Single-molecule positioning super-resolution imaging chromatic aberration correction method and system

Abstract

The invention relates to the technical field of image data processing or generation, in particular to a monomolecular positioning super-resolution imaging chromatic aberration correction method and system, and the method comprises the following steps: removing misregistration points from a plurality of groups of transverse moving images of two different wavelength channels; the reserved points are registered to obtain a chromatic aberration correction mapping function, and single-molecule positioning super-resolution imaging transverse chromatic aberration correction is completed; gaussian fitting and linear least square fitting are carried out on the multiple sets of axial height data of each wavelength channel, the average deviation of the two channels at different axial heights is obtained and compensated to a single-molecule positioning super-resolution imaging chromatic aberration correction system, and single-molecule positioning super-resolution imaging axial chromatic aberration correction is completed. According to the method and the system provided by the invention, the correction of transverse and axial chromatic aberration in multi-color single-molecule positioning super-resolution imaging is realized, the registration precision and the correction speed of the system are improved, and the correction cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing or generation, and in particular, to a method and system for chromatic aberration correction in single-molecule localization super-resolution imaging. Background Art

[0002] The resolution of traditional optical microscopes is limited by the diffraction limit of light, usually about 200 nanometers, which makes it impossible for them to effectively resolve subcellular structures. Single-molecule localization super-resolution imaging technology (Single-Molecule Localization Microscopy, SMLM) breaks through the diffraction limit of traditional optical microscopes and realizes the imaging ability with nanoscale resolution, providing a powerful tool for cell biology research. However, in multi-color SMLM imaging, although the optical system has been corrected for chromatic aberration, there are still nanoscale residual chromatic aberrations. These chromatic aberrations cause spatial offsets between imaging channels of different wavelengths, affecting the accuracy of co-localization analysis in multi-color imaging. For the lateral chromatic aberration correction of multi-color SMLM imaging, the conventional method is the Locally Weighted Mean (LWM) algorithm, which calculates the weighted average according to the distance of neighboring points, making the closer point pairs have a greater impact on the result, so as to achieve accurate spatial registration in the local area. However, this method relies on the preparation of complex fluorescent microsphere arrays, the preparation equipment is relatively expensive and the preparation process is relatively complex. At the same time, global registration needs to be carried out by means of interpolation, and the registration accuracy is relatively low. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for chromatic aberration correction in single-molecule localization super-resolution imaging, which realizes the correction of lateral chromatic aberration and axial chromatic aberration in multi-color single-molecule localization super-resolution imaging, improves the registration accuracy of the system, and has low cost and high speed.

[0004] A method for chromatic aberration correction in single-molecule localization super-resolution imaging, comprising the following steps: The lateral chromatic aberration correction includes the following steps: S11: Sequentially excite a lateral chromatic aberration correction sample on a glass slide with two different wavelengths of excitation light, and obtain the pixel information and lateral coordinate information of multiple groups of lateral movement images in two different wavelength channels; S12: Determine the localization accuracy of the lateral chromatic aberration correction sample in each wavelength channel according to the pixel information of multiple groups of lateral movement images in two different wavelength channels, eliminate the points with the localization accuracy of the lateral chromatic aberration correction sample in each wavelength channel greater than the set threshold, and eliminate the misregistration points between different wavelength channels; S13: Register the points retained in the two different wavelength channels to obtain a chromatic aberration correction mapping function, and apply the chromatic aberration correction mapping function to complete the lateral chromatic aberration correction of single-molecule localization super-resolution imaging; The longitudinal chromatic aberration correction includes the following steps: S21: Place the axial chromatic aberration correction sample on a glass slide treated with polylysine, sequentially excite the axial chromatic aberration correction sample with two different wavelengths of excitation light, and obtain the pixel information and axial height data of multiple groups of axial movement images in the two different wavelength channels; S22: Perform Gaussian fitting on the distribution of multiple groups of axial height data in each wavelength channel respectively to obtain the axial height data after fitting in the two channels; S23: Use linear least squares fitting based on the axial height data after fitting in the two channels to establish an axial chromatic aberration relationship between the channels, so as to obtain the average deviation at different axial heights in the two channels; S24: Compensate the average deviation at different axial heights in the two channels into the chromatic aberration correction system of single-molecule localization super-resolution imaging to complete the axial chromatic aberration correction of single-molecule localization super-resolution imaging.

[0005] Optimally, the lateral chromatic aberration correction sample described in step S11 is multiple fluorescent microspheres, and the axial chromatic aberration correction sample described in step 21 is a dual-color fluorescent secondary antibody.

