An intelligent control method, device, equipment, medium and product of functional microspheres
By improving the YOLOv11 target detection model and population path planning algorithm, intelligent manipulation of functional microspheres is achieved, solving the challenges of identification, manipulation, and delivery in the living microenvironment. This improves the accuracy of microsphere identification and delivery efficiency, making it suitable for targeted therapy in complex biological environments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JINAN UNIVERSITY
- Filing Date
- 2026-04-29
- Publication Date
- 2026-07-24
AI Technical Summary
Existing technologies struggle to achieve efficient targeted delivery of functional microspheres in living microenvironments, presenting challenges in identification, manipulation, and delivery, resulting in low delivery efficiency and success rates.
An improved YOLOv11 target detection model is used for multi-scale feature fusion and detection. Combined with an optical tweezers system and a group path planning algorithm, automated and intelligent manipulation of functional microspheres is achieved, including identification, coordinate transformation and obstacle avoidance path planning.
It significantly improves the recognition accuracy and delivery efficiency of functional microspheres, enabling rapid and precise delivery of multiple particles in complex biological environments, and is suitable for targeted therapy in complex microscopic environments.
Smart Images
Figure CN122449934A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of biomedical engineering and intelligent control, and in particular to a method, apparatus, equipment, medium and product for intelligent manipulation of functional microspheres. Background Technology
[0002] Targeted delivery of functional microspheres is a cutting-edge area in precision medicine. However, achieving efficient delivery of nanoscale particles in living microenvironments (such as blood vessels) faces multiple challenges: (1) Difficult to identify: The functional microspheres are extremely small in size and usually appear as spots of only a few pixels in microscopic images. They are easily confused with background noise, and traditional visual recognition methods have low accuracy.
[0003] (2) Difficult to control: There are differences in spatial scale and mapping relationship between the image coordinates under the microscope and the actual physical operation coordinates (such as optical tweezers control coordinates), which require high-precision coordinate transformation.
[0004] (3) Delivery difficulty: There are a large number of dynamic obstacles such as red blood cells and white blood cells in the blood vessels, and the delivery path is complex. The control system is required to have real-time obstacle avoidance and path planning capabilities to achieve fast and accurate targeted delivery.
[0005] Existing technologies typically struggle to address these issues simultaneously, resulting in low delivery efficiency and low success rates, which limits the actual therapeutic efficacy of functional microspheres. Summary of the Invention
[0006] The purpose of this application is to provide an intelligent control method, device, equipment, medium, and product for functional microspheres, which can realize automated, intelligent, and high-precision control of functional microspheres, and improve the stability and accuracy of microsphere control.
[0007] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for intelligently manipulating functional microspheres, comprising: Microscopic images containing functional microspheres are acquired, and a target detection model is used to identify the functional microspheres in the microscopic images to obtain the image coordinates of the functional microspheres. The target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling fusion and detection on the high-level features output by the backbone network, the detection of small and fine-grained targets is achieved. The image coordinates of the functional microsphere are converted into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere; Based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of the obstacle cells, and the coordinates of the blood vessel boundary, a population path planning algorithm is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the coordinates of the preset target point. The optical tweezers system is controlled to perform the capture and delivery of functional microspheres according to the optimal obstacle avoidance path.
[0008] Secondly, this application provides an intelligent control device for functional microspheres, comprising: The identification module is used to acquire microscopic images containing functional microspheres and to identify the functional microspheres in the microscopic images using a target detection model to obtain the image coordinates of the functional microspheres. The target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling fusion and detection on the high-level features output by the backbone network, it can detect small targets and fine-grained targets. The coordinate transformation module is used to convert the image coordinates of the functional microsphere into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere. The dynamic obstacle avoidance and delivery module is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the preset target coordinates based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of the obstacle cells and the blood vessel boundary, using a population path planning algorithm, and to control the optical tweezers system to perform capture and delivery operations on the functional microsphere according to the optimal obstacle avoidance path.
[0009] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize the above-described intelligent manipulation method for microspheres.
[0010] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described intelligent manipulation method for microspheres.
[0011] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described intelligent control method for microspheres.
