Spherical fluorescent microplastic counting method based on virtual digital slice scanning system and image recognition and application thereof

Abstract

The application discloses a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition, and relates to the technical field of microplastic research. The application discloses a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition, and relates to the technical field of microplastic research. The application discloses a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition, and relates to the technical field of microplastic research.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of fluorescent microplastic counting, and particularly relates to a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition and application thereof in preparation of a fluorescent microplastic standard solution. BACKGROUND

[0002] With the development of social economy, scientific progress and the improvement of people's living standards, people's drinking water safety awareness is also constantly improving. Microplastics are plastic particles with a three-dimensional size of 0.001-5mm, which are one of the emerging pollutants found in drinking water in recent years and have attracted widespread attention from academia and the public.

[0003] In a series of studies on microplastics in drinking water, the application of fluorescent microplastic model particles has become very common. Due to the constraints of manufacturing processes, the fluorescent microplastic model particles used by researchers at present are mostly spherical particles. In existing research, the detection methods for the abundance of fluorescent microplastics in water samples include flow cytometry, microscopic naked-eye counting and microscopic image recognition.

[0004] Flow cytometry is a device for automatic analysis and sorting of cells, and has been used for quantitative analysis of fluorescent microplastics in water samples in recent years. However, flow cytometry has certain requirements for the abundance of fluorescent microplastics in samples, generally requiring the abundance of fluorescent microplastics in water samples to be 105-107 / mL, which is much higher than the occurrence abundance of microplastics in drinking water (2-30 / mL) and the designed abundance of microplastics in drinking water-microplastic research (1-100 / mL). The above reasons lead to the need for large-volume sampling and tedious concentration operation for flow cytometry to quantitatively analyze fluorescent microplastics in drinking water samples. However, small-volume sampling methods are used in drinking water-related experiments (such as beaker experiments) carried out in laboratories, and the concentration process is also easy to cause loss and damage of samples, so flow cytometry is not suitable for quantitative detection of low-concentration and small-volume fluorescent microplastic samples.

[0005] Due to the low concentration and small sampling volume of the fluorescent microplastic samples in drinking water, the microscope-image recognition method is generally preferred in the research. Both the microscope-naked eye counting method and the microscope-image recognition method need to filter the microplastics in the water sample onto a filter membrane, and then use a microscope to observe the filter membrane. The microscope-naked eye counting method directly observes, manually judges and counts using the naked eye, which is low in efficiency and high in error rate, and is basically not used in the present research. The more common method is the microscope-image recognition method, which can be divided into the microscope-fluorescence intensity method and the microscope-connected area method according to different image processing logic. The microscope-fluorescence intensity method uses a microscope to take pictures of the fluorescent microplastics on the filter membrane, and the relative quantity of the fluorescent microplastics is represented by the different fluorescence intensities of the pictures. The microscope-fluorescence intensity method has high requirements for the stability, consistency of the microplastic fluorescence intensity and the consistency of the image shooting parameters, and cannot avoid the large error caused by the fluorescent impurities. The microscope-connected area method also uses a microscope to collect images of the fluorescent microplastics on the filter membrane, and uses Image J to identify and count the connected areas of the foreground pixels in the image. The microscope-connected area method cannot separate the agglomerated fluorescent microplastics, and also cannot avoid the interference of the fluorescent impurities. In addition, the existing microscope-image recognition method uses a microscope without a large field scanning function, which limits the shooting mode to random / zone random shooting. Such shooting mode will be seriously affected by the uneven distribution of microplastics on the filter membrane surface, thereby causing a large error. Therefore, the existing microscope-image recognition method cannot achieve high accuracy and high stability counting of the fluorescent microplastic standard spherical particles.

[0006] The OLYMPUS VS120 virtual digital section scanning system is a virtual section scanning large instrument applied to the research fields of histopathology, neurobiology and developmental biology. The VS120 uses a full-automatic microscope scanning system, combines with a virtual section software, scans and seamlessly splices the traditional sections to generate a whole field digital section, so as to resolve the contradiction between high resolution and large field.