[0006] Optimally, the two different wavelengths of excitation light described in step S11 and step S21 are respectively: excitation light with a wavelength of 647 nm and excitation light with a wavelength of 561 nm.

[0007] Optimally, when obtaining multiple groups of lateral movement images in two different wavelength channels in step S11, the exposure time of each frame of image is 100 milliseconds, and the acquisition range of a single image field of view is square micrometers. When obtaining multiple groups of axial movement images in two different wavelength channels in step S21, the exposure time of each frame of image is 10 milliseconds, and the acquisition range of a single image field of view is square micrometers.

[0008] Optimally, based on the pixel information of multiple groups of lateral movement images in two different wavelength channels, calculate the lateral chromatic aberration correction sample positioning accuracy of the two different wavelength channels respectively according to formula (1): (1); Where: represents the lateral chromatic aberration correction sample positioning accuracy of the two different wavelength channels, represents the expected value of the Gaussian distribution, represents the number of photons in the light spot, represents the pixel information of the lateral movement images in the two different wavelength channels, Indicates background noise.

[0009] Furthermore, in step S12, the DBSCAN clustering algorithm based on distance adaptive change is used to eliminate the misregistration points between different wavelength channels, which specifically includes the following steps: S121: Take the positioning points of all fluorescent microspheres in two different wavelength channels as control point pairs, and take the distance between the control point pairs formed by the fluorescent microspheres at the outermost edge in the two different wavelength channels as the search radius. With the center position of the image as the center of the circle, calculate the scale factor according to formula (2): (2); Where: represents the search radius, represents the scale factor, represents the distance from the current control point to the center of the image, represents the initial search radius; S122: Traverse all control points in one of the wavelength channels, calculate the search radius according to formula (2), and then, with this control point as the center of the circle, check whether there is a registration point in the other wavelength channel within the calculated search radius. If there is no registration point in the other wavelength channel, regard this control point as a noise point and eliminate it. If there are multiple registration points in the other wavelength channel, eliminate this control point and the multiple registration points. If there is a unique registration point in the other wavelength channel, retain this control point and the registration point as a control point pair.

[0010] Furthermore, in step S13, the following method is used to obtain the color difference correction mapping function: S131: Use a second-order polynomial to establish the mapping relationship between the reference channel and the channel to be registered as formula (3): (3); Where: represents the coordinate of the control point in the axis direction, represents the coordinate of the control point in the axis direction, represents the mapping function between the reference channel and the channel to be registered, represents the set of control points in the reference channel, represents the set of control points in the channel to be registered, represents axis direction mapping function, represents axis direction mapping function, represents axis direction mapping function of the first-order term coefficient of, represents The linear term coefficient of the axial direction mapping function, is represented by the linear term coefficient of the axial direction mapping function, is represented by the quadratic term coefficient of the axial direction mapping function, is represented by the cross term coefficient of the axial direction mapping function, is represented by the constant term coefficient of the axial direction mapping function, the linear term coefficient of the is represented by the linear term coefficient of the axial direction mapping function, is represented by the quadratic term coefficient of the axial direction mapping function, is represented by the quadratic term coefficient of the axial direction mapping function, is represented by the cross term coefficient of the axial direction mapping function, is represented by; the constant term coefficient of the axial direction mapping function; S132: Based on the reserved control point pairs, for any control point in the reference channel, find the control point closest to it in the channel to be registered, calculate the sum of the squares of the distances between the two control points, and calculate the error function according to Equation (4) based on the sum of the squares of the distances between the two control points: (4); Where: represents the error function, represents the influence weight of the th control point, represents the sum of the squares of the distances between the two control points, represents the Euclidean distance between the th control point and the registration point, represents the desired minimum target distance, represents the number of control points; S133: Iterate based on the constructed error function, and during the iteration process, compare the relationship between the current control point error and the set threshold. If the current control point error is greater than the set threshold, then change the coefficients in the mapping relation formula (3) and recalculate the error function until the current control point error is less than or equal to the set threshold, then obtain the color difference correction mapping function and output the mapping relation. If the current control point error is less than or equal to the set threshold, then directly obtain the color difference correction mapping function and output the current mapping relation.

[0011] Optimized, before iterating based on the constructed error function in step S133, first perform initial value estimation and convergence optimization according to formula (5): (5); Where: Denotes the particle velocity at the - th iteration, Denotes the particle velocity at the - th iteration, Denotes the inertia weight, Denotes the first learning factor, Denotes the first random number between 0 and 1, Denotes the historical optimal position of the current particle, Denotes the particle position at the - th iteration, Denotes the particle position at the - th iteration, Denotes the second learning factor, Denotes the second random number between 0 and 1, Denotes the historical optimal position of the particles in the whole group.

