Ultra-fast static image particle detection system and method

Abstract

The present invention relates to the technical field of particle detection, and discloses an ultra-fast static image method particle detection system and method, comprising a main control module, a light source, a platform, an optical path system, an image acquisition module and a data processing module. A filter membrane with a sample is provided on the platform. When detecting particles of the sample, an image acquisition area is divided, the image depth of field of a single acquisition area is fused, image detection and jigsaw puzzle are performed in parallel, and ultra-fast data acquisition and analysis are performed, so that the time of the entire test process is greatly shortened, and the detection result is clear and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of particle detection, and in particular to an ultra-fast static image particle detection system and method. Background Art

[0002] As one of the basic methods for detecting insoluble particles in the pharmaceutical and medical device industries, the microscopic counting method has the characteristic of "what you see is what you get" for the particles being detected, that is, while measuring the particles, you can also observe the shape and structure of the particles (such as solid, loose, porous, etc.), and you can also have a certain understanding of the surface morphology of the particles through vision. In the actual detection process, a microporous filter membrane is required to filter and retain the particles in the solution to be tested, and the filter membrane is dried before the test.

[0003] Due to the fluctuation of particle size and the warping of the filter membrane base during the drying process, the focal plane of the image needs to be adjusted during the test to ensure that the image can be clearly imaged within the fluctuation height difference of the particles and the warping range of the membrane, in order to ensure the accuracy and authenticity of the test results. When using a traditional microscope system, in order to meet the requirements of the accuracy and authenticity of the above test results, the test process often requires repeated adjustment of the field of view height range, which places high operational requirements on the operator and easily causes visual fatigue to the operator.

[0004] At present, a new type of automatic detection device has been developed to replace manual observation. Based on this, the invention patent with publication number CN115855758A discloses an insoluble particle detection device and a method of use. By setting up a shooting component, all areas are photographed, and the clearest images of all areas are spliced ​​by an image processing unit to obtain a clear image of the entire sample. Based on this image, the insoluble particles are analyzed. However, the time consumed by the detection is: acquisition time + puzzle time + detection time, resulting in a long overall time consumption. Summary of the invention

[0005] The purpose of the present invention is to provide an ultra-fast static image particle detection system and method to solve the problems existing in the prior art.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] An ultra-fast static image particle detection method comprises the following steps:

[0008] S1: Plan the collection of the filter membrane area. The specific operations are as follows:

[0009] S11: determine the filter membrane area to be collected;

[0010] S12: calculating a collection path and dividing the collection areas, each area including an overlapping area having an overlapping portion with an adjacent area and a non-overlapping area having no overlapping portion;

[0011] S2: Collect the filter membrane area image, and process and analyze the image. The specific operations are as follows:

[0012] According to the planned acquisition path, multiple images with different depths of field are acquired in the first acquisition area, depth of field fusion processing is performed, and the first fused image is analyzed; multiple images with different depths of field are acquired in the second acquisition area, depth of field fusion processing is performed, and the second fused image is analyzed, and then spliced ​​with the first fused image to obtain a new image, and the new image is analyzed; repeat the previous step until the last acquisition area in the planned acquisition path is completed, and the obtained new image is a clear and complete image of the entire filter membrane area, and the total analysis data is the data of the entire filter membrane area.

[0013] Furthermore, the collection path calculated in step S12 adopts an S shape, and the number of points collected is a minimum value.

[0014] Furthermore, using the area addition method, the measured area is S, the single point detection area is s, the total number of points collected is S / s, and the result should be rounded up;

[0015] The steps for determining the single-point detection area s are as follows: the optical magnification is denoted as I, the size of the detector target surface is determined, the length of the target surface is denoted as h, the width of the target surface is denoted as v, s=h*v / I and s*f≥S, f is the total number of single acquisitions.