[0007] The fluorescent microplastic solution after ultrasonic, magnetic stirring and vortex oscillation still has the problem of uneven distribution of microplastics, so the traditional solubility drug standard solution preparation method is not suitable for the preparation of the fluorescent microplastic standard solution. Since the existing counting method has a large error in counting the absolute quantity of the fluorescent microplastics in the sample solution, the preparation problem of the fluorescent microplastic standard solution has not been solved. SUMMARY

[0008] The application provides a spherical fluorescent microplastic counting method based on a virtual digital section scanning system and image recognition, which can quickly and accurately determine the abundance of the fluorescent microplastics in the drinking water sample.

[0009] A spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition, comprising:

[0010] (1) using a virtual digital slice scanning system VS120 to perform full-area scanning on a spherical fluorescent microplastic sample to obtain a full-area digital image, and performing lossless cutting on the full-area digital image to obtain multiple to-be-processed images;

[0011] (2) converting the to-be-processed images into gradient information grayscale images, performing binarization processing on the gradient information grayscale images to obtain binarized edge images, identifying the coordinates of the centers of circles from the binarized edge images based on a set radius range by using a Hough gradient method, and the radius and intensity of the circles corresponding to the coordinates of the centers of the circles; and arranging the identified multiple coordinates of the centers of the circles in ascending order of the horizontal coordinates to obtain a first sequence of coordinates of the centers of the circles;

[0012] (3) the step of removing overlapping or covering circles is: calculating the difference between the horizontal coordinates of the first coordinate of the center of the circle in the first sequence of coordinates of the centers of the circles and the second coordinate of the center of the circle to obtain an absolute value of the horizontal coordinate difference, when the absolute value of the horizontal coordinate difference is greater than the maximum radius of the first coordinate of the center of the circle and the second coordinate of the center of the circle, the first coordinate of the center of the circle and the second coordinate of the center of the circle are retained, when the absolute value of the horizontal coordinate difference is less than or equal to the maximum radius, the distance between the centers of the first coordinate of the center of the circle and the second coordinate of the center of the circle is obtained first, then when the distance between the centers is greater than the maximum radius, the first coordinate of the center of the circle and the second coordinate of the center of the circle are retained, and when the distance between the centers is less than or equal to the maximum radius, the intensity of the circle corresponding to the first coordinate of the center of the circle and the second coordinate of the center of the circle is compared, and the coordinate of the center of the circle with the maximum intensity is retained;

[0013] (4) comparing the first coordinate of the center of the circle with the remaining coordinates of the center of the circle in turn according to step (3) until the horizontal coordinate difference between the first coordinate of the center of the circle and the currently compared coordinate of the center of the circle is greater than the maximum radius, the comparison is completed, or the intensity of the circle of the first coordinate of the center of the circle is lower than the intensity of the circle of the currently compared coordinate of the center of the circle, the comparison is completed, or the first coordinate of the center of the circle is compared with the last coordinate of the center of the circle, the comparison is completed;

[0014] (5) after step (4) is completed, the first sequence of coordinates of the centers of the circles is arranged in ascending order of the horizontal coordinates, and steps (3)-(4) are performed on the first sequence of coordinates of the centers of the circles after the first coordinate of the center of the circle and the coordinates of the center of the circle removed in steps (3)-(4) are removed, until the last two coordinates of the center of the circle in the first sequence of coordinates of the centers of the circles are completed, and the comparison is stopped, to obtain a second sequence of coordinates of the centers of the circles after removing overlapping or covering impurities;

[0015] (6) arranging the circle center coordinates in the second circle center coordinate sequence in ascending order according to the corresponding circle intensity and circle radius respectively, removing the circle center coordinates corresponding to the smaller circle radius and smaller circle intensity based on a circle intensity threshold and a circle radius threshold to obtain a third circle center coordinate sequence, so as to remove the fluorescent impurities existing alone, and taking the number of the circle center coordinates in the third circle center coordinate sequence as the number of spherical fluorescent microplastics in the image to be processed.

[0016] The preparation method of the fluorescent microplastic sample is: filtering the fluorescent microplastics in the solution to be measured onto a polycarbonate filter membrane to obtain a fluorescent microplastic sample.

[0017] The full-area digital image is losslessly cut into multiple images to be processed, including: using the cut function of OlyVIA or Photoshop software to losslessly cut the full-area digital image into multiple images to be processed.

[0018] The image to be processed is converted into a gradient information gray image, including:

[0019] The image to be processed is converted into an 8-bit gray image, the Sobel edge detection operator is used to calculate the X-axis direction gradient value Gx and the Y-axis direction gradient value Gy of each pixel point in the 8-bit gray image, and the is taken as the total gradient value of each pixel point, and each pixel point with the total gradient value constructs a gradient information gray image.