[0012] According to the specific embodiments provided in this application, the present application achieves the following technical effects: By improving the YOLOv11 target detection model and combining optical tweezers with path planning, intelligent and precise manipulation of functional microspheres is achieved. First, a multi-scale feature fusion and detection branch is added to the model, and high-level features are upsampled and fused a second time, significantly improving the recognition accuracy and completeness of small-sized, fine-grained functional microspheres, solving the problems of missed detection and inaccurate positioning in traditional detection. Second, the image coordinates are accurately converted into the physical space coordinates of the optical tweezers, realizing the quantitative mapping of the microsphere position and providing a reliable data foundation for subsequent manipulation. At the same time, by integrating the microsphere coordinates, target points, obstacles, and blood vessel boundary information, a group path planning algorithm is used to generate the optimal obstacle avoidance path, which can effectively avoid intravascular obstacle cells and boundary restrictions, ensuring a safe and efficient delivery path. Finally, the capture and delivery are completed by optical tweezers according to the planned path, realizing automated, intelligent, and high-precision manipulation of functional microspheres, improving the stability and accuracy of microsphere manipulation, and is especially suitable for microsphere targeting operations in complex microscopic environments. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0014] Figure 1 This is a flowchart illustrating an intelligent control method for functional microspheres provided in an embodiment of this application.
[0015] Figure 2 This is a schematic diagram of the target detection model in one embodiment of this application.
[0016] Figure 3 This is a comparison chart of the detection accuracy of the target detection model and the original YOLOv11 in one embodiment of this application.
[0017] Figure 4 This is a schematic diagram of the operational logic of a group path planning algorithm in one embodiment of this application.
[0018] Figure 5 This is a schematic diagram of the functional modules of an intelligent control device for functional microspheres provided in an embodiment of this application.
[0019] Figure 6 This is a schematic diagram illustrating the effect of capturing microspheres in an aqueous solution in one embodiment of this application.
[0020] Figure 7 This is a schematic diagram of a functional microsphere for identification in a mixed solution according to one embodiment of this application.
[0021] Figure 8 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation
[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0023] This application provides an intelligent control method that can automatically, accurately, and efficiently complete the entire process of functional microsphere identification, capture, and delivery in complex biological environments. It aims to solve the technical problem of accurately identifying, capturing, and efficiently delivering small targets such as functional microspheres in complex biological environments (such as living blood vessels) in the prior art.
[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0025] In one exemplary embodiment, such as Figure 1 As shown, an intelligent control method for functional microspheres is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the intelligent control method for functional microspheres includes steps 101 to 104.
[0026] Step 101: Acquire a microscopic image containing functional microspheres, and use a target detection model to identify the functional microspheres in the microscopic image to obtain the image coordinates of the functional microspheres.
[0027] The proposed target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling, fusion, and detection on the high-level features output by the backbone network, it achieves the detection of small and fine-grained targets. Specifically, the multi-scale feature fusion and detection branch, through denser feature map sampling and more refined anchor box settings, is specifically designed to enhance the feature capture of pixel-level functional microspheres in the image, thereby significantly improving recognition accuracy.
[0028] The target detection model is pre-trained on a dataset containing images of in vitro and in vivo functional microspheres with data augmentation processing (geometric transformation and random cropping) to improve the recognition accuracy of functional microspheres.
[0029] In a specific application example, such as Figure 2As shown, the multi-scale feature fusion and detection branch includes an upsampling layer, a first connection layer, a first C3k2 module, a convolutional layer, a second connection layer, and a second C3k2 module connected in sequence; the multi-scale feature fusion and detection branch also includes a micro-target detection head. The micro-target detection head is connected to the convolutional layer.
[0030] The upsampling layer is also connected to the second C3k2 module in the neck network of YOLOv11. The first connection layer is also connected to the first C3k2 module in the backbone network of YOLOv11. The second connection layer is also connected to the second C3k2 module in the neck network of YOLOv11. The second C3k2 module is also connected to the first detection head and the first convolutional layer in the neck network of YOLOv11.