[0016] Furthermore, the image analysis and stitching after depth of field fusion in step S2 is specifically performed as follows:

[0017] Perform feature detection on the overlapping area of ​​the image of the first acquisition area, identify feature points, and detect particles in the non-overlapping area; perform feature detection on the overlapping area of ​​the image of the second acquisition area, identify feature points, and detect particles in the non-overlapping area, match the feature points of the image of the second acquisition area with the feature points of the image of the first acquisition area, splice the images, and detect particles in the overlapped area after splicing; repeat the previous step until the image of the last acquisition area is spliced;

[0018] Specifically, since particles in a single image field of view may be cut off by the edge of the image, that is, particles at the edge of the image may be incompletely collected in a certain acquisition image, it is necessary to divide the overlapping area and obtain a complete particle image by image stitching, that is, the filter membrane area covered by the current acquisition point and the filter membrane area covered by the next acquisition point have overlapping parts;

[0019] The specific method for setting the overlapping area is as follows: The overlapping width of the overlapping area is denoted as D pixels, and the range of D pixels can be selected from 50 to 200 pixels, preferably 150 pixels. According to the actual optical magnification I and the detector pixel width P μm, the reduction amount δ of the XY platform at each point in the path can be calculated as δ = D * P / I. According to the length h and width v of the target surface, when shooting along the path, the translation amount ΔX in the X direction for the same row is ΔX = h / I - δ, and when the acquisition point is in a different row from the previous point, the translation amount ΔY in the Y direction is ΔY = v / I – δ. When the XY platform moves to different points, an overlapping area is generated.

[0020] Furthermore, the method of collecting images with different depths of field multiple times in each acquisition area in step S2 is to perform scanning along the Z-axis direction, and the calculation method of the Z-axis scanning times is as follows:

[0021] The Z-axis scanning times n of a single acquisition area are determined by Δh and L: Δh is the undulation difference, Δh = h 2 - h 1 ,h 1 is the lowest height of the substrate of the particle to be detected, and h 2 is the highest height of the substrate of the particle to be detected;

[0022] The spacing L is determined by the depth of field DOF of the optical system.

[0023] Furthermore, when L = DOF, n = Δh / DOF, n is an integer, and when it contains a decimal, it is rounded up. That is, the reference height is h 1 ,scan upward and collect n times, and the acquisition spacing is DOF;

[0024] Furthermore, when L < DOF, L = Δh / (n + m), n = Δh / DOF, n is an integer, and when it contains a decimal, it is rounded up, and m is an integer not less than 1. That is, the reference height is h 1 ,scan upward and collect n + m times, and the acquisition spacing is L.

[0025] A super-fast static image method particle detection system includes:

[0026] A main control module for the overall control and scheduling of the system and realizing high-speed data acquisition in cooperation with the data processing module;

[0027] An image acquisition module for acquiring images;

[0028] An optical path system for focusing, and the optical path system includes an objective lens, a beam splitter, and a relay lens installed in sequence;

[0029] A light source for cooperating with the image acquisition module and the optical path system to acquire clear images;

[0030] A platform, used to move the collection area according to the planned collection path, wherein a filter membrane with a sample is provided on the platform, and the sample is a particle to be detected;

[0031] The data processing module is used for image fusion, image analysis and data processing after imaging.

[0032] Furthermore, the main control module controls the movement and position of the optical path system and the platform, and the main control module lights up the light source and triggers the image acquisition module to acquire images.

[0033] Furthermore, the optical path system and the platform are both provided with encoders, and the main control module, the optical path system and the platform adopt closed-loop control.

[0034] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0035] 1. The present invention ensures that the image data collected in each acquisition area is clearly focused and accurate by adopting image depth fusion, so that the data in the entire image area is clear and accurate.

[0036] 2. The present invention improves the entire detection speed by dividing the collection area and adopting a parallel approach of image detection and puzzle-solving, thereby shortening the entire test process from 20 to 30 minutes to about 3 minutes.

