First-aid hemostasis robot and control method thereof

Abstract

The invention provides a first-aid hemostasis robot and a control method thereof.The first-aid hemostasis robot comprises a robot body capable of walking on the ground and a mechanical arm which is arranged on the robot body and has multiple degrees of freedom, the mechanical arm is provided with a pressing head and a plugging piece, and the pressing head is used for conducting pressing hemostasis on blood vessels of a bleeding opening; the plugging piece is carried by a flexible continuum robot, the front end of the continuum robot is provided with a camera used for collecting information of a bleeding part, and the plugging piece can stretch into a bleeding opening through the continuum robot and expand after plugging the bleeding opening so as to seal a bleeding large blood vessel. The mechanical arm carries the pressing head and the plugging piece to stop bleeding of massive hemorrhage under different conditions, if the hemorrhage is split hemorrhage, pressing hemostasis is carried out at the proximal end of corresponding arterial hemorrhage or the distal end of corresponding venous hemorrhage through the pressing head, and if the hemorrhage is fractured hemorrhage, the plugging piece extends into a wound through the continuum robot, so that the hemorrhage is stopped. And sealing bleeding large blood vessels.

Description

Technical Field

[0001] This invention belongs to the field of medical auxiliary equipment technology, specifically relating to an emergency hemostasis robot and its control method. Background Technology

[0002] Massive bleeding from external injuries can easily lead to hemorrhagic shock, and many patients lose their chance of treatment before reaching a hospital. Moreover, when the amount of blood loss exceeds 40% of the total blood volume, it can be life-threatening. Therefore, hemostasis is an extremely important measure in first aid and must be performed quickly, accurately, and effectively.

[0003] With the development of science and technology, various hemostatic devices have emerged on the market. For massive bleeding, most hemostatic devices use compression to stop the bleeding, such as the emergency tourniquet disclosed in CN201210442160.2, the hemostatic clip disclosed in CN202111660284.3, and the rapid hemostatic device and tourniquet disclosed in CN202211496132.9. These methods of compressing blood vessels are effective for stopping bleeding from lacerations where the blood vessel is torn but not severed, but less effective for bleeding from ruptured blood vessels. Furthermore, when there is heavy traffic congestion on the way to the scene, ambulances have difficulty approaching the scene, requiring emergency personnel to carry stretchers and rescue equipment, wasting valuable time and reducing rescue efficiency. Summary of the Invention

[0004] To address the technical problems existing in the prior art, the first aspect of the present invention is to provide an emergency hemostasis robot. The second aspect, based on the same inventive concept, also provides a control method based on the aforementioned emergency hemostasis robot.

[0005] In this embodiment of the invention, the emergency hemostasis robot includes a robot body capable of walking on the ground and a robotic arm with multiple degrees of freedom mounted on the robot body. The robotic arm has a pressing head and a sealing component. The pressing head is used to press and stop bleeding at the blood vessels of the bleeding point. The sealing component is carried by a flexible continuous robot. The front end of the continuous robot has a camera for collecting information about the bleeding site. The sealing component can be inserted into the bleeding point through the continuous robot and expands after blocking the bleeding point to seal the large bleeding blood vessel.

[0006] The control method of this invention, based on the aforementioned emergency hemostasis robot, includes the following steps: S1, the emergency hemostasis robot arrives at the treatment site; S2, the emergency hemostasis robot acquires an image of the bleeding site of the injured person through medical detection equipment, determines whether it is laceration bleeding or rupture bleeding. If it is laceration bleeding, proceed to step S3; if it is rupture bleeding, proceed to step S4; S3, if it is laceration bleeding, determine whether it is arterial bleeding or venous bleeding, apply pressure with the compression head at the proximal end of the corresponding arterial bleeding to stop the bleeding, and apply pressure with the compression head at the distal end of the corresponding venous bleeding to stop the bleeding; S4, if it is rupture bleeding, the sealing member extends into the wound through the continuous robot along the bleeding opening. The camera at the front end of the continuous robot acquires an image of the bleeding site, determines the bleeding point, locates the major bleeding vessel, and inserts the sealing member into the bleeding opening. After the sealing member blocks the bleeding opening, it expands to seal the major bleeding vessel.

