System and method for classifying urine visible components

Abstract

The invention provides a urine visible component classification system and method, and the system comprises: a sample introduction module, which is used for mixing a urine sample with a preset flow of sheath fluid to form a to-be-detected liquid flow; the turbulent flow overturning chip is used for overturning the liquid flow to be detected to form overturning liquid flow; the form information acquisition module is used for acquiring a liquid flow image sequential sequence based on the turnover liquid flow and extracting multi-angle form information of a target visible component in the turnover liquid flow based on the liquid flow image sequential sequence; the visible component classification module is used for obtaining a classification result according to the multi-angle form information; wherein the turbulent flow overturning chip comprises a plurality of reducing sections, the flow speed of the to-be-measured liquid flow is increased when the to-be-measured liquid flow flows through the narrow flow channel sections of the reducing sections, and the flow speed of the to-be-measured liquid flow is decreased when the to-be-measured liquid flow flows through the wide flow channel sections of the reducing sections, so that periodic flow speed disturbance is generated when the to-be-measured liquid flow flows through the plurality of reducing sections, visible components in the to-be-measured liquid flow are continuously overturned, and overturning liquid flow is formed. According to the invention, the accuracy of organic component classification can be improved.

Description

Technical Field

[0001] This invention relates to the field of medical image analysis technology, and in particular to a urine formation classification system and method. Background Technology

[0002] Traditional urine formed element classification systems typically employ flow cytometry. This method mixes the urine sample with sheath fluid to form a monolayer, allowing the formed elements to pass sequentially through a fixed detection area. A high-speed camera captures static images of each formed element within the monolayer for classification. However, this traditional approach only captures a single image of each formed element, and this single image only captures a static image of the formed element from a fixed angle. Since different formed elements in urine may have similar morphologies at a specific angle, misclassification is highly likely. Furthermore, traditional methods are insufficient to accurately classify formed elements with unique morphologies from a single angle, thus failing to effectively classify formed elements.

[0003] To address the aforementioned issues, existing technologies calculate the corresponding morphological angle of a static image of formed elements captured at a fixed angle in a single imaging session. Based on this morphological angle, they estimate subsequent morphological changes in the formed elements and model the estimated morphology for classification. However, urine formed elements exhibit diverse morphologies, and their morphological changes during actual flow are complex. This leads to significant errors in the estimation of morphological changes in formed elements by existing technologies, resulting in low reliability of the subsequent modeling used for organic component classification and reducing the accuracy of organic component classification. Summary of the Invention

[0004] The present invention aims to provide a urine organic component classification system and method to solve the above-mentioned technical problems and improve the accuracy of organic component classification.

[0005] To address the aforementioned technical problems, the present invention provides a urine formation sorting system, comprising: The sample injection module is used to mix urine samples with a preset flow rate of sheath fluid to form the test fluid stream; A turbulence-reversing chip is used to reverse the liquid flow under test, forming a reversed liquid flow. The morphological information acquisition module is used to acquire a time sequence of liquid flow images based on the flipped liquid flow, and to extract multi-angle morphological information of the target formed elements in the flipped liquid flow based on the time sequence of liquid flow images. A formed element classification module is used to classify based on the multi-angle morphological information to obtain the classification result corresponding to the target formed element; The turbulence-reversing chip includes several variable-diameter sections. When the liquid to be tested flows through the narrow sections of the variable-diameter sections, the flow velocity increases, and when the liquid to be tested flows through the wide sections of the variable-diameter sections, the flow velocity decreases. This causes periodic flow velocity disturbances when the liquid to be tested flows through the variable-diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow. The formed elements in the reversed liquid flow are in a continuous reversing state.

[0006] In the above scheme, when the test liquid flows through the narrow section of a variable-diameter channel, the flow velocity increases as the channel width narrows. Subsequently, when the test liquid flows through the wide section of the variable-diameter channel, the flow velocity decreases as the channel width widens. Therefore, the test liquid undergoes periodic velocity changes as it flows through several variable-diameter sections. Furthermore, the velocity differences in the test liquid exert uneven fluid forces on the flowing organic components, causing them to tumble. Thus, this scheme ensures that the organic components continuously tumble as the test liquid flows through several variable-diameter sections, enabling the acquisition of multi-angle morphological information containing the target formed elements at different angles. This multi-angle morphological information is then used for classification. Compared to traditional schemes that classify based on static image information from a single angle, this scheme utilizes multi-angle morphological information of formed elements for classification, providing more comprehensive morphological information for organic component classification and thus improving the accuracy of organic component classification.