[0037] 3. The present invention adopts a stitching algorithm based on feature recognition, which avoids the defect and loss of particle images during the image stitching process and ensures the accuracy of the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0039] Figure 1 It is a schematic diagram of the steps of the present invention;

[0040] Figure 2 is a schematic diagram of an image acquisition path of a filter membrane region of the present invention;

[0041] Figure 3 is a schematic diagram of a single collection area of ​​the present invention;

[0042] Figure 4 is a schematic diagram of the height of the substrate of the present invention;

[0043] Figure 5 It is a connection schematic diagram of the detection system of the present invention;

[0044] In the figure: 1. Filter membrane; 2. Overlapping area; 3. Non-overlapping area. DETAILED DESCRIPTION

[0045] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] The invention discloses an ultra-fast static image method particle detection system, which comprises a main control module, an image acquisition module, an optical path system, a light source, a platform and a data processing module.

[0047] The main control module is responsible for the overall control and scheduling of the system, and collaborates with the data processing module to achieve high-speed data collection.

[0048] The image acquisition module is used to acquire images.

[0049] The optical path system is used for focusing, and includes an objective lens, a beam splitter, and a relay lens installed in sequence.

[0050] The light source is used to cooperate with the image acquisition module and the optical path system to acquire clear images.

[0051] A filter membrane 1 with a sample is arranged on the platform, and the sample is the particle to be detected.

[0052] The data processing module is used for image fusion, image analysis and data processing after imaging.

[0053] The main control module is also responsible for the motion control of the optical system and the platform. Preferably, the optical system and the platform are both equipped with encoders, and the main control module, the optical system and the platform adopt closed-loop control, and receive encoder feedback from the optical system and the platform to achieve accurate and rapid positioning of the optical system and the platform.

[0054] The main control module is connected to the light source and the image acquisition module. When the optical path system and the platform reach the target position, the light source is triggered to light up, and the image acquisition module is triggered to collect images at the same time. The data processing module obtains the image signal of the current position; the position is obtained from the main control system, and the light source is in high-brightness mode to meet the exposure time for ultra-short image acquisition, which is generally in the microsecond level. Preferably, the exposure time is 2μs to 10μs.

[0055] See also Figure 1-Figure 3 The present invention also discloses an ultra-fast static image particle detection method, comprising:

[0056] S1: Plan the collection of the filter membrane area. The specific operations are as follows:

[0057] S11: determine the filter membrane area to be collected;

[0058] S12: The main control module calculates the optimal acquisition path and divides the acquisition area to ensure that the number of points collected is the minimum value. That is, assuming that the area of ​​the area to be collected is S, that is, the measured area, the area of ​​a single acquisition area is s, that is, the single-point detection area, and the minimum value is the minimum number of s that need to be collected to cover S. Each area includes an overlapping area 2 that has an overlapping part with the adjacent area and a non-overlapping area 3 that does not have an overlapping part.

[0059] Specifically, the area addition method is adopted, the measured area is S, the single point detection area is s, the total number of points collected is S / s, and the result should be rounded up to ensure that the measured area is completely covered by the sum of the single collection area;

[0060] The step of determining the single point detection area s is: determining the size of the detector target surface, and s*f≥S, where f is the total number of single acquisitions;

[0061] Furthermore, the detector is usually a camera, the length of the camera target surface area is recorded as h, the width of the camera target surface area is recorded as v, and s=h*v / I is set;

[0062] Among them, the optical magnification I is 5 to 100 times;

[0063] Then, S is a perfect circle with a diameter of 10mm~25mm, h is 8.4mm~13.8mm, and v is 7.1mm~10.4mm;

[0064] Specifically, ① when S is a perfect circle with a diameter of 25 mm, a camera with a target area of ​​13.8 mm*10.4 mm is used, and the optical magnification I is 10 times, it is necessary to obtain 13.8*10.4*f≥12.5 2 *Π*10;

[0065] ② When S is a perfect circle with a diameter of 16mm, a camera with a target area of ​​13.8mm*10.4mm is used, and the optical magnification I is 10 times, it is necessary to obtain 13.8*10.4*f≥8 2 *Π*10;

[0066] ③ When S is a perfect circle with a diameter of 10 mm, a camera with a target area of ​​8.4 mm*7.1 mm is used, and the optical magnification I is 5 times, it is necessary to obtain 8.4*7.1*f≥5 2 *Π*10.