[0007] Compared with the prior art, the advantages of the superior technical solution of the present invention include:

[0008] 1. This invention utilizes a robotic arm carrying a compression head and a sealing device to stop bleeding in different situations of massive hemorrhage. Before stopping the bleeding, medical testing equipment acquires images of the bleeding site to determine whether it is laceration or rupture bleeding. If it is laceration bleeding, the compression head is applied to the proximal end of the corresponding arterial bleeding or the distal end of the corresponding venous bleeding to stop the bleeding. If it is rupture bleeding, a continuous robot inserts the sealing device into the wound to seal the major bleeding vessel. This invention uses different methods to stop bleeding in different types of massive hemorrhage, resulting in faster and more effective hemostasis.

[0009] 2. The support platform of this invention supports the injured person, and the protective device limits and protects the injured person. The emergency hemostasis robot runs on the ground via wheels at the bottom of its support legs. In traffic jams, the support legs are extended and retracted to tilt the support platform along its width direction and reduce the distance between the two support legs, thereby reducing the overall width of the emergency hemostasis robot. This allows the robot to move through the gaps between two vehicles in a traffic jam, achieving the goal of quickly rescuing the injured. The emergency hemostasis robot of this invention is small in size, highly maneuverable, and can move through the gaps between two vehicles in a traffic jam, greatly improving the speed of rescue and passage during traffic jams.

[0010] 3. This invention uses a convolutional neural network (CNN) to accurately extract subtle features from images of bleeding sites, effectively reducing subjective errors in human judgment and improving the accuracy of identifying lacerations and ruptures, especially suitable for complex bleeding scenarios. Simultaneously, the fully automated algorithm execution (from feature extraction to threshold judgment) significantly shortens the judgment time, buying crucial time for emergency treatment. The clearly standardized process reduces reliance on the professional background of operators, facilitating its promotion in non-medical settings, and the probability threshold... It can be flexibly adjusted to adapt to different emergency needs, ultimately providing a reliable decision-making basis for the subsequent matching of targeted hemostasis strategies and the realization of intelligent emergency care for the emergency hemostasis robot, and comprehensively improving the accuracy, efficiency and applicability of emergency care.

[0011] 4. This invention accurately locates active bleeding points through grayscale gradient recognition and activity verification, effectively eliminating grayscale interference from non-bleeding areas and improving the accuracy of bleeding point identification. Furthermore, by combining local optical flow and volume change calculations to determine real-time bleeding rate and total volume, it can quickly pinpoint the critical location with the largest bleeding volume, providing a basis for prioritizing the treatment of life-threatening bleeding. Subsequently, by fusing parameters such as velocity divergence and flow direction using multi-weighted coefficients, the bleeding vessel opening is determined. A control equation guided by the vessel direction drives the precise insertion of the occluder. Combined with position error judgment and stiffness coefficient control of the occlusion force, precise implantation of the occluder and effective hemostasis are achieved. Simultaneously, by cyclically detecting whether the bleeding volume is below a threshold, the hemostasis effect can be dynamically judged and corrected in a timely manner. The overall process balances the accuracy of bleeding location, the accuracy of occlusion operation, and the reliability of hemostasis, significantly improving the efficiency and success rate of emergency hemostasis, and is particularly suitable for the rapid vascular occlusion needs in emergency situations. Attached Figure Description

[0012] Figure 1 This is a three-dimensional structural diagram of the emergency hemostasis robot in this embodiment, with the robotic arm housed in a box at the bottom of the support platform.

[0013] Figure 2 This is a side view of the emergency hemostasis robot in the embodiment along its width. Figure 1 The robotic arm extends above the support platform.

[0014] Figure 3 This is a side view of the emergency hemostasis robot in the embodiment along its width. Figure 2 The support platform is in a horizontal position, and the robotic arm is stored in the box at the bottom of the support platform.

[0015] Figure 4 This is a side view of the emergency hemostasis robot in the embodiment along its width. Figure 3 The support platform is tilted, and the robotic arm is stored in the box at the bottom of the support platform.