[0012] Furthermore, during the error function iteration in step S133, accelerate the registration index according to the following method: S1331: During the iteration, based on the reserved control points, construct a KD - tree index structure. Store a data point at each node of the KD - tree index structure binary tree, and each time divide the data points in one dimension. Divide the data points whose values are less than or equal to the value of the data point in this dimension into the left subtree, and divide the data points whose values are greater than the value of the data point in this dimension into the right subtree; S1332: When performing the nearest neighbor search, first compare the data value of the current node with its value in the current dimension. If the value of the current data point is less than or equal to its value in the current dimension, then search in the left subtree. If the value of the current data point is greater than its value in the current dimension, then search in the right subtree, and calculate the distance between the current node and the query point, so as to find the nearest neighbor node; S1333: Recursively search until reaching the leaf nodes. When backtracking, update the nearest neighbor nodes, then search in another subtree and determine whether there are nearest neighbor nodes. If there are no nearest neighbor nodes in the other subtree, save the nearest neighbor nodes updated during backtracking. If there are nearest neighbor nodes in the other subtree, update the nearest neighbor nodes.

[0013] A single-molecule localization super-resolution imaging chromatic aberration correction system for performing any of the single-molecule localization super-resolution imaging chromatic aberration correction methods described above, which includes an excitation light source module, an excitation light path regulation module, a sample, a fluorescence collection and filtering module, a light detection module, and a data acquisition and processing module; The excitation light source module is used to excite the sample with excitation lights of different wavelengths; The sample includes a lateral chromatic aberration correction sample and an axial chromatic aberration correction sample; The excitation light path regulation module is used to laterally move the target surface or axially move the target surface, thereby obtaining multiple groups of laterally moved images and multiple groups of axially moved images in two different wavelength channels; The fluorescence collection and filtering module is used to collect and filter the fluorescence; The light detection module is used to detect the fluorescence; The data acquisition and processing module is used to collect the fluorescence detection situation, multiple groups of laterally moved images and multiple groups of axially moved images and perform data processing to achieve single-molecule localization super-resolution imaging chromatic aberration correction.

[0014] Advantages of the invention: A single-molecule localization super-resolution imaging chromatic aberration correction method and system provided by the present invention have the following advantages: By sequentially exciting the lateral chromatic aberration correction sample with excitation lights of two different wavelengths, collecting the image data of two wavelength channels at different lateral positions, and then performing point registration by combining the local weighted average and the iterative closest point method, the correction of lateral chromatic aberration in multi-color single-molecule localization super-resolution imaging is realized. By sequentially exciting the axial chromatic aberration correction sample with excitation lights of two different wavelengths, collecting the image data of two wavelength channels at different axial positions, and using Gaussian fitting and least squares fitting to obtain the average deviation of the two channels at the axial height and perform compensation, the correction of axial chromatic aberration in multi-color single-molecule localization super-resolution imaging is realized, improving the registration accuracy of the system, and the correction cost is relatively low and the correction speed is relatively fast. Brief description of the drawings

[0015] Figure 1 It is a schematic diagram of the lateral chromatic aberration correction process of the present invention.

[0016] Figure 2 It is a schematic diagram of the axial chromatic aberration correction process of the present invention.

[0017] Figure 3 It is a Gaussian fitting peak position diagram of the axial height of the two channels of the present invention.

[0018] Figure 4 It is a least square fitting curve diagram of the axial positions of the two channels of the present invention.

[0019] Figure 5 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0020] A method for chromatic aberration correction of single-molecule localization super-resolution imaging, the schematic diagram of the lateral chromatic aberration correction process is as follows: Figure 1 As shown, the schematic diagram of the axial chromatic aberration correction process is as follows Figure 2 shown.

[0021] Lateral chromatic aberration correction includes the following steps: S11: using two different wavelengths of excitation light to sequentially excite the lateral chromatic aberration correction sample on the glass slide, and obtaining pixel information and lateral coordinate information of multiple groups of lateral movement images of two different wavelength channels; Specifically, a plurality of fluorescent microspheres may be selected as the sample for lateral chromatic aberration correction.

[0022] The two excitation lights of different wavelengths may preferably be: an excitation light of 647 nanometers and an excitation light of 561 nanometers.

[0023] When obtaining multiple sets of lateral motion images of two different wavelength channels, the exposure time of each frame of the image can be preferably 100 milliseconds, and the field of view acquisition range of a single image is Square micrometers.

[0024] The target surface can be laterally moved by using a two-dimensional translation stage, thereby obtaining multiple groups of lateral movement images of two different wavelength channels.