[0020] The gradient information gray image is binarized to obtain a binarized edge image, including:

[0021] The global image threshold of the gradient information gray image is calculated using the law of the large numbers, and the binarized edge image is obtained by binarizing the gradient information gray image using the global image threshold.

[0022] The gradient information gray image is binarized, including: the pixel point value is set to 1 when the total gradient value is greater than or equal to the global image threshold, and the pixel point value is set to 0 when the total gradient value is less than the global image threshold, thereby obtaining the binarized edge image.

[0023] The initial circle center coordinates are identified from the binarized edge image based on the set radius range using the Hough gradient method, as well as the circle radius and the circle intensity of the initial circle center coordinates, including:

[0024] The circular edges are identified from the binarized edge image based on the set radius range, the gradient direction of the circular edge pixel points and the set radius range are obtained to obtain multiple candidate circle center coordinates, the candidate circle center coordinates are voted, the candidate circle center coordinates with the highest total vote value are taken as the circle center coordinates, and the voting result of the circle center coordinates is taken as the circle intensity.

[0025] The pixel point of the circle center coordinate voting and the multiple radius values corresponding to the pixel point are obtained, the multiple radius values are voted, and the radius value with the highest total vote value is taken as the radius of the circle center coordinate.

[0026] The multiple candidate circle center coordinates are voted and the multiple radius values are voted, and the single vote value p0 of the voting is:

[0027]

[0028] R is a set radius value passing through the pixel point.

[0029] An application of a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition to the preparation of a fluorescent microplastic standard solution, comprising:

[0030] (1) The microplastic dispersion liquid is subjected to suction filtration, the fluorescent microplastics are enriched on a black polycarbonate filter membrane, and the number of spherical fluorescent microplastics on the filter membrane is obtained by the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition according to any one of claims 1-8;

[0031] (2) The black polycarbonate filter membrane enrichment surface is washed, the eluent is retained, the filter membrane after elution is counted by the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition according to any one of claims 1-8, and the number of spherical fluorescent microplastics on the filter membrane that are not eluted is obtained;

[0032] (3) The number of spherical fluorescent microplastics obtained in step (1) and the number of spherical fluorescent microplastics after elution obtained in step (2) are subtracted to obtain the number of spherical fluorescent microplastics in the elution solution, the number of spherical fluorescent microplastics in the elution solution is compared with the volume of the elution solution to obtain the abundance of the fluorescent microplastics in the elution solution, thereby completing the preparation of the fluorescent microplastic standard solution, and the eluent can also be diluted into a fluorescent microplastic standard solution with lower abundance.

[0033] Compared with the prior art, the beneficial effects of the present application are:

[0034] (1) The present application realizes full-area photographing of the filter membrane through the virtual digital slice scanning system VS120, and avoids random errors caused by the uneven distribution of fluorescent microplastics on the surface of the filter membrane due to random photographing.

[0035] (2) The application proposes to use shape discrimination algorithm, to use Hough transform to preliminarily identify the spherical edge in the image, and then to increase the algorithm to remove overlapping / covered circles. The application can avoid generating a large amount of calculation process, and also realize satisfactory processing effect and high processing speed. The addition of shape discrimination algorithm, overlapping / covered circle removal algorithm and small radius and low intensity circle removal algorithm makes the algorithm involved in the application not sensitive to R1 (the minimum radius of the target recognition circle) and R2 (the maximum radius of the target recognition circle) parameters when processing images, so that a large number of images can be processed, and satisfactory accuracy and stability can be ensured.