[0031] The process of identifying functional microspheres in microscopic images using a target detection model includes the following steps: First, the microscopic image enters the target detection model through the input layer. The Conv and C3k2 modules extract low-level features such as basic edges and textures. Then, the SPPF module performs multi-scale feature fusion, and the C2PSA module enhances spatial attention, automatically focusing on the functional microsphere region and weakening background noise and biological interference. Subsequently, a high-resolution multi-scale feature pyramid is constructed using Upsample and Concat. The enhanced feature map is then input into the multi-scale feature fusion and detection branch, which connects to the shallow layers of the network. This branch provides higher resolution and richer feature information, combined with smaller anchor boxes with aspect ratios more suitable for speckled targets like functional microspheres. It also receives deeper semantic information, thus determining that the small specks are functional microspheres rather than noise. Finally, the pixel coordinates, category confidence, and bounding box information of the functional microspheres in the image are output, providing high-precision data support for subsequent coordinate transformation and optical tweezers manipulation. The target detection model used in this application is compared with the detection accuracy of the original YOLOv11. Figure 3 As shown, the original YOLOv11 has no correct data and cannot detect the target microsphere.
[0032] Step 102: Convert the image coordinates of the functional microsphere into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere.
[0033] In a specific application example, the image coordinates of the functional microsphere are converted into the physical space coordinates of the optical tweezers system through affine transformation to obtain the current coordinates of the functional microsphere, realizing a seamless connection from visual recognition to physical manipulation. At the same time, the converted physical coordinates are written into a standardized text file in real time for the optical tweezers system to call.
[0034] Specifically, first determine the position of the origin of the ROI coordinate system in the image pixel coordinate system: ; ; in, The x-coordinate of the upper left corner point in the microscope's field of view. The y-coordinate of the point at the top left corner of the microscope field of view. The x-coordinate of the lower right corner point in the microscope's field of view. This represents the y-coordinate of the lower right corner point in the microscope's field of view. The x-coordinate of the center point in the microscope's field of view. The y-axis coordinates of the center point under the microscope's field of view are given.
[0035] Then, the following formula is used to convert it to an ROI coordinate system with the center point of the ROI as the origin: ; ; in, For the functional microsphere image coordinates, The current coordinates of the functional microsphere. The x-axis coordinates of the center point of the functional microsphere returned by the target detection model. The y-axis coordinates of the center point of the functional microsphere returned by the target detection model. Here are the x-axis coordinates of the functional microspheres under the microscope's field of view. The y-axis coordinates of the functional microspheres are shown in the microscope field of view.
[0036] Step 103: Based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of obstacle cells (such as red blood cells and white blood cells), and the coordinates of the blood vessel boundary, a population path planning algorithm is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the preset target point coordinates. The optimal obstacle avoidance path is a delivery path planned by the optical tweezers system that avoids all obstacle cells and is constrained by the blood vessel boundary, thereby controlling the optical tweezers system to capture the functional microsphere and deliver it quickly and non-destructively to the target point along the planned path.
[0037] In a specific application example, this application adopts an improved A Group path planning algorithm, improved A Group path planning algorithms in traditional A Based on the swarm path planning algorithm, a cost function dynamically updated for multiple obstacles is introduced, and it supports planning conflict-free parallel delivery paths for multiple functional microspheres. Path calculation can be completed within seconds, and path planning for multiple functional microspheres can be performed simultaneously. For example... Figure 4 As shown, step 103 includes steps 31 to 35.
[0038] Step 31: Calculate the unit vector from the current coordinates of the functional microsphere to the coordinates of the preset target point to obtain the attraction vector, the purpose of which is to guide the functional microsphere toward the preset target point.
[0039] Specifically, using the formula Calculate the attraction vector; where, For the direction of attraction, The current coordinates of the functional microsphere. The coordinates of the preset target point.
[0040] Step 32: Traverse all obstacle cells, calculate the repulsive force from the coordinates of each obstacle cell to the current coordinates of the functional microsphere, and sum all the repulsive forces to obtain the total repulsive force vector. The closer the functional microsphere is to the obstacle cell, the greater the repulsive force. The repulsive forces generated by all obstacle cells will be summed to form the total repulsive force vector.
[0041] Specifically, using the formula Calculate the total repulsive force vector; where, The total repulsive force vector is... I The number of barrier cells, For the first i The coefficient of each obstacle cell, , r For the first i The radius of each obstacle cell To establish a safe distance, The current coordinates of the functional microsphere and the first i The actual distance between each obstacle cell; the closer the distance, the better. The larger the value, the stronger the repulsive force. For the first i The coordinates of each obstacle cell.