[0025] S12: determining the positioning accuracy of the lateral chromatic aberration correction sample in each wavelength channel according to the pixel information of the multiple groups of lateral movement images of the two different wavelength channels, removing the points whose positioning accuracy of the lateral chromatic aberration correction sample in each wavelength channel is greater than a set threshold, and removing the misaligned points between different wavelength channels; Since fluorescent microspheres may be bleached, aggregated, or blocked in actual imaging, mismatched point pairs must be eliminated first.

[0026] Specifically, when the points in each wavelength channel where the positioning accuracy of the lateral chromatic aberration correction sample is greater than the set threshold are eliminated, Based on the pixel information of multiple groups of lateral movement images of two different wavelength channels, the sample positioning accuracy of the lateral chromatic aberration correction of the two different wavelength channels can be calculated according to formula (1): (1); Wherein: represents the lateral chromatic aberration correction sample positioning accuracy of two different wavelength channels, represents the expected value of the Gaussian distribution, represents the number of photons in the light spot, represents the pixel information of the laterally shifted images of two different wavelength channels, represents the background noise.

[0027] After calculating the lateral chromatic aberration correction sample positioning accuracy of two different wavelength channels, compare it with the set threshold, and remove the points where the lateral chromatic aberration correction sample positioning accuracy in each wavelength channel is greater than the set threshold, while retaining the points where the lateral chromatic aberration correction sample positioning accuracy in each wavelength channel is less than or equal to the set threshold.

[0028] The set threshold here can be preferably 5 nanometers, so that more than 90% of the point pairs can be retained.

[0029] Specifically, the steps of removing the misregistration points between different wavelength channels by using the DBSCAN clustering algorithm based on distance adaptive change include the following: S121: Take the positioning points of all fluorescent microspheres in two different wavelength channels as control point pairs, and take the distance between the control point pairs formed by the fluorescent microspheres at the outermost edge positions in the two different wavelength channels as the search radius. With the center position of the image as the center of the circle, calculate the scale factor according to Equation (2): (2); Wherein: represents the search radius, represents the scale factor, represents the distance from the current control point to the center of the image, represents the initial search radius; Initial search radius The selection of the initial search radius is related to the set threshold of the positioning accuracy. In the system we are using currently, the drift caused by consecutive two-frame acquisitions will not exceed 5 nanometers, and the set threshold of the positioning accuracy is limited within the range of 5 nanometers. The positioning range between the points with the same position at the center of the target surface in different channels will not exceed 25 nanometers. Therefore, the initial search radius can be preferably 25 nanometers.

[0030] The selection of the scale factor is related to the lateral chromatic aberration of the system itself, and the lateral chromatic aberration is larger at positions farther from the center of the target surface. Therefore, by collecting the fluorescent microspheres at the outermost edge position of the target surface and performing separate positioning in two channels, taking the distance between this control point pair as the threshold, under the target surface with a size of square micrometers, this value is about 80 nanometers, and set this value as the search radius upper limit, and thus the scale factor can be calculated according to Equation (2). .

[0031] S122: Traverse all the control points in one of the wavelength channels, calculate the search radius according to Equation (2), and then, with this control point as the center, check whether there are registration points in the other wavelength channel within the calculated search radius. If there are no registration points in the other wavelength channel, consider this control point as a noise point and eliminate it. If there are multiple registration points in the other wavelength channel, eliminate this control point and the multiple registration points. If there is a unique registration point in the other wavelength channel, retain this control point and the registration point as a control point pair.

[0032] The above method makes full use of the physical property that the lateral chromatic aberration increases with the centrifugal distance, pre-eliminates the mismatched point pairs, thereby reducing the subsequent workload and improving the accuracy of point pair registration, which is conducive to improving the chromatic aberration correction accuracy of the system.