[0036] (3) The application can assist in preparing a microplastic standard solution with a known concentration. The application counts the number of fluorescent microplastics enriched on the black polycarbonate filter film LM1 to obtain the number N1 of fluorescent microplastics enriched thereon, uses a syringe to flush the enrichment surface of the filter film to flush the microplastics into the eluent to obtain the filter film LM2 after elution. Counting operation is performed on LM2 to obtain the number N2 of residual fluorescent microplastics thereon. N=N1-N2 is the number of fluorescent microplastics in the eluent. The application avoids the problem of difficult preparation of a standard solution caused by uneven distribution of fluorescent microplastics in an aqueous phase, and realizes the preparation of a fluorescent microplastic standard solution with a clear abundance based on the spherical fluorescent microplastic counting method of the application. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 The flowchart of the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition provided for the specific embodiment of the application;

[0038] Figure 2 The image of the sample to be measured filtered onto a polycarbonate filter film provided for the specific embodiment of the application;

[0039] Figure 3 The full-area digital image taken by the virtual digital slice scanning system VS120 provided for the specific embodiment of the application;

[0040] Figure 4 The image to be processed obtained by lossless cutting of the full-area digital image provided for the specific embodiment of the application;

[0041] Figure 5 The edge image obtained after edge recognition of the image to be processed provided for the specific embodiment of the application;

[0042] Figure 6 The effect diagram after screening the circular edge by the Hough gradient method provided for the specific embodiment of the application;

[0043] Figure 7An effect diagram after the overlapping / covering circle removal processing provided for the specific embodiment of the present application;

[0044] Figure 8 An effect diagram after the low-brightness impurity and irregular impurity removal processing provided for the specific embodiment of the present application;

[0045] Figure 9 A marking diagram of the spherical fluorescent microplastic counting method of the virtual digital slice scanning system and image recognition for the to-be-processed image and an Image-Pro Plus counting marking diagram provided for the specific embodiment of the present application, wherein, Figure 9 (a) is a to-be-processed image, Figure 9 (b) is a marking diagram of the spherical fluorescent microplastic counting method of the virtual digital slice scanning system and image recognition for the to-be-processed image, Figure 9 (c) is an Image-Pro Plus counting marking diagram;

[0046] Figure 10 A digital image using random / partition photographing provided for the specific embodiment of the present application;

[0047] Figure 11 A fluorescent microplastic standard solution preparation flowchart provided for the specific embodiment of the present application. DETAILED DESCRIPTION

[0048] The present application will be further described below in combination with specific embodiments and the accompanying drawings.

[0049] The present application provides a spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition, as shown in the accompanying drawings, comprising: Figure 1

[0050] (1) Obtain a solution to be detected, and filter the microplastics in the solution onto a polycarbonate filter membrane with a diameter of 25 mm and a pore size of 0.8 μm.

[0051] (2) Place the filter membrane flat on a clean glass slide without fluorescent substance pollution with the to-be-detected surface of the filter membrane facing up, complete the sample preparation, as shown in the accompanying drawings. Figure 2

[0052] ​​(3) Place the sample to be tested prepared in step (2) on the motorized stage of the VS120 virtual digital slide scanning system and fix it in the slide slot. Open the CellSens software (the operating software of the VS120 virtual digital slide scanning system), select the fluorescence mode, select the corresponding slide slot, fill in the save path and slide name, select low magnification (×2), preview the scanning of the entire slide in bright field or fluorescence environment, select an appropriate objective magnification and select the area to be photographed, set an appropriate number and position of focus points, select the required fluorescence channel and exposure time, further, manually focus the selected focus points, and finally click "Scan Now" to start scanning and photographing the target area. After the photographing is completed, save the full-area digital image obtained. The full-area digital image format is VSI, see Figure 3 .

[0053] (4) Due to limitations in overall computer memory requirements and processing speed, it is necessary to first use OlyVIA software to divide the full-area digital image into four equal parts, save the divided images in JPG / JPEG / BMP format, and then use Photoshop's image slicing function to losslessly cut the four-part image into multiple images of appropriate size to be processed, such as... Figure 4 As shown.

[0054] (5) Convert the image to be processed into an 8-bit grayscale image, and use the Sobel edge detection operator to calculate the gradient value Gx in the X-axis direction and the gradient value Gy in the Y-axis direction for each pixel in the 8-bit grayscale image, and then... As the total gradient value for each pixel, each pixel with a total gradient value constructs a gradient information grayscale image. A global image threshold for the gradient information grayscale image is calculated using the big rule algorithm. Pixels with a total gradient value greater than or equal to the global image threshold are set to 1 (Logical), while pixels with a total gradient value less than the global image threshold are set to 0 (Logical), thus obtaining a binarized edge image, such as... Figure 5 As shown.