[0016] The reference numerals in the accompanying drawings include: 1. support platform; 2. support leg; 3. travel wheel; 4. strap; 5. baffle; 6. counterweight; 7. support wheel; 8. telescopic rod; 9. hinge seat; 10. base; 11. image detection device; 12. distance detection device; 13. vehicle controller; 14. robotic arm; 15. pressing head; 16. sealing component; 17. continuous robot; 18. camera; 19. medical testing equipment; 20. rotating telescopic arm; 21. box. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] Example 1

[0019] This embodiment provides an emergency hemostasis robot, such as Figure 1 and Figure 2 As shown, in a preferred embodiment, the emergency hemostasis robot includes a robot body capable of walking on the ground and a robotic arm 14 with multiple degrees of freedom mounted on the robot body, such as a robotic arm 14 with multiple rotational joints (including rotation in the X, Y and Z directions). The robotic arm 14 has a pressing head 15 and a sealing member 16.

[0020] The pressing head 15 is installed at the end of the robotic arm 14 and is used to press and stop bleeding at the blood vessel at the bleeding point. The occlusion component 16 is carried by a flexible continuum robot 17, which is installed on one segment of the end arm of the robotic arm 14. The front end of the continuum robot 17 has a camera 18 to collect information about the bleeding site, determine the bleeding point, locate the major blood vessel, and the occlusion component 16 can be inserted into the bleeding point through the continuum robot 17. After blocking the bleeding point, it expands to seal the major bleeding blood vessel, achieving rapid hemostasis. Specifically, the occlusion component 16 is an airbag or a hemostatic gel that can expand after absorbing blood. When the occlusion component 17 is an airbag, the airbag is connected to an inflation device through a trachea. The inflation device injects air into the airbag through the trachea to inflate the airbag, and the inflated airbag seals the major bleeding blood vessel.

[0021] like Figures 1-4As shown, in another preferred embodiment of the present invention, the robot body includes a support platform 1, four support legs 2 located at the bottom of the support platform 1 for supporting it, a traveling wheel 3 located at the bottom of each support leg 2 for running on the ground, and a protective device installed on the support platform 1 for limiting and protecting the injured. The support platform 1 can be a stretcher or a support plate, and the traveling wheels 3 are mounted on the bottom of the support legs 2 via a base 10. The traveling wheels 3 are either drive wheels or steering wheels, both existing technologies, and their structure and principles are not detailed here.

[0022] The front end of the carrying platform 1 is equipped with an image detection device 11 for detecting obstacles and a distance detection device 12 for detecting the distance between the front end of the carrying platform 1 and obstacles. The signal output terminals of the image detection device 11 and the distance detection device 12 are connected to the vehicle controller 13, which acquires the distance between two parallel vehicles during traffic jams. The image detection device 11 is a rotatable image sensor mounted on the front end of the carrying platform 1, and the distance detection device 12 is a rotatable distance sensor mounted on the front end of the carrying platform 1. The image sensor and the distance sensor can rotate synchronously and face the same direction to acquire obstacles in different directions and their distances, facilitating a more accurate determination of the distance between two parallel vehicles.

[0023] Four support legs 2 are symmetrically arranged on the left and right sides of the support platform 1 in the width direction. The upper ends of the four support legs 2 are rotatably connected to the support platform 1 through hinge seats 9. The support legs 2 on both sides of the support platform 1 can extend and retract independently to adjust the height of the support legs 2. The two support legs 2 on the left side of the support platform 1 can extend and retract synchronously, and the two support legs 2 on the right side of the support platform 1 can also extend and retract synchronously. The variable pitch output control terminal of the vehicle controller 13 is connected to the extension and retraction enable terminal of the support legs 2. By extending and retracting the support legs 2 on both sides of the support platform 1, the support platform 1 can tilt along its width direction and reduce the distance between the two support legs 2, thereby reducing the width of the entire emergency hemostasis robot, so that the emergency hemostasis robot can pass through the gap between two vehicles in traffic jams.