[0033] S13: Perform registration between the points retained in the two different wavelength channels to obtain the chromatic aberration correction mapping function, and apply the chromatic aberration correction mapping function to complete the lateral chromatic aberration correction of single molecule localization super-resolution imaging; Specifically, the chromatic aberration correction mapping function can be obtained by the following method: S131: Use a second-order polynomial to establish the mapping relationship between the reference channel and the channel to be registered as Equation (3): (3); Where: represents the coordinate of the control point in the axis direction, represents the coordinate of the control point in the axis direction, represents the mapping function between the reference channel and the channel to be registered, represents the set of control points in the reference channel, represents the set of control points in the channel to be registered, represents axis direction mapping function, that is at the axis direction derivative, represents axis direction mapping function, that is at the axis direction derivative, represents axis direction mapping function of the first-order term coefficient of, represents axis direction mapping function of the first-order term coefficient of, Represents The quadratic term coefficient of the axial direction mapping function, Represents The quadratic term coefficient of the axial direction mapping function, Represents The cross-term coefficient of the axial direction mapping function, Represents The constant term coefficient of the axial direction mapping function, Represents The linear term coefficient of the axial direction mapping function, Represents The linear term coefficient of the axial direction mapping function, Represents The quadratic term coefficient of the axial direction mapping function, Represents The quadratic term coefficient of the axial direction mapping function, Represents The cross-term coefficient of the axial direction mapping function, Represents The constant term coefficient of the axial direction mapping function; S132: Based on the reserved control point pairs, for any control point in the reference channel, find the closest control point in the channel to be registered, calculate the sum of the squares of the distances between the two control points, and calculate the error function according to Equation (4) based on the sum of the squares of the distances between the two control points: (4); Where: Represents the error function, Represents the influence weight of the th control point, Represents the sum of the squares of the distances between the two control points, Represents the Euclidean distance between the th control point and the registration point, ; Where: Represents the control point in the channel to be registered that is closest to the th control point in the reference channel, Represents the norm calculation, Indicates the th control point in the reference channel, represents the coordinate index of the control point in the channel to be registered, represents the function of taking the minimum value.

[0034] The sum of the squares of the distances between two control points can be calculated according to the following formula , The meaning of ; In the scenario of correcting lateral chromatic aberration, the farther the point pair is from the center, the greater the influence of lateral chromatic aberration. Therefore, the correction of this part of the point pair is the focus during registration, and the error function is calculated according to Equation (4) based on the sum of the squares of the distances between two control points.

[0035] The expected minimum target distance here can be set to 5 nanometers.

[0036] represents the penalty for the uncertainty between the registered point pairs, ensuring that the registered distance of the transformed point pairs is as close as possible to the set final registration distance, while increasing the stability of the transformation.

[0037] Indicates the influence weight of the th control point, which can be defined as the following formula: where: represents the normalized distance, ; where: represents the coordinate of the center of the target surface in the axis direction, represents the coordinate of the center of the target surface in the axis direction, Indicates the th control point in the axis direction coordinate, Indicates the th control point in the axis direction coordinate, Indicates the distance between the th control point in the reference channel and the nearest neighbor among all control points, Indicates the th control point to the distance from the center of the target surface.

[0038] S133: Iterate based on the constructed error function, and during the iteration process, compare the relationship between the current control point error and the set threshold. If the current control point error is greater than the set threshold, then change the coefficients in the mapping relation formula (3) and recalculate the error function until the current control point error is less than or equal to the set threshold, then obtain the color difference correction mapping function and output the mapping relation. If the current control point error is less than or equal to the set threshold, directly obtain the color difference correction mapping function and output the current mapping relation.

[0039] The change of the coefficients in the mapping relation formula (3) here includes 、 、 、 、 、 、 、 、 、 、 、 。

[0040] The above point registration method combines local weighted average (LWM) and iterative closest point (ICP) algorithms, and is more suitable for non-rigid transverse color difference transformation.

[0041] Optimally, before iterating based on the constructed error function in step S133, first perform initial value estimation and convergence optimization according to formula (5): (5); Where: represents the particle velocity at the -th iteration, represents the particle velocity at the -th iteration, represents the inertia weight, represents the first learning factor, represents the first random number between 0 and 1, represents the historical optimal position of the current particle, that is, the individual optimal position, represents the particle position at the -th iteration, represents the particle position at the -th iteration, represents the second learning factor, represents the second random number between 0 and 1, represents the historical optimal position of the particles in the whole group, that is, the global optimal position.

[0042] Optimizing the initial value through the above particle swarm-based optimization algorithm can avoid falling into local optima.

[0043] Furthermore, during the iterative process of the error function in step S133, the registration index is accelerated according to the following method: S1331: During the iterative process, based on the reserved control point pairs, construct a KD-tree index structure. Each node on the binary tree of the KD-tree index structure stores a data point, and each time the data points are divided in one dimension. The data points with values less than or equal to the value of the data point in this dimension are divided into the left subtree, and the data points with values greater than the value of the data point in this dimension are divided into the right subtree; S1332: When performing the nearest neighbor search, first compare the data value of the current node with its value in the current dimension. If the value of the current data point is less than or equal to its value in the current dimension, search in the left subtree; if the value of the current data point is greater than its value in the current dimension, search in the right subtree, and calculate the distance between the current node and the query point to find the nearest neighbor node; S1333: Recursively search until reaching the leaf node. When backtracking, update the nearest neighbor node, then search in the other subtree, and determine whether there is a nearest neighbor node. If there is no nearest neighbor node in the other subtree, save the nearest neighbor node updated during backtracking; if there is a nearest neighbor node in the other subtree, update the nearest neighbor node.