[0055] (6) The effect obtained by filtering the circular edges using the Hough gradient method is shown in the figure below. Figure 6 As shown: The Hough gradient method is used to identify the center coordinates of circles from a binary edge image based on a user-specified radius range [R1, R2], as well as the circle radius and circle intensity corresponding to the center coordinates, and to obtain the first circle center coordinate sequence. The specific steps are as follows:

[0056] The setting radius is traversed by steps of 0.5, a circular edge is identified from the binary edge image based on the set radius, a plurality of candidate center coordinates of the circle are calculated according to a gradient direction of a pixel point of the circular edge and the preset radius, voting is performed on the plurality of candidate center coordinates of the circle, the candidate center coordinate with the highest total vote value is taken as the center coordinate of the circle, and a voting result of the center coordinate of the circle is taken as a circle intensity; a pixel point that votes for the center coordinate of the circle and a plurality of radius values corresponding to the pixel point are obtained, voting is performed on the plurality of radius values, and the radius value with the highest total vote value is taken as a circle radius of the center coordinate of the circle.

[0057] In the voting performed on the plurality of candidate center coordinates of the circle and the voting performed on the plurality of radius values, a single vote value p0 of the voting is as follows:

[0058]

[0059] R is a set radius value passing through the pixel point. Through the setting of the single vote value, the circle radius with more voting times in the set radius, that is, the radius with a larger radius value, is set to have a lower single vote value, so that the radius value is larger and more votes are obtained, and fairness is achieved by reducing the single vote value.

[0060] The plurality of identified center coordinates of the circle are arranged in ascending order of the horizontal coordinates to obtain a first center coordinate sequence, and the first center coordinate sequence includes a first center coordinate, a second center coordinate,..., and an Nth center coordinate.

[0061] (7) removing overlapping or covering circles in the image to be processed, and the specific steps are as follows:

[0062] (7.1) calculating a horizontal coordinate difference absolute value by performing difference calculation on the horizontal coordinates of the first center coordinate and the second center coordinate of the first center coordinate sequence, when the horizontal coordinate difference absolute value is greater than a maximum circle radius of the first center coordinate and the second center coordinate, the first center coordinate and the second center coordinate are retained, when the horizontal coordinate difference absolute value is less than or equal to the maximum circle radius, a center distance of the first center coordinate and the second center coordinate is obtained first, then when the center distance is greater than the maximum circle radius, the first center coordinate and the second center coordinate are retained, and when the center distance is less than or equal to the maximum circle radius, a circle intensity corresponding to the first center coordinate and the second center coordinate is compared, and the center coordinate with the maximum circle intensity is retained;

[0063] (7.2) according to step (7.1) sequentially compare the first circle center coordinates with the remaining circle center coordinates until the horizontal coordinate difference between the first circle center coordinates and the currently compared circle center coordinates is greater than the maximum radius, indicating that the remaining circle center coordinates in the first circle center sequence have a horizontal coordinate difference greater than the maximum circle radius value, or the circle intensity of the first circle center coordinates is lower than the circle intensity of the currently compared circle center coordinates, indicating that the circle corresponding to the first circle center coordinates is a fluorescent impurity, thereby eliminating the first circle center coordinates to complete the comparison, or the first circle center coordinates are compared with the last circle center coordinates in the first circle center sequence to complete the comparison;

[0064] (7.3) after step (7.2) completes the comparison, according to the ascending order of the horizontal coordinates, the first circle center coordinates in the first circle center sequence are executed according to steps (7.1)-(7.2) except for the first circle center coordinates and the circle center coordinates removed by (3)-(4), until the last order circle center coordinates remaining in the first circle center sequence complete steps (7.1)-(7.2), stop comparison, to obtain a second circle center sequence that removes overlapping or covering impurities, and the processed effect diagram is as shown in Figure 7

[0065] (8) removing low-brightness and irregular impurities in the image to be processed: the circle center coordinates in the second circle center sequence are arranged in ascending order according to the corresponding circle intensity and circle radius, and the circles with a radius less than the first a% level and a circle intensity less than the first b% level are removed. The values of a and b are flexibly adjusted according to the size of the impurity content in the image to obtain a third circle center sequence, thereby removing the fluorescent impurities that exist alone, and the effect after removing the low-brightness impurities and irregular impurities is as shown in Figure 8 The number of circle center coordinates in the third circle center sequence is taken as the number of spherical fluorescent microplastics in the image to be processed.