[0024] In this embodiment, a box 21 with an open end is fixed to the bottom of the support platform 1. Under normal conditions, the multiple degrees of freedom robotic arms 14 are retracted and stored in the box 21, without increasing the width of the emergency hemostasis robot or affecting its passage. When hemostasis is required, the robotic arms 14 are extended, extending from the opening of the box 21, and carrying the pressing head 15 and the sealing component 16, and are positioned above the support platform 1 to facilitate hemostasis of the injured.

[0025] In another preferred embodiment, the robotic arm 14 also carries a medical detection device 19 capable of acquiring images of the bleeding site of the injured person. The medical detection device 19 can acquire images of the bleeding site of the injured person by ultrasound or X-ray. Specifically, the medical detection device 19 can be mounted on one arm of the robotic arm 14 via a rotating telescopic arm 20 (which can rotate horizontally and extend). For example, it can be mounted on the fifth arm from the end of the robotic arm 14. When the robotic arm 14 moves to perform hemostasis by pressing the head 15 and the sealing member 16, the fifth arm from the end does not move, while the last four arms move. The rotating telescopic arm 20 delivers the medical detection device 19 to the detection site, which does not hinder the operation of the medical detection device 19 and can also perform hemostasis on the injured person.

[0026] In this invention, when the support platform 1 is tilted along its width, a protective device is used to limit the movement of the injured person, ensuring that the injured person will not slip off the tilted support platform 1. Specifically, the protective device includes several straps 4 provided on the support platform 1, which are used to secure the injured person.

[0027] More preferably, the protective device also includes a baffle 5 located on the side of the support platform 1. The baffle 5 is located at the lower end when the support platform 1 is tilted. The baffle 5 is an arc-shaped baffle that curves upward and inward from the side of the support platform 1. When the support platform 1 is tilted in the width direction, the baffle 5 supports and protects the lower side of the tilted injured person, preventing the injured person from slipping off the support platform 1 and further improving safety. Moreover, when the support platform 1 is tilted, the outer edge of the baffle 5 will not extend beyond the outer edge of the support platform 1, and will not increase the overall width of the emergency hemostasis robot.

[0028] In another preferred embodiment, the robot body also includes a counterweight device, which improves the stability of the emergency hemostasis robot when the supporting platform 1 is tilted. Specifically, the counterweight device adopts one or a combination of the following structures, preferably both structure one and structure two are provided simultaneously.

[0029] Structure 1: The counterweight device includes a counterweight block 6 located on the higher side of the support platform 1 when it is tilted, or the counterweight block 6 is located on the support leg 2 or the travel wheel 3 on the higher side of the support platform 1 when it is tilted. Preferably, the counterweight block 6 is fixed to the base 10 on the higher side of the support platform 1 where the travel wheel 3 is mounted. By setting the counterweight block 6, when the support platform 1 is tilted, the counterweight block 6 causes the center of gravity of the emergency hemostasis robot to move towards the higher side of the support platform, thereby improving the stability of the emergency hemostasis robot.

[0030] Structure 2: The counterweight device includes a support wheel 7 located inside the traveling wheel 3. The support wheel 7 is a swivel wheel capable of supporting the ground. The support wheel 7 is telescopically connected to the inside of the traveling wheel 3 via a telescopic rod 8. The telescopic rod 8 is a horizontally arranged multi-stage electric push rod, and its housing is fixed to the base 10 on which the traveling wheel 3 is mounted. By setting the support wheel 7, when the supporting platform 1 tilts, the telescopic rod 8 extends, causing the support wheel 7 to move inward toward the traveling wheel 3, thereby enhancing the stability of the emergency hemostasis robot.

[0031] Example 2

[0032] This embodiment provides a control method for the emergency hemostasis robot based on Embodiment 1, such as... Figures 1-4 As shown, the control method includes the following steps:

[0033] S1, the emergency hemostasis robot arrives at the scene.

[0034] S2. The emergency hemostasis robot obtains an image of the bleeding site of the injured person through the medical detection device 19 (ultrasound or X-ray detection) carried by the robotic arm 14, and determines whether it is laceration bleeding or rupture bleeding. If it is laceration bleeding, proceed to step S3; if it is rupture bleeding, proceed to step S4.