[0044] By adopting the above method of constructing a KD-tree index structure during iteration, the complexity of finding the nearest neighbor can be reduced, thereby improving the calculation efficiency.

[0045] Longitudinal chromatic aberration correction includes the following steps: S21: Place the axial chromatic aberration correction sample on a glass slide treated with polylysine, sequentially excite the axial chromatic aberration correction sample with two different wavelengths of excitation light, and obtain the pixel information and axial height data of multiple groups of axial movement images in two different wavelength channels; Specifically, the axial chromatic aberration correction sample can preferably be a dual-color fluorescent secondary antibody.

[0046] The two different wavelengths of excitation light can preferably be: excitation light with a wavelength of 647 nm and excitation light with a wavelength of 561 nm.

[0047] When obtaining multiple groups of axial movement images in two different wavelength channels, the exposure time of each frame of image can preferably be 10 milliseconds, and the acquisition range of a single image field of view can preferably be square micrometers.

[0048] S22: Respectively perform Gaussian fitting on the distribution of multiple groups of axial height data in each wavelength channel to obtain the axial height data after fitting for the two channels. Specifically, the Gaussian fitting peak position diagrams of the axial heights of the two channels are as Figure 3As shown, in the figure, 1 represents the graph of the peak position of the Gaussian fitting of the axial height corresponding to the excitation light of 561 nm, and the axial height corresponding to the Gaussian fitting peak is 203 nm.

[0049] 2 represents the graph of the peak position of the Gaussian fitting of the axial height corresponding to the excitation light of 647 nm, and the axial height corresponding to the Gaussian fitting peak is 304 nm.

[0050] S23: Based on the axial height data after fitting of the two channels, use linear least squares fitting to establish the relationship formula of the axial chromatic aberration between the channels, so as to obtain the average deviation between the two channels at different axial heights; specifically, the least squares fitting curve graph of the axial positions of the two channels is as Figure 4 shown, and the unit nm in the figure is nanometer.

[0051] The relationship formula of the axial chromatic aberration between the channels established here is as follows: ; Among them: represents the axial height of the reference channel, represents the axial height of the channel to be registered, represents the average deviation between the two channels at different axial heights.

[0052] S24: Compensate the average deviation between the two channels at different axial heights into the single-molecule localization super-resolution imaging chromatic aberration correction system to complete the axial chromatic aberration correction of the single-molecule localization super-resolution imaging.

[0053] A single-molecule localization super-resolution imaging chromatic aberration correction system for performing the single-molecule localization super-resolution imaging chromatic aberration correction method as described in any one of the above, and its system structure schematic diagram is as Figure 5 shown, including an excitation light source module, an excitation light path regulation module, a sample, a fluorescence collection and filtering module, a light detection module, and a data acquisition and processing module; The excitation light source module is used to excite the sample with excitation light of different wavelengths; The sample includes a lateral chromatic aberration correction sample and an axial chromatic aberration correction sample; The excitation light path regulation module is used to laterally move the target surface or axially move the target surface, so as to obtain multiple groups of laterally moved images and multiple groups of axially moved images of two different wavelength channels; The fluorescence collection and filtering module is used to collect and filter the fluorescence; The light detection module is used to detect the fluorescence; The data acquisition and processing module is used to collect the fluorescence detection situation, multiple groups of laterally moved images and multiple groups of axially moved images and perform data processing to realize the chromatic aberration correction of the single-molecule localization super-resolution imaging.

[0054] Specifically, the excitation light path control module includes a two-dimensional translation stage and a microscope axial adjustment device. The two-dimensional translation stage is used to laterally move the target surface, and the microscope axial adjustment device is used to axially move the target surface.

[0055] In summary, the single-molecule localization super-resolution imaging chromatic aberration correction method and system provided by the present invention perform point registration by combining the point registration method of local weighted average and iterative closest point, realizing the correction of lateral chromatic aberration in multi-color single-molecule localization super-resolution imaging. By using Gaussian fitting and least squares fitting, the average deviation at the axial height of the two channels is obtained and compensated, realizing the correction of axial chromatic aberration in multi-color single-molecule localization super-resolution imaging, thereby improving the registration accuracy of the system, reducing the cost of correction, and increasing the correction speed.