[0007] Further, the morphology information acquisition module is used to acquire a liquid flow image time sequence based on the flipped liquid flow, and to extract multi-angle morphology information of the target formed elements in the flipped liquid flow based on the liquid flow image time sequence. The extraction of multi-angle morphology information of the target formed elements in the flipped liquid flow based on the liquid flow image time sequence includes: acquiring the first frame in which the target formed element first appears in the liquid flow image time sequence, and acquiring the first position corresponding to the target formed element based on the liquid flow image corresponding to the first frame in the liquid flow image time sequence; taking the first position as the current position and the first frame as the current frame; acquiring the acquisition frame rate corresponding to the liquid flow image time sequence, and performing a morphology position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; integrating the first frame, the first position, and several motion positions to obtain a target position time sequence; and extracting a set of local images of formed elements from the liquid flow image time sequence based on the target position time sequence, and taking the set of local images of formed elements as multi-angle morphology information.

[0008] In the above scheme, the position of the target formed element changes over time. The first position is the position of the target formed element in the fluid flow image corresponding to the first frame. Starting from this first position, the positions of the target formed element in the fluid flow images corresponding to subsequent frames are obtained, i.e., the motion positions are obtained. This scheme integrates several motion positions with the first position to reflect the positional changes of the target formed element. Integrating the first frame further confirms the corresponding fluid flow images of each target position in the time sequence of the target positions. Subsequently, this scheme extracts a local image centered on the target position from the fluid flow images corresponding to each target position. This local image can present the morphology of the target formed element in the current frame. The final local image set contains local images of the target formed element's morphology from different angles. Compared to traditional schemes that can only obtain static images from a single angle for classification, this scheme provides multi-angle morphological information that contains richer morphological information of the target formed element, thereby improving the accuracy of subsequent organic component classification.

[0009] Further, in the step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions, the morphological position acquisition step includes: obtaining the current flow velocity corresponding to the current position, and obtaining the current moving distance based on the acquisition frame rate and the current flow velocity; generating an initial motion position based on the current moving distance and the current position; confirming that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then performing the motion position acquisition step; otherwise, stopping the execution of the morphological position acquisition step, and obtaining several motion positions corresponding to each motion position acquisition step; the motion position acquisition step includes: extracting the fluid flow image corresponding to the next frame from the time sequence of the fluid flow image as a subsequent fluid flow image based on the next frame corresponding to the current frame; filtering based on the subsequent fluid flow image and the initial motion position to obtain motion positions; using the motion position as the current position, using the next frame as the current frame, and re-executing the morphological position acquisition step.

[0010] Further, the step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: obtaining a set of formed component positions with the initial motion position as the center within a preset search radius based on the subsequent fluid flow image; generating an error distance corresponding to each formed component position in the set of formed component positions based on the initial motion position and the set of formed component positions; and taking the formed component position corresponding to the smallest error distance as the motion position.

[0011] In the above scheme, the time interval between adjacent frames of the liquid flow image time sequence can be obtained based on the acquisition frame rate. Therefore, according to the current flow velocity and the interval time corresponding to the acquisition frame rate, the distance that the target formed element needs to move from the current frame to the next frame can be obtained, i.e., the current moving distance. Then, based on the current moving distance and the current position, the ideal position of the target formed element in the next frame can be obtained, i.e., the initial motion position. However, the initial motion position is only an estimated ideal position. There may not be a formed element at the initial motion position in the liquid flow image of the next frame. Therefore, it is necessary to search for all formed element positions within a preset search radius with the initial motion position as the center, and take the formed element position with the smallest error distance between the formed element position and the initial motion position as the motion position, i.e., the position of the target formed element in the liquid flow image of the next frame. This scheme can find the matching position of the target formed element from liquid flow images of different frames when organic components are flipped and moved, realizing effective tracking of the position of the target formed element, improving the reliability of the multi-angle morphological information of the target formed element obtained later, and thus improving the accuracy of subsequent organic component classification.

[0012] Furthermore, the formed element classification module is used to classify based on the multi-angle morphological information to obtain the classification result corresponding to the target formed element, including: extracting the three-dimensional information of the target formed element based on the multi-angle morphological information; comparing the three-dimensional information of the target formed element with a preset formed element classification library to obtain the classification result corresponding to the target formed element.

[0013] Furthermore, the turbulence-reversing chip includes several variable-diameter sections, and the flow velocity of the liquid to be tested increases when it flows through the narrow sections of the variable-diameter sections and decreases when it flows through the wide sections of the variable-diameter sections, causing periodic flow velocity disturbances when the liquid to be tested flows through the variable-diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow; the formed elements in the reversed liquid flow are in a continuous reversing state, and the cross-sectional shape of the narrow section of the variable-diameter section is triangular, trapezoidal, arc-shaped, or rectangular.