[0067] S2: Collect the filter membrane area image, and process and analyze the image. The specific operations are as follows:

[0068] According to the planned acquisition path, images with different depths of field are acquired multiple times in the first acquisition area, depth-of-field fusion processing is performed, and feature detection is performed on the first fused image to identify feature points and detect particles in the non-overlapping area 3; images with different depths of field are acquired multiple times in the second acquisition area, depth-of-field fusion processing is performed, and feature detection is performed on the second fused image to identify feature points and detect particles in the non-overlapping area 3. The feature points of the images in the second acquisition area are matched with the feature points of the images in the first acquisition area, the images are stitched, and particles are detected in the overlapping area 2 after stitching; the above steps are repeated until the last acquisition area in the planned acquisition path is completed. The new image obtained is a clear and complete image of the entire filter membrane area, and the total analysis data obtained is the data of the entire filter membrane area.

[0069] Please refer to Figure 4 , to ensure that the image data acquired in each acquisition area is clearly and accurately focused, the following specific implementation methods are used for multiple acquisitions: First step, rapid pre-acquisition of the sample to determine the lowest height h 1 and the highest height h 2 , the undulation difference Δh = h 2 - h 1 ; Second step, determine the number of Z-axis scans n and the acquisition interval L. The acquisition interval L is determined by the depth of field DOF of the optical system. The acquisition interval L should not exceed the depth of field of the optical system, that is, L ≤ DOF;

[0070] Method 1, adopt the acquisition strategy of L = DOF. The number of Z-axis scans n is determined by Δh and L, n = Δh / L, n is an integer, otherwise round up. That is, the reference height is h 1 , and scan upward and acquire n times;

[0071] Method 2, adopt the acquisition strategy of L < DOF. At this time, L = Δh / (n + m), n = Δh / DOF, n is an integer, round up when containing decimals, m is an integer not less than 1, and the reference height is h 1 , scan upward and acquire n + m times, and the acquisition interval is L;

[0072] The above rounding up ensures that the detection times are completely detected;

[0073] Example 1: Adopt the acquisition strategy of L = DOF, h 1 = -1.695mm, h 2 = -1.607mm, Δh = 88um, depth of field DOF = 10um, acquisition interval L is taken as 10um, then n = Δh / L = 8.8, n is rounded up to 9, and acquire upward 9 times;

[0074] Example 2: Adopt the acquisition strategy of L = DOF, h 1= -1.705 mm, h 2 = -1.636 mm, Δh = 69 μm, depth of field DOF = 6 μm, acquisition pitch L is taken as 6 μm, n = Δh / L = 11.5, n is rounded up to 12, and acquire upward 12 times;

[0075] Example 3: Adopt an acquisition strategy with L < DOF, h 1 = -1.705 mm, h 2 = -1.636 mm, Δh = 69 μm, depth of field DOF = 6 μm, n = Δh / DOF = 11.5, n is rounded up to 12, m is taken as 2, acquisition pitch L = Δh / (n + m) = 4.9 μm, and acquire upward n + m = 14 times.

[0076] In summary, the image acquisition module acquires each acquisition area in an S shape, and the data processing module performs depth-of-field fusion processing on the images, completing the detection and stitching of the images in each acquisition area. When the detection and stitching of the images in the last acquisition area are completed, the particle detection and image stitching of the entire detection area are completed.

[0077] The entire detection process takes a short time, improving work efficiency. The images are complete and clear, improving the accuracy of the detection results.

[0078] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be construed as limiting the claimed rights.