[0056] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for chromatic aberration correction of single-molecule localization super-resolution imaging, characterized in that: Lateral chromatic aberration correction includes the following steps: S11: using two different wavelengths of excitation light to sequentially excite the lateral chromatic aberration correction sample on the glass slide, and obtaining pixel information and lateral coordinate information of multiple groups of lateral movement images of two different wavelength channels; S12: determining the positioning accuracy of the lateral chromatic aberration correction sample in each wavelength channel according to the pixel information of the multiple groups of lateral movement images of the two different wavelength channels, removing the points whose positioning accuracy of the lateral chromatic aberration correction sample in each wavelength channel is greater than a set threshold, and removing the misaligned points between different wavelength channels; S13: aligning the points retained in the two different wavelength channels to obtain a chromatic aberration correction mapping function, and applying the chromatic aberration correction mapping function to complete the lateral chromatic aberration correction of single-molecule localization super-resolution imaging; Longitudinal chromatic aberration correction includes the following steps: S21: placing the axial chromatic aberration correction sample on a glass slide treated with poly-lysine, sequentially exciting the axial chromatic aberration correction sample with two different wavelengths of excitation light, and acquiring pixel information and axial height data of multiple groups of axial movement images of two different wavelength channels; S22: performing Gaussian fitting on the distribution of multiple groups of axial height data of each wavelength channel respectively to obtain fitted axial height data of two channels; S23: Based on the axial height data of the two channels after fitting, a linear least squares fitting is used to establish a relationship between the axial chromatic aberrations of the channels, thereby obtaining the average deviations of the two channels at different axial heights; S24: Compensate the average deviation of the two channels at different axial heights into the single-molecule localization super-resolution imaging chromatic aberration correction system to complete the single-molecule localization super-resolution imaging axial chromatic aberration correction.

2. The single-molecule localization super-resolution imaging chromatic aberration correction method according to claim 1, wherein: The lateral chromatic aberration correction sample in step S11 is a plurality of fluorescent microspheres, and the axial chromatic aberration correction sample in step S21 is a dual-color fluorescent secondary antibody.

3. A method for chromatic aberration correction in single molecule localization super-resolution imaging according to claim 1, characterized in that: The two excitation lights of different wavelengths in step S11 and step S21 are respectively: an excitation light of wavelength 647 nanometers and an excitation light of wavelength 561 nanometers.

4. A chromatic aberration correction method for single molecule localization super-resolution imaging according to claim 1, characterized in that: When obtaining multiple sets of laterally moving images in two different wavelength channels in step S11, the exposure time of each frame of image is 100 milliseconds, and the acquisition range of a single image field of view is square micrometers. When obtaining multiple sets of axially moving images in two different wavelength channels in step S21, the exposure time of each frame of image is 10 milliseconds, and the acquisition range of a single image field of view is square micrometers.

5. A method for chromatic aberration correction in single molecule localization super-resolution imaging according to claim 2, characterized in that: In step S12, based on the pixel information of the multiple groups of lateral movement images of the two different wavelength channels, the positioning accuracy of the lateral chromatic aberration correction samples of the two different wavelength channels is calculated according to formula (1): （1）； Wherein: represents the lateral chromatic aberration correction sample positioning accuracy of two different wavelength channels, represents the expected value of the Gaussian distribution, represents the number of photons in the light spot, represents the pixel information of the laterally shifted images of two different wavelength channels, represents the background noise.

6. The method for chromatic aberration correction in single molecule localization super-resolution imaging according to claim 2, wherein: In step S12, the DBSCAN clustering algorithm based on adaptive distance change is used to remove misaligned points between different wavelength channels, which specifically includes the following steps: S121: The positioning points of all fluorescent microspheres in two different wavelength channels are taken as control point pairs, and the distance between the control point pairs formed by the fluorescent microspheres at the most edge positions in two different wavelength channels is taken as the search radius. The center of the image is taken as the center of the circle, and the scale factor is calculated according to formula (2): （2）； Wherein: represents the search radius, represents the scale factor, represents the distance from the current control point to the center of the image, represents the initial search radius; S122: Traverse all control points in one of the wavelength channels, calculate the search radius according to Equation (2), then, with this control point as the center, check whether there are registration points in the other wavelength channel within the calculated search radius. If there are no registration points in the other wavelength channel, consider this control point as a noise point and eliminate it. If there are multiple registration points in the other wavelength channel, eliminate this control point and the multiple registration points. If there is a unique registration point in the other wavelength channel, retain this control point and the registration point as a control point pair.