[0014] This invention also provides a method for classifying urine formed elements, applied to a urine formed element classification system as described in any of the preceding claims, comprising: mixing a urine sample with a sheath fluid of a preset flow rate to form a test flow; flipping the test flow to form a flipped flow; acquiring a time-series sequence of flow images based on the flipped flow, and extracting multi-angle morphological information of target formed elements in the flipped flow based on the time-series sequence of flow images; classifying according to the multi-angle morphological information to obtain a classification result corresponding to the target formed elements.

[0015] Further, the step of obtaining a time sequence of liquid flow images based on the flipped liquid flow, and extracting multi-angle morphological information of the target formed elements in the flipped liquid flow based on the time sequence of liquid flow images, includes: taking the frame in which the target formed element first appears in the time sequence of liquid flow images as the first frame, and obtaining the first position corresponding to the target formed element based on the liquid flow image corresponding to the first frame in the time sequence of liquid flow images; taking the first position as the current position, and taking the first frame as the current frame; obtaining the acquisition frame rate corresponding to the time sequence of liquid flow images, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; integrating the first frame, the first position, and several motion positions to obtain a target position time sequence; extracting a set of local images of formed elements from the time sequence of liquid flow images based on the target position time sequence, and obtaining multi-angle morphological information based on the set of local images of formed elements.

[0016] Further, in the step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions, the morphological position acquisition step includes: obtaining the current flow velocity corresponding to the current position, and obtaining the current moving distance based on the acquisition frame rate and the current flow velocity; generating an initial motion position based on the current moving distance and the current position; confirming that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then performing the motion position acquisition step; otherwise, stopping the execution of the morphological position acquisition step, and obtaining several motion positions corresponding to each motion position acquisition step; the motion position acquisition step includes: extracting the fluid flow image corresponding to the next frame from the time sequence of the fluid flow image as a subsequent fluid flow image based on the next frame corresponding to the current frame; filtering based on the subsequent fluid flow image and the initial motion position to obtain motion positions; using the motion position as the current position, using the next frame as the current frame, and re-executing the morphological position acquisition step.

[0017] Further, the step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: obtaining a set of formed component positions with the initial motion position as the center within a preset search radius based on the subsequent fluid flow image; generating an error distance corresponding to each formed component position in the set of formed component positions based on the initial motion position and the set of formed component positions; and taking the formed component position corresponding to the smallest error distance as the motion position.

[0018] In the above scheme, the organic components in the test liquid flow are flipped to form a flipped liquid flow, enabling the acquisition of multi-angle morphological information containing the target formed components at different angles, and classification is performed based on this multi-angle morphological information. Compared with traditional schemes that classify based on static image information from a single angle, this scheme uses multi-angle morphological information of formed components for classification, which can provide more comprehensive morphological information for organic component classification, thereby improving the accuracy of organic component classification. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a urine formation and sorting system according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the flow velocity of the liquid to be measured in a turbulence-reversing chip of a urine formation classification system according to an embodiment of the present invention; Figure 3 A schematic diagram of a triangular cross-section of a variable-diameter flow channel narrow section of a urine formation sorting system provided in an embodiment of the present invention; Figure 4 A schematic diagram of the cross-sectional shape of a trapezoidal turbulence-flipping chip in a narrow section of a variable-diameter flow channel of a urine formation sorting system provided in an embodiment of the present invention; Figure 5 A schematic diagram of the cross-sectional shape of a turbulence-flipping chip in a narrow section of a variable-diameter flow channel of a urine formation sorting system provided in an embodiment of the present invention; Figure 6 A schematic diagram of a rectangular cross-section of a variable-diameter flow channel narrow section of a urine formation sorting system provided in an embodiment of the present invention; Figure 7 This is a flowchart of a urine formation classification method according to an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 This embodiment provides a urine formation classification system, including: The sample injection module is used to mix urine samples with a preset flow rate of sheath fluid to form the test fluid stream; A turbulence-reversing chip is used to reverse the liquid flow under test, forming a reversed liquid flow. The morphological information acquisition module is used to acquire a time sequence of liquid flow images based on the flipped liquid flow, and to extract multi-angle morphological information of the target formed elements in the flipped liquid flow based on the time sequence of liquid flow images. A formed element classification module is used to classify based on the multi-angle morphological information to obtain the classification result corresponding to the target formed element; The turbulence-reversing chip includes several variable-diameter sections. When the liquid to be tested flows through the narrow sections of the variable-diameter sections, the flow velocity increases, and when the liquid to be tested flows through the wide sections of the variable-diameter sections, the flow velocity decreases. This causes periodic flow velocity disturbances when the liquid to be tested flows through the variable-diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow. The formed elements in the reversed liquid flow are in a continuous reversing state.