[0042] Step 33: The attractive force vector and the total repulsive force vector are weighted and summed, then normalized to obtain a unit movement direction vector. Specifically, the summation is performed according to preset weights (e.g., the weight of the attractive force vector is 0.3, and the weight of the total repulsive force vector is 0.5) to obtain the resultant force vector. Since the length of the resultant force vector is not necessarily 1, it needs to be normalized to convert it into a unit direction vector, which serves as the final direction of the particle's current movement. The formula is: , ;in, The resultant force vector, The weights of the attraction vector, The weights of the total repulsive force vector are... The direction vector for unit movement.
[0043] Step 34: Calculate the new position of the functional microsphere based on the unit movement direction vector and the fixed step size. ;in, For the new location, For fixed step size.
[0044] Step 35: Determine whether the new position will collide with the obstacle cell (i.e., whether the new position falls within the obstacle cell region) and whether the new position is within the blood vessel boundary coordinates. If the new position will collide with the obstacle cell or the new position is not within the blood vessel boundary coordinates, then keep the current coordinates of the functional microsphere unchanged; otherwise, update the current coordinates of the functional microsphere to the new position until the current coordinates of the functional microsphere coincide with the coordinates of the preset target point, thus obtaining the optimal obstacle avoidance path.
[0045] Regardless of whether the current coordinates of the functional microsphere change, its current coordinates (either the updated new position or the original position) will be recorded as a waypoint in the motion trajectory list. The entire simulation cycle ends when the functional microsphere reaches the preset target point (or reaches the maximum number of simulation steps). All recorded waypoints constitute the complete motion trajectory.
[0046] Step 104: Control the optical tweezers system to perform the capture and delivery operation of the functional microspheres according to the optimal obstacle avoidance path.
[0047] This application significantly improves the recognition accuracy and delivery efficiency of functional microspheres, overcomes the interference of complex obstacles in blood vessels, and achieves parallel, rapid, and precise automated delivery of multiple particles, providing an efficient and reliable technical solution for targeted therapy.
[0048] Based on the same inventive concept, this application also provides an apparatus for implementing the method described above. The solution provided by this apparatus is similar to the solution described in the above method; therefore, specific limitations in one or more apparatus embodiments provided below can be found in the limitations of the method described above, and will not be repeated here.
[0049] In one exemplary embodiment, such as Figure 5 As shown, an intelligent control device for a functional microsphere is provided, comprising: an identification module 501, a coordinate transformation module 502, and a dynamic obstacle avoidance delivery module 503.
[0050] The identification module 501 is used to acquire microscopic images containing functional microspheres and to identify the functional microspheres in the microscopic images using a target detection model to obtain the image coordinates of the functional microspheres. The target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling fusion and detection on the high-level features output by the backbone network, it achieves the detection of small and fine-grained targets.
[0051] In a specific application example, the recognition module 501 is an improvement upon the YOLOv11 model architecture. First, an initial dataset of functional microspheres was constructed, containing 2000 in vitro sample images and 1000 in vivo sample images. Through data augmentation techniques such as geometric transformation and random cropping, the dataset was expanded to 5000 labeled images for model training. To address the low recognition rate of small targets (functional microspheres) in the original model, a dedicated micro-target detection head was added to the model's detection head. This micro-target detection head enhances the extraction and classification capabilities of weak target features in images by designing denser feature map sampling and finer anchor boxes. After multiple rounds of hyperparameter optimization training, the recognition accuracy (e.g., mAP) of functional microspheres was ultimately improved from approximately 40% to over 70%.
[0052] The coordinate transformation module 502 is used to convert the image coordinates of the functional microsphere into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere.
[0053] In a specific application example, the core of the coordinate transformation module 502 lies in constructing an accurate affine transformation matrix. This affine transformation matrix is obtained through calibration using a calibration plate or markers of known size, and it accurately maps the coordinates of the center point of the functional microsphere bounding box in the image coordinate system (in pixels) to the real physical world coordinate system (in micrometers), which is consistent with the operating plane of the optical tweezers. The transformed coordinates are written in real time to a standardized text file (TXT format) for downstream optical tweezers systems to read.