7. A method for chromatic aberration correction in single molecule localization super-resolution imaging according to claim 6, characterized in that: In step S13, the following method is used to obtain the color difference correction mapping function: S131: Use a second-order polynomial to establish the mapping relationship between the reference channel and the channel to be registered as Equation (3): （3）； Wherein: represents the coordinate of the control point in the axis direction, represents the coordinate of the control point in the axis direction, represents the mapping function between the reference channel and the channel to be registered, represents the set of control points of the reference channel, represents the set of control points of the channel to be registered, represents the axis direction mapping function, represents the axis direction mapping function, represents the coefficient of the first-order term of the axis direction mapping function of, represents the coefficient of the first-order term of the axis direction mapping function of, represents the coefficient of the second-order term of the axis direction mapping function of, represents the coefficient of the second-order term of the axis direction mapping function of, represents the cross-term coefficient of the axis direction mapping function, represents the constant-term coefficient of the axis direction mapping function, represents the coefficient of the first-order term of the axis direction mapping function of, represents the coefficient of the first-order term of the axis direction mapping function of, represents the coefficient of the second-order term of the axis direction mapping function of, represents the coefficient of the second-order term of the axis direction mapping function of, represents the cross-term coefficient of the axis direction mapping function, represents the constant-term coefficient of the axis direction mapping function; S132: Based on the retained control point pairs, for any control point in the reference channel, find the control point closest to it in the channel to be registered, calculate the sum of the squares of the distances between the two control points, and calculate the error function according to Equation (4) based on the sum of the squares of the distances between the two control points: （4）； Wherein: represents the error function, represents the influence weight of the th control point, represents the sum of squares of the distances between two control points, represents the Euclidean distance between the th control point and the registration point, represents the desired minimum target distance, represents the number of control points; S133: Iterate based on the constructed error function, and during the iteration process, compare the relationship between the current control point error and the set threshold. If the current control point error is greater than the set threshold, change the coefficients in the mapping relationship Equation (3) and recalculate the error function until the current control point error is less than or equal to the set threshold, then obtain the color difference correction mapping function and output the mapping relationship. If the current control point error is less than or equal to the set threshold, directly obtain the color difference correction mapping function and output the current mapping relationship.

8. A method for chromatic aberration correction in single molecule localization super-resolution imaging according to claim 7, characterized in that: Before iterating based on the constructed error function in step S133, first perform initial value estimation and convergence optimization according to Equation (5): （5）； Wherein: represents the particle velocity at the -th iteration, represents the particle velocity at the -th iteration, represents the inertia weight, represents the first learning factor, represents the first random number between 0 and 1, represents the historical optimal position of the current particle, represents the particle position at the -th iteration, represents the particle position at the -th iteration, represents the second learning factor, represents the second random number between 0 and 1, represents the historical optimal position of the particles in the whole swarm.

9. A method for chromatic aberration correction in single-molecule localization super-resolution imaging according to claim 7, characterized in that: During the error function iteration process in step S133, accelerate the registration index according to the following method: S1331: During the iteration process, construct a KD-tree index structure based on the retained control point pairs. Store a data point at each node of the binary tree of the KD-tree index structure, and divide the data points in one dimension each time. Divide the data points with values less than or equal to the value of the data point in this dimension into the left subtree, and divide the data points with values greater than the value of the data point in this dimension into the right subtree; S1332: When performing the nearest neighbor search, first compare the data value of the current node with its value in the current dimension. If the value of the current data point is less than or equal to its value in the current dimension, search in the left subtree. If the value of the current data point is greater than its value in the current dimension, search in the right subtree, and calculate the distance between the current node and the query point to find the nearest neighbor node; S1333: Recursively search until reaching the leaf node. When backtracking, update the nearest neighbor node, then search in the other subtree, and determine whether there is a nearest neighbor node. If there is no nearest neighbor node in the other subtree, save the nearest neighbor node updated during backtracking. If there is a nearest neighbor node in the other subtree, update the nearest neighbor node.

10. A single-molecule localization super-resolution imaging chromatic aberration correction system for performing a single-molecule localization super-resolution imaging chromatic aberration correction method according to any one of claims 1 to 9, characterized in that: It includes an excitation light source module, an excitation light path regulation module, a sample, a fluorescence collection and filtering module, a detection light module, and a data acquisition and processing module; The excitation light source module is used to excite the sample with excitation light of different wavelengths; The sample includes a lateral chromatic aberration correction sample and an axial chromatic aberration correction sample; The excitation light path regulation module is used to laterally move the target surface or axially move the target surface, so as to obtain multiple groups of laterally moved images and multiple groups of axially moved images of two different wavelength channels; The fluorescence collection and filtering module is used to collect and filter the fluorescence; The light detection module is used to detect the fluorescence; The data acquisition and processing module is used to collect the fluorescence detection situation, multiple groups of laterally moved images and multiple groups of axially moved images and perform data processing to achieve chromatic aberration correction for single molecule localization super-resolution imaging.

Images