[0022] In the above embodiments, when the test liquid flows through the narrow section of a variable-diameter channel, the flow velocity increases as the channel width narrows. Subsequently, when the test liquid flows through the wide section of the variable-diameter channel, the flow velocity decreases as the channel width widens. Therefore, the test liquid undergoes periodic velocity changes as it flows through several variable-diameter sections. Furthermore, the velocity differences in the test liquid exert uneven fluid forces on the flowing organic components, causing them to tumble. Thus, in this embodiment, the organic components continuously tumble as the test liquid flows through several variable-diameter sections, enabling the acquisition of multi-angle morphological information containing different angular shapes of the target formed elements, and classification based on this multi-angle morphological information. Compared to traditional methods that classify based on static image information from a single angle, this embodiment utilizes multi-angle morphological information of formed elements for classification, providing more comprehensive morphological information for organic component classification and thus improving the accuracy of organic component classification.

[0023] In one embodiment, a preset flow rate can be set according to the required mixing ratio of urine sample and sheath fluid, and the urine sample is mixed with the preset flow rate of sheath fluid to obtain the fluid flow to be tested. In one embodiment, the acquisition of fluid flow images in the fluid flow image time sequence is achieved by high-speed camera acquisition.

[0024] In one embodiment, a schematic diagram of the flow rate of the liquid to be measured in the turbulence-reversing chip is shown below. Figure 2 As shown in the figure, the arrows do not indicate the direction of the flow velocity of the liquid being measured at each location. The closer the color of a location is to red, the faster the flow velocity at that location; conversely, the closer the color is to blue, the slower the flow velocity at that location. From... Figure 2It can be seen that when the liquid to be tested flows through the wide section of the variable diameter section, the color corresponding to the flow velocity changes from red to blue because the channel width changes from narrow to wide, that is, the flow velocity of the liquid to be tested slows down; when the liquid to be tested flows through the narrow section of the variable diameter section, the color corresponding to the flow velocity changes from blue to red because the channel width changes from wide to narrow, that is, the flow velocity of the liquid to be tested slows down. The difference in flow velocity between different positions in the liquid to be tested will exert uneven fluid forces on the organic components flowing through it, causing the organic components to flip.

[0025] Further, the morphology information acquisition module is used to acquire a liquid flow image time sequence based on the flipped liquid flow, and to extract multi-angle morphology information of the target formed elements in the flipped liquid flow based on the liquid flow image time sequence. The extraction of multi-angle morphology information of the target formed elements in the flipped liquid flow based on the liquid flow image time sequence includes: acquiring the first frame in which the target formed element first appears in the liquid flow image time sequence, and acquiring the first position corresponding to the target formed element based on the liquid flow image corresponding to the first frame in the liquid flow image time sequence; taking the first position as the current position and the first frame as the current frame; acquiring the acquisition frame rate corresponding to the liquid flow image time sequence, and performing a morphology position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; integrating the first frame, the first position, and several motion positions to obtain a target position time sequence; and extracting a set of local images of formed elements from the liquid flow image time sequence based on the target position time sequence, and taking the set of local images of formed elements as multi-angle morphology information.

[0026] In the above embodiments, the position of the target formed element changes over time. The first position is the position of the target formed element in the fluid flow image corresponding to the first frame. Starting from this first position, the positions of the target formed element in the fluid flow images corresponding to subsequent frames are obtained, i.e., the motion positions are obtained. In this embodiment, the integration of several motion positions and the first position can reflect the positional changes of the target formed element. Integrating the first frame can confirm the corresponding fluid flow image of each target position in the time sequence of the target position in the fluid flow image time sequence. Subsequently, in this embodiment, a local image centered on the target position is extracted from the fluid flow image corresponding to each target position. This local image can present the morphology of the target formed element in the current frame. The final local image set contains local images of the target formed element at different angles. Compared with traditional schemes that can only obtain static images from a single angle for classification, the multi-angle morphology information provided in this embodiment can contain richer morphology information of the target formed element, thereby improving the accuracy of subsequent organic component classification.

[0027] Further, in the step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions, the morphological position acquisition step includes: obtaining the current flow velocity corresponding to the current position, and obtaining the current moving distance based on the acquisition frame rate and the current flow velocity; generating an initial motion position based on the current moving distance and the current position; confirming that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then performing the motion position acquisition step; otherwise, stopping the execution of the morphological position acquisition step, and obtaining several motion positions corresponding to each motion position acquisition step; the motion position acquisition step includes: extracting the fluid flow image corresponding to the next frame from the time sequence of the fluid flow image as a subsequent fluid flow image based on the next frame corresponding to the current frame; filtering based on the subsequent fluid flow image and the initial motion position to obtain motion positions; using the motion position as the current position, using the next frame as the current frame, and re-executing the morphological position acquisition step.