[0054] The dynamic obstacle avoidance delivery module 503 is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the preset target coordinates based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of the obstacle cells and the blood vessel boundary, using a population path planning algorithm, and to control the optical tweezers system to perform capture and delivery operations on the functional microsphere according to the optimal obstacle avoidance path.
[0055] In a specific application example, the dynamic obstacle avoidance delivery module 503 is the "decision center" of the device. During each delivery task, the dynamic obstacle avoidance delivery module receives a real-time data stream from the identification module: the coordinates of the functional microsphere to be manipulated (starting point), the preset potential trap position (target point) of the treatment target, a list of coordinates of all non-target cells identified within the same field of view (as dynamic obstacles), and the boundary coordinates of the blood vessel wall (as an impassable area). Improved A Group path planning algorithm, in A Based on the swarm path planning algorithm, a computational function for dynamic updates of multiple obstacles is introduced, and a path conflict avoidance strategy is considered when optical tweezers simultaneously manipulate multiple particles. The algorithm can plan a safe and efficient path that satisfies all constraints in a very short time (e.g., within seconds). According to this path instruction, the optical tweezers system moves its generated optical trap, first capturing the target functional microsphere, and then smoothly and quickly transporting it to the target location along the planned path. The entire process is completed automatically.
[0056] This application achieves intelligent, precise, and efficient delivery of functional microspheres in complex in vivo environments through collaborative innovation in three key aspects: identification, coordinate mapping, and path planning, providing a powerful technical tool for targeted therapy of diseases.
[0057] Two application examples are provided below: Application Example 1: Functionalized microspheres to be identified are added to an aqueous solution and uniformly distributed using a stirrer, resulting in a mixed solution of functionalized microspheres. The improved YOLOv11 model is used to identify the functionalized microspheres in the aqueous solution, and the coordinates of the microspheres are obtained through a coordinate mapping algorithm, which is then used to capture them with optical tweezers. The detection results are as follows: Figure 6 As shown, the captured particles identified by green and blue fluorescence verification are indeed the desired target microspheres. After successful capture of the functionalized microspheres, a target location A was pre-defined in the field of view. The group path planning algorithm calculates the optimal path and then uses an optical tweezers system to deliver it to the designated location. The final result is as follows: Figure 6 As shown, the functional microsphere successfully reached the target position, and the path was the shortest path.
[0058] Application Example 2: The functional microspheres were thoroughly mixed with a zebrafish blood solution, such as... Figure 7As shown, the mixed solution contains not only target microspheres but also red blood cells and platelets. Using the improved YOLOv11 model to identify objects in the field of view, the target microspheres and cells were effectively distinguished. Subsequently, using the coordinates of the target microspheres as the starting position, the coordinates of the platelets as the target position, and the red blood cells as obstacles, a path planning algorithm was used to plan the optimal path, successfully delivering the target microspheres around the obstacles to the target location. This process demonstrates the device's precise control and therapeutic potential.
[0059] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 8 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores microscopic images of the functional microspheres. The I / O interfaces allow the processor to exchange information with external devices. The communication interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an intelligent manipulation method for the functional microspheres.
[0060] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment to which the present application is applied. Specific computer equipment may include, for example, [the following is a list of possible additional structures]. Figure 8 The diagram shows more or fewer components, or combinations of certain components, or different component arrangements.
[0061] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0062] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0063] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0064] In this application, all actions to acquire signals, information, or data are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with the authorization granted by the owner of the relevant device.
[0065] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).
[0066] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, etc., and are not limited to these.
[0067] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0068] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for intelligently manipulating functional microspheres, characterized in that, include: Microscopic images containing functional microspheres are acquired, and a target detection model is used to identify the functional microspheres in the microscopic images to obtain the image coordinates of the functional microspheres. The target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling fusion and detection on the high-level features output by the backbone network, the detection of small and fine-grained targets is achieved. The image coordinates of the functional microsphere are converted into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere; Based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of the obstacle cells, and the coordinates of the blood vessel boundary, a population path planning algorithm is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the coordinates of the preset target point. The optical tweezers system is controlled to perform the capture and delivery of functional microspheres according to the optimal obstacle avoidance path.
