exp[(K - γ) / K];

[0091] The calibration unit also takes the adjusted excitation wavelength range and the calibrated excitation power as the calibration scheme; by combining the analysis results of the equipment status and the imaging quality through the calibration unit, the actual error components of the fluorescence microscope reflected by the two are specifically analyzed and calibrated, effectively improving the calibration efficiency and calibration accuracy of the fluorescence microscope.

[0092] Specifically, in this embodiment, the value of the preset light source stability index H is not specifically limited, and those skilled in the art can freely set it as long as it meets the value requirements of the preset light source stability index H. The best value of the preset light source stability index H in this embodiment is 0.3; the acquisition method of the actual excitation power P of the excitation light source in this embodiment is obtained through the connected computer of the fluorescence microscope.

[0093] Please continue to refer to Figure 4 As shown, the calibration module further includes an output unit, which is connected to the calibration unit, and the output unit is used to output the calibration scheme of the fluorescence microscope to the user.

[0094] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A fluorescence microscope calibration system based on machine learning, characterized in that: include: An imaging data acquisition module, used to acquire original imaging data of the standard sample and target imaging data of the standard sample; Equipment data acquisition module, used to collect imaging environment data and equipment data of fluorescence microscope; A sample data processing module is used to process the original imaging data of the standard sample and the target imaging data of the standard sample to obtain original standard data, target image parameters, original processed images and target processed images; A sample data analysis module is used to analyze the imaging state of the target imaging data according to the target image parameters; An imaging quality analysis module is used to analyze the clarity and light intensity abnormality of the target imaging data according to the imaging state analysis results of the target imaging data and the target processed image, and to analyze the imaging quality of the target imaging data according to the analysis results; Equipment abnormality analysis module, used to analyze equipment abnormality based on equipment data from the fluorescence microscope; A calibration module, used to generate a calibration scheme for the fluorescence microscope according to an imaging quality analysis result of the target imaging data and an equipment abnormality analysis result of the fluorescence microscope; The imaging quality analysis module includes an image light intensity analysis unit, which is used to grayscale the original processed image and the target processed image, and calculate the light intensity difference index gz when the imaging state of the target imaging data is normal. At this time, if gz exceeds the preset light intensity difference index GZ, the image light intensity analysis unit determines that the light intensity of the target imaging data is abnormal; Otherwise, it is determined that the light intensity of the target imaging data is normal; The imaging quality analysis module further includes an ambient light intensity analysis unit, which is used to analyze the ambient light abnormality according to the imaging environment data and the equipment data of the fluorescence microscope, wherein the ambient light abnormality analysis result includes normal and abnormal, and when the ambient light is abnormal, the preset light intensity difference index is adjusted to GZ'; The imaging quality analysis module also includes an exposure analysis unit, which is used to analyze exposure abnormality according to the exposure time t of the target imaging data. The analysis results of the exposure abnormality include normal, first abnormality and second abnormality. When the first abnormality occurs, the preset interference light intensity is optimized to YQ'; when the second abnormality occurs, the preset interference light intensity is optimized to YQ".

2. The fluorescence microscope calibration system based on machine learning according to claim 1, characterized in that: The sample data processing module includes a raw data processing unit, which is used to identify the stained pixel position coordinates of the stained object in the raw imaging data of the standard sample, and eliminate the background pixels of the raw imaging data according to the stained pixel position coordinates of the stained object to obtain an original processed image; The raw data processing unit is also used to calculate the standard uniformity α1 according to the raw processed image; The raw data processing unit is further used to obtain a standard contrast β1 of the raw processed image, and use the standard contrast β1 and the standard uniformity α1 as raw standard data.

3. The fluorescence microscope calibration system based on machine learning according to claim 2, characterized in that: The sample data analysis module is used to analyze the imaging state of the target imaging data according to the target image parameters, and send an imaging result signal to the user according to the analysis result: if α2<α1×η1 and β1×η2≤β2<β1×η3, the sample data analysis module determines that the imaging state of the target imaging data is normal, and sends an imaging success signal to the user; if α2≥α1×η1 or β2 (β1×η2, β1×η3), the sample data analysis module determines that the imaging state of the target imaging data is abnormal, and sends an imaging failure signal to the user; Wherein, η1 is a preset fluorescence imaging uniformity, η2 is a first preset contrast, η3 is a second preset contrast, η2<η3, α2 is a target uniformity, and β2 is a target contrast.

4. The fluorescence microscope calibration system based on machine learning according to claim 3, characterized in that: The imaging quality analysis module is also provided with a clarity analysis unit, which is used to calculate the clarity index ▽T of the target processed image when the imaging state of the target imaging data is normal, and analyze the clarity of the target imaging data according to the clarity index ▽T of the target processed image. The clarity analysis results of the target imaging data include abnormal clarity and normal clarity.

5. The fluorescence microscope calibration system based on machine learning according to claim 4, characterized in that: The imaging quality analysis module is also provided with an imaging quality analysis unit, which is used to analyze the imaging quality of the target imaging data according to the clarity analysis results and the light intensity abnormality analysis results of the target imaging data. The imaging quality analysis results of the target imaging data include good, unqualified and qualified.

6. The fluorescence microscope calibration system based on machine learning according to claim 5, characterized in that: The device abnormality analysis module is used to calculate the device abnormality index γ according to the device data of the fluorescence microscope, and analyze the device abnormality according to the device abnormality index. The analysis result of the device abnormality includes the device being normal and the device being abnormal.

7. The fluorescence microscope calibration system based on machine learning according to claim 6, characterized in that: The calibration module includes a calibration unit, which is used to generate a calibration scheme for the fluorescence microscope according to an imaging quality analysis result of the target imaging data and an equipment abnormality analysis result of the fluorescence microscope: When the imaging quality of the target imaging data is good, the calibration unit does not calibrate the fluorescence microscope; When the imaging quality is qualified and the device is normal, if a1×(▽T-YT) / YT+a2×(gz-GZ) / GZ<H, the calibration unit determines that the light source stability of the fluorescence microscope is abnormal, and calibrates the excitation power of the excitation light source to P'; wherein H is a preset light source stability index, a1 is a light intensity weight, a2 is a clarity weight, and a1+a2=1; When the imaging quality is unqualified, the calibration unit determines that the fluorescence microscope device is faulty and recommends that the user replace the filter and the excitation light source; When the imaging quality is qualified and the device is abnormal, the calibration unit determines the filter error of the fluorescence microscope and adjusts the excitation wavelength range to DG'; The calibration unit also uses the adjusted excitation wavelength range and the calibrated excitation power as a calibration scheme, and YT is a preset image clarity.

Images