[0028] Further, the step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: obtaining a set of formed component positions with the initial motion position as the center within a preset search radius based on the subsequent fluid flow image; generating an error distance corresponding to each formed component position in the set of formed component positions based on the initial motion position and the set of formed component positions; and taking the formed component position corresponding to the smallest error distance as the motion position.

[0029] In the above embodiments, the time interval between adjacent frames of the liquid flow image time sequence can be obtained based on the acquisition frame rate. Therefore, according to the current flow velocity and the interval time corresponding to the acquisition frame rate, the distance that the target formed element needs to move from the current frame to the next frame can be obtained, i.e., the current moving distance. Then, based on the current moving distance and the current position, the ideal position of the target formed element in the next frame can be obtained, i.e., the initial motion position. However, the initial motion position is only an estimated ideal position. There may not be a formed element at the initial motion position in the liquid flow image of the next frame. Therefore, it is necessary to search for all formed element positions within a preset search radius with the initial motion position as the center, and take the formed element position with the smallest error distance between the formed element position and the initial motion position as the motion position, i.e., as the position of the target formed element in the liquid flow image of the next frame. This embodiment can find the matching position of the target formed element from liquid flow images of different frames when organic components are flipped and moved, realizing effective tracking of the position of the target formed element, improving the reliability of the multi-angle morphological information of the target formed element obtained subsequently, and thus improving the accuracy of subsequent organic component classification.

[0030] Furthermore, the formed element classification module is used to classify according to the multi-angle morphological information to obtain the classification result corresponding to the target formed element, including: extracting the three-dimensional information of the target formed element based on the multi-angle morphological information; comparing the three-dimensional information of the target formed element with a preset formed element classification library to obtain the classification result corresponding to the target formed element.

[0031] Furthermore, the turbulence-reversing chip includes several variable-diameter sections, and the flow velocity of the liquid to be tested increases when it flows through the narrow sections of the variable-diameter sections and decreases when it flows through the wide sections of the variable-diameter sections, causing periodic flow velocity disturbances when the liquid to be tested flows through the variable-diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow; the formed elements in the reversed liquid flow are in a continuous reversing state, and the cross-sectional shape of the narrow section of the variable-diameter section is triangular, trapezoidal, arc-shaped, or rectangular.

[0032] In one embodiment, a turbulence-reversing chip with a triangular cross-sectional shape in the narrow section of a variable-diameter flow channel is used. Figure 3 As shown in the cross-sectional schematic diagram, the flow-changing chip in this embodiment includes a first flow channel inlet 11, a first flow channel outlet 41, and a first variable diameter section 51, and the first variable diameter section 51 includes a first narrow flow channel section 21 and a first wide flow channel section 31. In this embodiment, the flow channel width of the first flow channel inlet 11 is wider than the flow channel width of the first narrow flow channel section 21. When the liquid to be tested flows from the first flow channel inlet 11 to the first narrow flow channel section 21, the change in flow channel width from wide to narrow will accelerate the flow velocity of the liquid to be tested, and the first flow channel inlet 11 serves as the inlet for the liquid to be tested to flow into the flow-changing chip. The cross-sectional shape of the first narrow flow channel section 21 is triangular, and its inclined straight sidewalls can generate a relatively strong and concentrated fluid force in the flow field of the liquid to be tested. The process is conducive to generating a strong rotational torque to flip the target formed element, which is suitable for application scenarios that require strong and rapid flipping of formed elements; when the liquid to be tested flows from the narrow section 21 of the first flow channel to the wide section 31 of the first flow channel, the change in flow channel width from narrow to wide will slow down the flow rate of the liquid to be tested; when the liquid to be tested flows from the wide section 31 of the first flow channel to the next narrow section of the flow channel, the change in flow channel width from wide to narrow will speed up the flow rate of the liquid to be tested; the outlet 41 of the first flow channel is the outlet for the liquid to be tested to flow out of the turbulence flipping chip.

[0033] In one embodiment, a turbulence-reversing chip with a trapezoidal cross-sectional shape in the narrow section of a variable-diameter flow channel is used. Figure 4As shown in the cross-sectional schematic diagram, the flow-changing chip in this embodiment includes a second flow channel inlet 12, a second flow channel outlet 42, and a second variable diameter section 52. The second variable diameter section 52 includes a second narrow flow channel section 22 and a second wide flow channel section 32. In this embodiment, the flow channel width of the second flow channel inlet 12 is wider than the flow channel width of the second narrow flow channel section 22. When the liquid to be tested flows from the second flow channel inlet 12 to the second narrow flow channel section 22, the change in flow channel width from wide to narrow will accelerate the flow velocity of the liquid to be tested. The second flow channel inlet 12 serves as the inlet for the liquid to be tested to flow into the flow-changing chip. The cross-sectional shape of the second narrow flow channel section 22 is trapezoidal, and its inclined straight sidewalls can generate a relatively strong and directional concentrated fluid force change in the flow field of the liquid to be tested. This is beneficial for generating a strong rotational torque to flip the target formed element, and is suitable for application scenarios where formed elements require strong and rapid flipping. When the liquid to be tested flows from the narrow section 22 of the second flow channel to the wide section 32 of the second flow channel, the change in flow channel width from narrow to wide will slow down the flow rate of the liquid to be tested. When the liquid to be tested flows from the wide section 32 of the second flow channel to the next narrow section of the flow channel, the change in flow channel width from wide to narrow will speed up the flow rate of the liquid to be tested. The outlet 42 of the second flow channel is the outlet for the liquid to be tested to flow out of the turbulence flipping chip.