[0035] S3. If it is laceration bleeding, determine whether it is arterial bleeding (arterial bleeding will spurt blood in spurts) or venous bleeding (venous bleeding will flow continuously, but will not spurt blood). Apply pressure with the compression head to the proximal end of the corresponding arterial bleeding to stop the bleeding, and apply pressure with the compression head to the distal end of the corresponding venous bleeding to stop the bleeding.

[0036] S4. If it is rupture bleeding, the sealing component is inserted into the wound through the bleeding opening by the continuous robot. The camera at the front end of the continuous robot collects images of the bleeding site, determines the bleeding point, finds the major blood vessel, and inserts the sealing component into the bleeding opening. After the sealing component blocks the bleeding opening, it expands and seals the bleeding major blood vessel.

[0037] The control method for stopping the expansion of the sealing component is as follows:

[0038] S41. The emergency hemostasis robot detects whether the bleeding has stopped using medical testing equipment. If the bleeding stops, the sealing component stops expanding and remains sealed. If the bleeding does not stop, proceed to step S42.

[0039] S42. If the thickness of the blood vessel wall is within the safe threshold, the occlusion device continues to expand and the thickness of the blood vessel wall is continuously monitored. If the thickness of the blood vessel wall reaches the safe threshold, step S43 is executed.

[0040] S43. Use the pressure head in conjunction with pressure to stop bleeding.

[0041] In step S1, the method for the emergency hemostasis robot to reach the treatment site includes: during the movement of the emergency hemostasis robot, the image detection device 11 and the distance detection device 12 at the front end of the robot body rotate and detect to obtain the distance between two parallel vehicles when there is a traffic jam. Based on the distance between the two parallel vehicles, the tilt angle of the bearing platform 1 along the width direction and the spacing of the two side support legs 2 are controlled to reduce the overall width of the emergency hemostasis robot.

[0042] The methods for controlling the tilt angle of the bearing platform 1 and the spacing between the two supporting legs 2 include:

[0043] The distance between two vehicles during a traffic jam is detected by the image detection device 11 and the distance detection device 12. , The distance between the two vehicles at position S for the emergency hemostasis robot is calculated based on the workshop distance. Control the tilt angle of the support platform 1.

[0044] like Then the supporting platform 1 remains horizontal, and the distance between the supporting legs 2 on both sides of the supporting platform 1 remains unchanged. The distance between the support legs 2 on both sides of the platform 1 when it is horizontal. The safe distance coefficient for the emergency hemostasis robot to travel between two vehicles can be set to 1.2 or 1.5 to ensure that the emergency hemostasis robot will not scratch the vehicles.

[0045] like The tilt angle of the bearing platform 1 can be adjusted by extending the two left support legs 2 or shortening the two right support legs 2. for:

[0046]

[0047] At this moment, the support platform 1 tilts, and the spacing between its two supporting legs 2 automatically adjusts to... Simultaneously, by extending the telescopic rod 8, the support wheel 7 on the inner side of the travel wheel 3 on the lower end of the bearing platform 1 is adjusted to move inward, and the distance of inward movement is: Where h is the height of the hinge seat 9 on the lower side of the bearing platform 1 to the ground.

[0048] This invention reduces the width of the entire emergency hemostasis robot by tilting the support platform along its width and narrowing the distance between the two supporting legs. This allows the robot to navigate through gaps between vehicles in traffic jams, achieving the goal of quickly rescuing the injured. Specifically, when... ≥ When the clearance between the two vehicles is large enough, the platform can pass through while remaining level; when ⟨ When the gap between the two vehicles is too small, it indicates that the overall width of the emergency hemostasis robot needs to be reduced so that the robot can move through the gap between the two vehicles in the traffic jam.

[0049] In this invention, the specific method for obtaining the distance between two parallel vehicles during traffic jams is as follows:

[0050] S11. Obtain an image of the front of the emergency hemostasis robot through the image detection device 11;

[0051] S12. Select the images of the two vehicles closest to the emergency hemostasis robot from the images obtained in S11;

[0052] S13. Determine the shortest distance d between the boundaries of two parallel vehicles in the filtered image.