[0066] (9) processing output data and calculating sample abundance: the fluorescent microplastic count results and the recognition effect diagram of each picture are output by the present application, so that the user can master the processing effect and adjust the parameters in time, and it is also convenient for the user to process data. The specific processing method of the data is as follows: the number of fluorescent microplastics corresponding to each small picture to be processed after the full field of view scanning picture is divided is summed, that is, the total number of fluorescent microplastics corresponding to the complete scanning picture is obtained, and finally the abundance of fluorescent microplastics in the target sample solution is calculated according to the volume of the target sample.

[0067] ​Based on this, the OLYMPUS VS120 virtual digital slice scanning system is first applied to counting of the fluorescent microplastic model particles, full-area photographing of the filter membrane is realized, and errors caused by random / partition photographing and uneven distribution of the fluorescent microplastics on the surface of the filter membrane are avoided; the shape discrimination algorithm is used to identify the spherical fluorescent microplastics in the image, and the identification software only taking the connected domain as the counting object such as Image J has higher accuracy; the algorithm for eliminating overlapping / covered circles and eliminating low-brightness fluorescent impurities is proposed, the function of batch processing of the fluorescent microplastic pictures is realized under the condition of ensuring the identification accuracy, the picture processing efficiency is greatly improved, and the time of the scientific researchers is saved.

[0068] The application further provides a method for preparing a fluorescent microplastic standard solution by elution and quantification, as shown in the following formula: Figure 11 The specific steps are as follows:

[0069] (1) The microplastic dispersion liquid is subjected to suction filtration to obtain a black polycarbonate filter membrane LM1 rich in fluorescent microplastics;

[0070] (2) The black polycarbonate filter membrane LM1 rich in fluorescent microplastics is subjected to counting operation by using the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition, the number N1 of the fluorescent microplastics rich thereon is obtained, the filter membrane rich surface is washed by using a syringe, the microplastics are washed into an eluent, and an eluted filter membrane LM2 is obtained;

[0071] (3) The LM2 is subjected to counting operation by using the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition to obtain the number N2 of the residual fluorescent microplastics thereon. 洗脱液 N1-N2, that is, the number of the fluorescent microplastics in the eluent. 洗脱液 Compared with the volume of the eluent or the volume V0 of the eluent after dilution, the abundance C0 of the fluorescent microplastics in the elution solution is obtained, and thus the preparation of the fluorescent microplastic standard solution with a known concentration is completed.

[0072] The total number (N1) of the microplastics rich on the black polycarbonate filter membrane from the microplastic dispersion liquid is 146, and the number (N2) of the fluorescent microplastics remaining on the filter membrane after elution is 9. In order to master the number of the fluorescent microplastics lost in the elution process, the eluent obtained in this embodiment is subjected to suction filtration and counting, and the result shows that the number (N3) of the fluorescent microplastics rich on the filter membrane (LM3) after suction filtration of the eluent is 135. The calculation formula of the total number N4 of the fluorescent microplastics lost in the elution and suction filtration of the eluent is as follows:

[0073] N4=N1-N2-N3

[0074] In this embodiment, N4 is 2. The formula for calculating the fluorescent microplastic loss rate η during elution and filtration of the eluent is as follows:

[0075]

[0076] In this embodiment, the fluorescent microplastic loss rate η is 1.36%, which meets the experimental accuracy requirements. Therefore, N1-N2 can be directly used as the number of fluorescent microplastics in the eluent. Furthermore, the eluent volume used in this embodiment is 75 mL. If a fluorescent microplastic standard solution with higher abundance is required, the abundance of fluorescent microplastics in the eluent can be increased by increasing the eluent flow rate and decreasing the eluent volume.

[0077] Comparative Example

[0078] Use random / regional photography to photograph the filtered samples, such as Figure 10 As shown, microplastic spillage exists on the right side, exhibiting significant non-uniformity. Five sections were selected from the original image: center, top, bottom, left, and right, numbered 1, 2, 3, 4, and 5. The number of spherical fluorescent microplastic model particles within each section was accurately determined by visual counting. The criteria for visually identifying fluorescent microplastic model particles were: particles with a diameter of approximately 10 μm and whose center was within the image. The table below shows the accurate values ​​and statistical indicators of spherical fluorescent microplastic model particles in the five sections:

[0079]

[0080] As can be seen from the data in the table above, the coefficient of variation among the five zones reached 0.20, which indicates that the distribution of fluorescent microplastics on the filter membrane is highly uneven, and also indicates that random / zone-based photography will introduce significant errors.