[0034] In one embodiment, a turbulence-reversing chip with an arc-shaped cross-section in the narrow section of a variable-diameter flow channel is used. Figure 5 As shown in the cross-sectional schematic diagram, the turbulence-reversing chip in this embodiment includes a third flow channel inlet 13, a third flow channel outlet 43, and a third variable diameter section 53. The third variable diameter section 53 includes a third narrow flow channel section 23 and a third wide flow channel section 33. In this embodiment, the flow channel width of the third flow channel inlet 13 is wider than the flow channel width of the third narrow flow channel section 23. When the liquid to be tested flows from the third flow channel inlet 13 to the third narrow flow channel section 23, the change in flow channel width from wide to narrow will accelerate the flow velocity of the liquid to be tested. The third flow channel inlet 13 serves as the inlet for the liquid to be tested to flow into the turbulence-reversing chip. The cross-sectional shape of the third narrow flow channel section 23 is arc-shaped, and its smooth transition contour is conducive to the smooth change of the flow field of the liquid to be tested, reducing flow separation and turbulence, and reducing the risk of flow separation and turbulence. The formation of components may be damaged or aggregated during drastic changes, and may also generate more complex multi-directional disturbances, which helps the formed components to achieve multi-axis rotation; when the liquid to be tested flows from the narrow section 23 of the third flow channel to the wide section 33 of the third flow channel, the change in flow channel width from narrow to wide will slow down the flow rate of the liquid to be tested; when the liquid to be tested flows from the wide section 33 of the third flow channel to the next narrow section of the flow channel, the change in flow channel width from wide to narrow will accelerate the flow rate of the liquid to be tested; the outlet 43 of the third flow channel is the outlet of the liquid to be tested flowing out of the disturbance flip chip.

[0035] In one embodiment, a turbulence-reversing chip with a rectangular cross-sectional shape in the narrow section of a variable-diameter flow channel is used. Figure 6As shown in the cross-sectional schematic diagram, the turbulence-reversing chip in this embodiment includes a fourth flow channel inlet 14, a fourth flow channel outlet 44, and a fourth variable diameter section 54, and the fourth variable diameter section 54 includes a fourth flow channel narrow section 24 and a fourth flow channel wide section 34. In this embodiment, the width of the fourth channel inlet 14 is wider than the width of the fourth channel narrow section 24. When the liquid to be tested flows from the fourth channel inlet 14 to the first channel narrow section 21, the narrowing of the channel width will accelerate the flow rate of the liquid to be tested. The fourth channel inlet 14 serves as the inlet for the liquid to be tested to flow into the turbulence flip chip. The cross-sectional shape of the fourth channel narrow section 24 is rectangular, and its structure is the simplest compared to the above embodiment. It is easy to manufacture with high precision using standard micromachining processes, which can significantly reduce the processing difficulty and cost of the turbulence flip chip. When the liquid to be tested flows from the fourth channel narrow section 24 to the fourth channel wide section 34, the widening of the channel width will slow down the flow rate of the liquid to be tested. When the liquid to be tested flows from the fourth channel wide section 34 to the next channel narrow section, the narrowing of the channel width will accelerate the flow rate of the liquid to be tested. The fourth channel outlet 44 serves as the outlet for the liquid to be tested to flow out of the turbulence flip chip.

[0036] Please see Figure 7 This embodiment also provides a urine formed element classification method, applied to a urine formed element classification system as described in any of the above embodiments, comprising: step S1: mixing a urine sample with a preset flow rate of sheath fluid to form a test flow; step S2: flipping the test flow to form a flipped flow; step S3: acquiring a flow image time sequence based on the flipped flow, and extracting multi-angle morphological information of the target formed elements in the flipped flow based on the flow image time sequence; step S4: classifying according to the multi-angle morphological information to obtain the classification result corresponding to the target formed elements.