2. The intelligent control method for functional microspheres according to claim 1, characterized in that, The multi-scale feature fusion and detection branch includes an upsampling layer, a first connection layer, a first C3k2 module, a convolutional layer, a second connection layer, and a second C3k2 module connected in sequence; the multi-scale feature fusion and detection branch also includes a micro-target detection head; the micro-target detection head is connected to the convolutional layer; The upsampling layer is also connected to the second C3k2 module in the neck network of YOLOv11. The first connection layer is also connected to the first C3k2 module in the backbone network of YOLOv11. The second connection layer is also connected to the second C3k2 module in the neck network of YOLOv11. The second C3k2 module is also connected to the first detection head and the first convolutional layer in the neck network of YOLOv11.
3. The intelligent control method for functional microspheres according to claim 1, characterized in that, Converting the image coordinates of the functional microspheres into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microspheres includes: The image coordinates of the functional microsphere are converted into the physical space coordinates of the optical tweezers system through affine transformation to obtain the current coordinates of the functional microsphere.
4. The intelligent control method for functional microspheres according to claim 1, characterized in that, Based on the current coordinates of the functional microspheres, the coordinates of the preset target point, the coordinates of the obstacle cells, and the coordinates of the blood vessel boundary, a population path planning algorithm is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microspheres to the preset target coordinates, including: Calculate the unit vector from the current coordinates of the functional microsphere to the coordinates of the preset target point to obtain the attraction vector; Traverse all obstacle cells, calculate the repulsive force from the coordinates of each obstacle cell to the current coordinates of the functional microsphere, and sum all the repulsive forces to obtain the total repulsive force vector; The attractive force vector and the total repulsive force vector are weighted and summed, and then normalized to obtain the unit movement direction vector; The new position of the functional microsphere is calculated based on the unit movement direction vector and the fixed step size. Determine whether the new position will collide with the obstacle cell and whether the new position is within the coordinates of the blood vessel boundary. If the new position will collide with the obstacle cell or the new position is not within the coordinates of the blood vessel boundary, then keep the current coordinates of the functional microsphere unchanged; otherwise, update the current coordinates of the functional microsphere to the new position until the current coordinates of the functional microsphere coincide with the coordinates of the preset target point, and obtain the optimal obstacle avoidance path.
5. The intelligent control method for functional microspheres according to claim 4, characterized in that, Using formula Calculate the attraction vector; where, For the direction of attraction, The current coordinates of the functional microsphere. The coordinates of the preset target point.
6. The intelligent control method for functional microspheres according to claim 4, characterized in that, Using formula Calculate the total repulsive force vector; where, The total repulsive force vector is... I The number of barrier cells, For the first i The coefficient of each obstacle cell, , r For the first i The radius of each obstacle cell To establish a safe distance, The current coordinates of the functional microsphere and the first i The actual distance between each obstacle cell For the first i The coordinates of each obstacle cell These are the current coordinates of the functional microsphere.
7. An intelligent control device for functional microspheres, executing the intelligent control method for functional microspheres according to any one of claims 1-6, characterized in that, The device includes: The identification module is used to acquire microscopic images containing functional microspheres and to identify the functional microspheres in the microscopic images using a target detection model to obtain the image coordinates of the functional microspheres. The target detection model adds a multi-scale feature fusion and detection branch to the YOLOv11 target detection model. By performing secondary upsampling fusion and detection on the high-level features output by the backbone network, it can detect small targets and fine-grained targets. The coordinate transformation module is used to convert the image coordinates of the functional microsphere into the physical space coordinates of the optical tweezers system to obtain the current coordinates of the functional microsphere. The dynamic obstacle avoidance and delivery module is used to determine the optimal obstacle avoidance path from the current coordinates of the functional microsphere to the preset target coordinates based on the current coordinates of the functional microsphere, the coordinates of the preset target point, the coordinates of the obstacle cells and the blood vessel boundary, using a population path planning algorithm, and to control the optical tweezers system to perform capture and delivery operations on the functional microsphere according to the optimal obstacle avoidance path.
8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the intelligent manipulation method of the functional microspheres according to any one of claims 1-6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent manipulation method of the functional microspheres as described in any one of claims 1-6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent manipulation method of the functional microspheres as described in any one of claims 1-6.