[0081] As for the full-area imaging achieved by the VS120 in this patent, the uniformity of the distribution of fluorescent microplastics on the filter membrane will not affect the final counting result, completely avoiding the large random errors caused by random / regional shooting.

[0082] Effect description:

[0083] The spherical fluorescent microplastic counting method based on a virtual digital slice scanning system and image recognition provided in this invention, along with the counting function of Image-Pro Plus software, were used to count particles from the same fluorescent microplastic model image. The accuracy of the results was compared. Figure 9 a- Figure 9The accurate value of the fluorescent microplastic model particles in the picture is identified by the human eye. The judgment basis and standard for identifying the fluorescent microplastic model particles by the human eye are that the fluorescent microplastic model particles in the picture have a particle size of about 10 microns and a circle center in the image.

[0084] In the present experiment, the accurate number of the fluorescent microplastic model particles in the picture is 248, as shown in FIG. 8c. Figure 1 The accurate number of the fluorescent microplastic model particles in the picture is 248, as shown in FIG. 8c. Figure 9 The recognition effect of the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition provided by the present application reaches an accuracy of 100%, while the Image-Pro Plus software shown in FIG. 8b only reaches an accuracy of 76.21%. Figure 9 b shown in FIG. 8b only reaches an accuracy of 76.21%.

Claims

1. A method for counting spherical fluorescent microplastics based on a virtual digital slice scanning system and image recognition, characterized in that, The method comprises the following steps: (1) using a virtual digital slice scanning system VS120 to take full-area digital images by full-area scanning of a spherical fluorescent microplastic sample, and cutting the full-area digital images into multiple to-be-processed images without loss; (2) converting the to-be-processed images into gradient information gray images, performing binaryzation processing on the gradient information gray images to obtain binary edge images, identifying the coordinates of the centers of the circles from the binary edge images based on a set radius range by using a Hough gradient method, and the radius and intensity of the circles corresponding to the coordinates of the centers of the circles; and arranging the identified multiple coordinates of the centers of the circles in ascending order of the horizontal coordinates to obtain a first sequence of the coordinates of the centers of the circles; (3) the step of removing overlapping or covering circles is as follows: calculating the absolute value of the difference between the horizontal coordinates of the first coordinate of the center of the circle in the first sequence of the coordinates of the centers of the circles and the horizontal coordinate of the second coordinate of the center of the circle to obtain a horizontal coordinate difference absolute value, when the horizontal coordinate difference absolute value is greater than the maximum radius of the first coordinate of the center of the circle and the second coordinate of the center of the circle, the first coordinate of the center of the circle and the second coordinate of the center of the circle are retained, when the horizontal coordinate difference absolute value is less than or equal to the maximum radius, the distance between the centers of the first coordinate of the center of the circle and the second coordinate of the center of the circle is obtained, then when the distance between the centers is greater than the maximum radius, the first coordinate of the center of the circle and the second coordinate of the center of the circle are retained, when the distance between the centers is less than or equal to the maximum radius, the intensity of the circle corresponding to the first coordinate of the center of the circle and the second coordinate of the center of the circle is compared, and the coordinate of the center of the circle with the maximum intensity is retained; (4) the first coordinate of the center of the circle is compared with the remaining coordinates of the center of the circle in sequence according to step (3) until the horizontal coordinate difference between the first coordinate of the center of the circle and the currently compared coordinate of the center of the circle is greater than the maximum radius, the comparison is completed, or the intensity of the first coordinate of the center of the circle is lower than the intensity of the currently compared coordinate of the center of the circle, the comparison is completed, or the first coordinate of the center of the circle is compared with the last coordinate of the center of the circle, the comparison is completed, and the comparison is completed; (5) after the comparison in step (4) is completed, the first coordinate of the center of the circle is arranged in ascending order of the horizontal coordinates in the first sequence of the coordinates of the centers of the circles, and steps (3)-(4) are performed on the first coordinate of the center of the circle after the first coordinate of the center of the circle and the coordinates of the center of the circle removed in steps (3)-(4) in the first sequence of the coordinates of the centers of the circles, until the last coordinate of the center of the circle in the first sequence of the coordinates of the centers of the circles is completed, the comparison is stopped, and a second sequence of the coordinates of the centers of the circles removed of overlapping or covering impurities is obtained; (6) the coordinates of the centers of the circles in the second sequence of the coordinates of the centers of the circles are arranged in ascending order according to the corresponding circle intensity and circle radius respectively, the coordinates of the centers of the circles corresponding to smaller circle radii and smaller circle intensities are removed based on a circle intensity threshold and a circle radius threshold to obtain a third sequence of the coordinates of the centers of the circles, so that the single existing fluorescent impurities are removed, and the number of the coordinates of the centers of the circles in the third sequence of the coordinates of the centers of the circles is taken as the number of the spherical fluorescent microplastics in the to-be-processed image.

2. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 1, wherein, The preparation method of the fluorescent microplastic sample is as follows: the fluorescent microplastics in the to-be-tested solution are filtered onto a polycarbonate filter membrane to obtain the fluorescent microplastic sample.

3. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 1, wherein, The full-area digital images are cut into multiple to-be-processed images without loss, including: using the cut function of OlyVIA or Photoshop software to cut the full-area digital images into multiple to-be-processed images without loss.

4. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 1, wherein, The to-be-processed images are converted into gradient information gray images, including: The image to be processed is converted into an 8-bit gray scale image, and a Sobel edge detection operator is used to calculate the gradient value Gx in the X-axis direction and the gradient value Gy in the Y-axis direction of each pixel point in the 8-bit gray scale image, and As the total gradient value of each pixel point, each pixel point with a total gradient value constructs a gradient information gray scale image.

5. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 4, characterized in that, The binarization processing of the gradient information gray image obtains a binarized edge image, comprising: The global image threshold of the gradient information gray image is calculated by using the common law, and the binarization processing of the gradient information gray image is performed by using the global image threshold to obtain a binarized edge image.

6. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 5, wherein, The binarization processing of the gradient information gray image comprises: the pixel point value of the total gradient value greater than or equal to the global image threshold is set to 1, and the pixel point value of the total gradient value less than the global image threshold is set to 0, thereby obtaining a binarized edge image.

7. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 1, wherein, The initial center coordinates of the circle are identified from the binarized edge image based on the set radius range by using the Hough gradient method, and the circle radius and the circle intensity of the initial center coordinates, comprising: The circular edge is identified from the binarized edge image based on the set radius range, the gradient direction of the circular edge pixel point and the set radius range are obtained, a plurality of candidate center coordinates are obtained, the voting of the plurality of candidate center coordinates is performed, the candidate center coordinates with the highest total vote value are taken as the center coordinates, and the voting result of the center coordinates is taken as the circle intensity; The pixel points voted for the center coordinates and the plurality of radius values corresponding to the pixel points are obtained, and the voting of the plurality of radius values is performed, and the radius value with the highest total vote value is taken as the circle radius of the center coordinates.

8. The method of spherical fluorescent microplastic counting based on virtual digital slice scanning system and image recognition according to claim 7, wherein, The single vote value p0 of the voting of the plurality of candidate center coordinates and the voting of the plurality of radius values is: R is the set radius value passing through the pixel point.

9. The application of a spherical fluorescent microplastic counting method based on virtual digital slice scanning system and image recognition according to any one of claims 1-8 in the preparation of a fluorescent microplastic standard solution, characterized in that, Comprising: (1) The microplastic dispersion liquid is subjected to suction filtration, and the fluorescent microplastics are enriched on the black polycarbonate filter membrane, and the number of spherical fluorescent microplastics on the black polycarbonate filter membrane is obtained by the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition according to any one of claims 1-8; (2) The black polycarbonate filter membrane enrichment surface is washed, and the eluent is retained, and the washed black polycarbonate filter membrane is counted by the spherical fluorescent microplastic counting method based on the virtual digital slice scanning system and image recognition according to any one of claims 1-8 to obtain the number of spherical fluorescent microplastics on the filter membrane that are not eluted; (3) The number of spherical fluorescent microplastics obtained in step (1) and the number of spherical fluorescent microplastics obtained in step (2) are subtracted to obtain the number of spherical fluorescent microplastics in the elution solution, and the number of spherical fluorescent microplastics in the elution solution is compared with the volume of the elution solution to obtain the abundance of fluorescent microplastics in the elution solution, thereby completing the preparation of the fluorescent microplastic standard solution.

Images