[0037] Further, the step of obtaining a time sequence of liquid flow images based on the flipped liquid flow, and extracting multi-angle morphological information of the target formed elements in the flipped liquid flow based on the time sequence of liquid flow images, includes: taking the frame in which the target formed element first appears in the time sequence of liquid flow images as the first frame, and obtaining the first position corresponding to the target formed element based on the liquid flow image corresponding to the first frame in the time sequence of liquid flow images; taking the first position as the current position, and taking the first frame as the current frame; obtaining the acquisition frame rate corresponding to the time sequence of liquid flow images, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; integrating the first frame, the first position, and several motion positions to obtain a target position time sequence; extracting a set of local images of formed elements from the time sequence of liquid flow images based on the target position time sequence, and obtaining multi-angle morphological information based on the set of local images of formed elements.

[0038] Further, in the step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions, the morphological position acquisition step includes: obtaining the current flow velocity corresponding to the current position, and obtaining the current moving distance based on the acquisition frame rate and the current flow velocity; generating an initial motion position based on the current moving distance and the current position; confirming that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then performing the motion position acquisition step; otherwise, stopping the execution of the morphological position acquisition step, and obtaining several motion positions corresponding to each motion position acquisition step; the motion position acquisition step includes: extracting the fluid flow image corresponding to the next frame from the time sequence of the fluid flow image as a subsequent fluid flow image based on the next frame corresponding to the current frame; filtering based on the subsequent fluid flow image and the initial motion position to obtain motion positions; using the motion position as the current position, using the next frame as the current frame, and re-executing the morphological position acquisition step.

[0039] Further, the step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: obtaining a set of formed component positions with the initial motion position as the center within a preset search radius based on the subsequent fluid flow image; generating an error distance corresponding to each formed component position in the set of formed component positions based on the initial motion position and the set of formed component positions; and taking the formed component position corresponding to the smallest error distance as the motion position.

[0040] In the above embodiments, the organic components in the test liquid flow are flipped to form a flipped liquid flow, enabling the acquisition of multi-angle morphological information containing the target formed components at different angles, and classification based on this multi-angle morphological information. Compared with traditional methods that classify based on static image information from a single angle, this embodiment utilizes the multi-angle morphological information of formed components for classification, providing more comprehensive morphological information for organic component classification, thereby improving the accuracy of organic component classification.

[0041] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A urine formation classification system, characterized in that, include: The sample injection module is used to mix urine samples with a preset flow rate of sheath fluid to form the test fluid stream; A turbulence-reversing chip is used to reverse the liquid flow under test, forming a reversed liquid flow. The morphological information acquisition module is used to acquire a time sequence of liquid flow images based on the flipped liquid flow, and to extract multi-angle morphological information of the target formed elements in the flipped liquid flow based on the time sequence of liquid flow images. A formed element classification module is used to classify based on the multi-angle morphological information to obtain the classification result corresponding to the target formed element; The turbulence-reversing chip includes several variable diameter sections. When the liquid to be tested flows through the narrow section of the variable diameter section, the flow velocity increases, and when the liquid to be tested flows through the wide section of the variable diameter section, the flow velocity decreases. This causes periodic flow velocity disturbances when the liquid to be tested flows through several variable diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow. The constituent elements in the tumbling liquid flow are in a continuous tumbling state.

2. The urine formation sorting system according to claim 1, characterized in that, The morphology information acquisition module is used to acquire a time sequence of liquid flow images based on the flipped liquid flow, and to extract multi-angle morphology information of the target formed elements in the flipped liquid flow based on the time sequence of the liquid flow images. The extraction of multi-angle morphology information of the target formed elements in the flipped liquid flow based on the time sequence of the liquid flow images includes: The first frame in which the target formed element first appears in the liquid flow image time sequence is obtained, and the first position corresponding to the target formed element is obtained based on the liquid flow image corresponding to the first frame in the liquid flow image time sequence. Take the first position as the current position and the first frame as the current frame; Obtain the acquisition frame rate corresponding to the time sequence of the fluid flow image, and perform a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; Based on the first frame, the first position, and several motion positions, a target position time sequence is obtained; Based on the target location time sequence, a set of local images of formed elements is extracted from the fluid flow image time sequence, and the set of local images of formed elements is used as multi-angle morphological information.

3. A urine formation sorting system according to claim 2, characterized in that, The step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain a plurality of motion positions, includes: Obtain the current flow rate corresponding to the current location, and based on the acquisition frame rate and the current flow rate, obtain the current moving distance; Based on the current movement distance and the current position, an initial movement position is generated; If it is confirmed that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then the motion position acquisition step is executed; otherwise, the morphological position acquisition step is stopped, and several motion positions corresponding to each motion position acquisition step are obtained. The motion position acquisition step includes: Based on the next frame corresponding to the current frame, extract the fluid flow image corresponding to the next frame from the fluid flow image time sequence as the subsequent fluid flow image; The motion position is obtained by filtering based on the subsequent fluid flow image and the initial motion position; The motion position is taken as the current position, the next frame is taken as the current frame, and the shape position acquisition step is re-executed.