[0053]

[0054] in, The distance between the detection device 12 and the right edge of the vehicle on the left in front. The distance detection device 12 moves forward from the front edge of the vertical support platform 1 until it detects the rotation angle of the right edge of the vehicle on the left in front of it. The distance between the detection device 12 and the left edge of the vehicle on the right in front. The distance detection device 12 moves forward from the front edge of the vertical support platform 1 until it detects the rotation angle of the left edge of the vehicle on the right.

[0055] In step S2 of this invention, the emergency hemostasis robot acquires an image of the bleeding site of the injured person through medical detection equipment, and the method for determining whether it is laceration bleeding or rupture bleeding is as follows:

[0056] Feature extraction is performed on images of bleeding sites in wounded patients acquired by medical testing equipment, specifically through convolution operations and pooling processes:

[0057] ,

[0058] in, The feature map obtained after processing. For deep learning models, For the input image, This means that a convolutional neural network is used to extract features from the input image. The extracted features can include morphological features (such as the shape of the crack, the angle of the break, the size of the blood vessels, etc.).

[0059] For each input image The classifier module of a CNN deep learning model (such as a fully connected layer, typically a Softmax classifier). Output This indicates the score for laceration bleeding and rupture bleeding;

[0060] ,

[0061] in, The score indicates the severity of laceration bleeding. The score indicates the extent of rupture bleeding.

[0062] Using the Softmax function Standardization yields the probabilities of laceration hemorrhage and rupture hemorrhage:

[0063]

[0064]

[0065] in, This indicates the probability (expressed as a percentage) that the image represents a laceration hemorrhage. The probability (expressed as a percentage) that the image belongs to ruptured hemorrhage is represented by e, where e is the natural constant.

[0066] In conclusion and After considering these two probabilities, we determine the final classification result using a predetermined probability threshold. Let's assume the probability threshold is... ,if If the bleeding is severe, it is considered a laceration hemorrhage; otherwise, it is considered a rupture hemorrhage.

[0067] ,in This is the probability threshold.

[0068] In step S4 of the present invention, the method of determining the bleeding point and locating the major bleeding vessel by using an image of the bleeding site, and inserting the sealing member into the bleeding opening is as follows:

[0069] S401, locate all bleeding points;

[0070] Traumatic bleeding sites often appear as abrupt changes in grayscale in medical images. These abrupt changes are identified by analyzing the image's grayscale gradient. Specifically:

[0071]

[0072] in, Image of the bleeding site The grayscale value of the location; The gray-level gradient at that point represents the intensity of the local gray-level change.

[0073] Gray-scale gradient is higher than gradient threshold The points are suspected bleeding points, namely:

[0074] ,

[0075] When the suspected bleeding point is located at ( , ), obtain its activity level, when its activity level is greater than 0, that is:

[0076]

[0077] Furthermore, the gradient change increases in multiple consecutive frames, or the grayscale gradient continuously increases, or remains above the second threshold. ( Greater than If the point is identified as an active bleeding point, then that point is considered an active bleeding point.

[0078] S402, find the point with the largest bleeding volume among all active bleeding points;

[0079] The real-time bleeding rate of each active bleeding point is calculated using local optical flow or volume changes. (t):

[0080]

[0081] in, Let be the surface area of ​​the bleeding region of the i-th active bleeding point; Let be the local blood flow velocity field of the i-th active bleeding point. Specifically, this can be achieved through image registration, calculating the cloud top trajectory of gray-level feature points, calculating the motion velocity, and obtaining the blood flow velocity; n is the surface normal; k is a scaling constant, which depends on sensor calibration. Among these, the parameters... , All measurements are taken using the International System of Units (SI). n is dimensionless, and the dimension of k is determined by the dimensions of other parameters in the formula. The specific value must satisfy the uniformity of the physical properties on both sides of the equation.