Claims

1. An ultra-fast static image particle detection method, characterized in that: It includes the following steps: S1: Plan the collection of the filter membrane area. The specific operations are as follows: S11: Determine the filter membrane area to be collected; S12: Calculate the collection path, divide the collection area, and each area includes an overlapping area (2) with an overlapping part with an adjacent area and a non-overlapping area (3) without an overlapping part; S2: Collect images of the filter membrane area, and process and analyze the images. The specific operations are as follows: According to the planned collection path, collect images with different depths of field multiple times in the first collection area, perform depth-of-field fusion processing, and analyze the first fused image; collect images with different depths of field multiple times in the second collection area, perform depth-of-field fusion processing, and analyze the second fused image, then splice it with the first fused image to obtain a new image, and analyze the new image; repeat the above steps until the last collection area in the planned collection path is completed. The obtained new image is a clear and complete image of the entire filter membrane area, and the total analysis data obtained is the data of the entire filter membrane area; Analysis and splicing of the depth-of-field fused images. The specific operations are as follows: Perform feature detection on the overlapping area (2) of the image in the first collection area to identify feature points, and detect particles in the non-overlapping area (3); perform feature detection on the overlapping area (2) of the image in the second collection area to identify feature points, detect particles in the non-overlapping area (3), match the feature points of the image in the second collection area with the feature points of the image in the first collection area, splice the images, and detect particles in the overlapping area (2) after splicing; Repeat the above steps until the images of the last collection area are spliced.

2. The ultra-fast static image particle detection method according to claim 1, characterized in that: The collection path calculated in step S12 adopts an S shape, and the number of collection points is the minimum value.

3. The ultra-fast static image particle detection method according to claim 2, characterized in that: Adopt the area addition method. The measured area is S, the single-point detection area is s, and the total number of collection points is S / s. The result should be rounded up; The steps for determining the single-point detection area s are as follows: The optical magnification is denoted as I, determine the size of the detector target surface, the length of the target surface is denoted as h, the width of the target surface is denoted as v, then s = h * v / I, and s * f ≥ S, where f is the total number of single-point collections.

4. The ultra-fast static image particle detection method according to claim 1, characterized in that: In step S2, the method of collecting images with different depths of field in each collection area is to scan along the Z-axis direction. The calculation method of the Z-axis scan times is as follows: The Z-axis scan times n of a single collection area are determined by Δh and L: Δh is the undulation difference, Δh = h2 - h1, h1 is the lowest height of the base of the particle to be detected, and h2 is the highest height of the base of the particle to be detected; The collection pitch L is determined by the depth of field DOF of the optical system.

5. The ultra-fast static image particle detection method according to claim 4, characterized in that: When L = DOF, n = Δh / DOF, n is an integer, and when it contains a decimal, it is rounded up. That is, the reference height is h1, and it is scanned and collected n times upward, and the collection pitch is DOF.

6. The ultra-fast static image particle detection method according to claim 4, characterized in that: When L < DOF, L = Δh / (n + m), n = Δh / DOF, n is an integer, and when it contains a decimal, it is rounded up, and m is an integer not less than 1. That is, the reference height is h1, and it is scanned and collected n + m times upward, and the collection pitch is L.

7. A detection system using the ultra-fast static image particle detection method according to any one of claims 1 to 6, characterized in that: The system includes: The main control module is used for overall control and scheduling of the system and to coordinate with the data processing module to achieve high-speed data collection; An image acquisition module, used for acquiring images; An optical path system, used for focusing, the optical path system comprising an objective lens, a beam splitter and a relay lens installed in sequence; A light source, used to cooperate with the image acquisition module and the optical path system to acquire clear images; A platform, used to move the collection area according to the planned collection path, wherein a filter membrane (1) with a sample is provided on the platform, and the sample is a particle to be detected; Data processing module, used for image fusion, image analysis and data processing after imaging; The main control module controls the movement and position of the optical path system and the platform, and the main control module lights up the light source and triggers the image acquisition module to acquire images at the same time.

8. The detection system according to claim 7, characterized in that: The optical path system and the platform are both provided with encoders, and the main control module, the optical path system and the platform adopt closed-loop control.

Images