4. A urine formation sorting system according to claim 3, characterized in that, The step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: Based on the subsequent fluid flow image, a set of formed component positions is obtained with the initial motion position as the center within a preset search radius; Based on the initial motion position and the set of formed component positions, generate the error distance corresponding to each formed component position in the set of formed component positions; The position of the formed element corresponding to the smallest error distance is taken as the motion position.

5. A urine formation sorting system according to claim 1, characterized in that, The formed element classification module is used to classify based on the multi-angle morphological information to obtain the classification result corresponding to the target formed element, including: Based on the multi-angle morphological information, the three-dimensional information of the target's morphological components is extracted; Based on the three-dimensional information of the target formed elements, the classification result corresponding to the target formed elements is obtained by comparing it with a preset formed element classification library.

6. A urine formation sorting system according to claim 1, characterized in that, Among them The turbulence-reversing chip includes several variable-diameter sections. When the liquid to be tested flows through the narrow sections of the variable-diameter sections, the flow velocity increases, and when it flows through the wide sections of the variable-diameter sections, the flow velocity decreases. This causes periodic flow velocity disturbances when the liquid to be tested flows through the variable-diameter sections, thereby causing the formed elements in the liquid to be tested to continuously reverse, forming the reversed liquid flow. The formed elements in the reversed liquid flow are in a continuous reversing state, and the cross-sectional shape of the narrow section of the variable-diameter section is triangular, trapezoidal, arc-shaped, or rectangular.

7. A method for classifying urine by its formation components, characterized in that, An application to a urine formation sorting system as described in any one of claims 1 to 6, comprising: The urine sample is mixed with a pre-set flow rate of sheath fluid to form the test fluid stream; The liquid flow to be tested is reversed to form a reversed liquid flow; Based on the flipped liquid flow, a time sequence of liquid flow images is obtained, and based on the time sequence of liquid flow images, multi-angle morphological information of the target formed elements in the flipped liquid flow is extracted; Based on the multi-angle morphological information, classification results are obtained for the formed elements of the target.

8. A urine formation classification method according to claim 7, characterized in that, The step of obtaining a time-series sequence of liquid flow images based on the flipped liquid flow, and extracting multi-angle morphological information of target formed elements in the flipped liquid flow based on the time-series sequence of liquid flow images, includes: The first frame in which the target formed element first appears in the liquid flow image time sequence is taken as the first frame, and the first position corresponding to the target formed element is obtained based on the liquid flow image corresponding to the first frame in the liquid flow image time sequence. Take the first position as the current position and the first frame as the current frame; Obtain the acquisition frame rate corresponding to the time sequence of the fluid flow image, and perform a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain several motion positions; By integrating the first frame, the first position, and several motion positions, a target position time sequence is obtained; Based on the target location time sequence, a set of local images of formed components is extracted from the fluid flow image time sequence, and multi-angle morphological information is obtained based on the set of local images of formed components.

9. A urine formation classification method according to claim 8, characterized in that, The step of obtaining the acquisition frame rate corresponding to the time sequence of the fluid flow image, and performing a morphological position acquisition step based on the acquisition frame rate, the current position, and the current frame to obtain a plurality of motion positions, includes: Obtain the current flow rate corresponding to the current location, and based on the acquisition frame rate and the current flow rate, obtain the current moving distance; Based on the current movement distance and the current position, an initial movement position is generated; If it is confirmed that the initial motion position does not exceed the end position of the flow channel corresponding to the time sequence of the fluid flow image, then the motion position acquisition step is executed; otherwise, the morphological position acquisition step is stopped, and several motion positions corresponding to each motion position acquisition step are obtained. The motion position acquisition step includes: Based on the next frame corresponding to the current frame, extract the fluid flow image corresponding to the next frame from the fluid flow image time sequence as the subsequent fluid flow image; The motion position is obtained by filtering based on the subsequent fluid flow image and the initial motion position; The motion position is taken as the current position, the next frame is taken as the current frame, and the shape position acquisition step is re-executed.

10. A urine formation classification method according to claim 9, characterized in that, The step of filtering based on the subsequent fluid flow image and the initial motion position to obtain the motion position includes: Based on the subsequent fluid flow image, a set of formed component positions is obtained with the initial motion position as the center within a preset search radius; Based on the initial motion position and the set of formed component positions, generate the error distance corresponding to each formed component position in the set of formed component positions; The position of the formed element corresponding to the smallest error distance is taken as the motion position.

Images