[0082] Bleeding volume within a certain period of time :

[0083]

[0084] Choose the point with the greatest bleeding. ; This is the end point of the measurement period;

[0085] S403, Locate the bleeding vessel at this bleeding point;

[0086]

[0087] in, These are weighting coefficients, representing the importance of velocity divergence, volumetric source term, and flow direction, respectively. This refers to the amount of bleeding. For gradient operators; This is a bleeding gradient used to determine the direction of blood flow; d is the velocity divergence, used to determine the presence of bleeding points; d is the unit vector of the flow direction, used to determine the main direction of blood flow. The location of the blood vessel outlet is the point that satisfies the condition of maximum divergence and the starting point of the flow direction; The position of the independent variable is determined so that the function reaches its maximum value later.

[0088] S404, Insert the sealing element into the bleeding opening;

[0089] The target point is the outlet of the blood vessel. The governing equations are as follows:

[0090]

[0091] in, The trajectory of the sealing component at time t; For sealing components The trajectory of movement at any moment For time increments; For step control gain; This is the depth coefficient along the direction of the blood vessel; It is a unit vector in the direction of the blood vessel, used to guide the occlusion device deeper into the blood vessel.

[0092] Final sealing conditions:

[0093]

[0094] in, The final position of the sealing component. For positional error,

[0095] And apply sealing force or expansion:

[0096]

[0097] in, The difference in radius between the sealing component and the blood vessel opening, This is the sealing stiffness coefficient.

[0098] S405, if the bleeding amount is less than the threshold, exit; otherwise, proceed to step S401.

[0099] In this invention, dividing interconnected regions into independent bleeding points requires combining spatial connectivity, feature consistency, and medical / image-based region segmentation rules, based on the "active bleeding point determination." The specific steps are as follows:

[0100] Step 1: First, filter out the set of pixels that meet the activity criteria.

[0101] According to the previous judgment rules, first find all those that satisfy the criteria. And the gradient change meets the requirements (increase / greater than) Candidate pixels, which are the basic units that make up the bleed point.

[0102] Step 2: Divide the region into independent areas based on spatial connectivity

[0103] Bleeding points are spatially continuous regions in an image. Therefore, the candidate pixel set is segmented according to connectivity rules (commonly 4-connectivity or 8-connectivity): 4-connectivity means that a pixel is considered connected only with its four adjacent pixels (top, bottom, left, and right); 8-connectivity means that a pixel is considered connected with its eight adjacent pixels (top, bottom, left, right, and diagonal).

[0104] Connected candidate pixels are grouped into the same region, while disconnected candidate pixel groups are divided into different independent regions. Each independent connected region is a preliminary "bleed point".

[0105] Step 3: Verify consistency by combining bleeding characteristics

[0106] For the initially segmented connected regions, further verify the consistency of the bleed features to ensure that the pixels within the region satisfy:

[0107] 1) Real-time bleeding rate The computational characteristics of (t) are unified (e.g., local blood flow velocity field). (The surface normal n does not have a significant abrupt change within the region).

[0108] 2) The trends in bleeding volume and bleeding gradient remain consistent within the region;

[0109] 3) Surface area of ​​the region It conforms to the medical morphological characteristics of a single bleeding point (without excessive dispersion or splicing).

[0110] If the characteristics of a connected region are consistent, then the region is ultimately determined as an independent active bleed point; if the characteristics are too different, further splitting or merging is required.

[0111] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A first-aid hemostatic robot, characterized by, The robot body capable of walking on the ground, and the mechanical arm with multiple degrees of freedom provided on the robot body, the mechanical arm is provided with a pressing head and a blocking piece, the pressing head is used for pressing the blood vessel of the bleeding port to stop bleeding, the blocking piece is carried by the flexible continuum robot, the front end of the continuum robot is provided with a camera for collecting information of the bleeding site, and the blocking piece can be inserted into the inside of the bleeding port by the continuum robot and expanded after blocking the bleeding port to seal the large blood vessels of the bleeding.

2. The first aid hemostasis robot according to claim 1, characterized in that, The blocking piece is a gas bag or a hemostatic gel that can be inflated after blood absorption, and when the blocking piece is a gas bag, the gas bag is connected with an inflation device through a gas pipe.

3. The emergency hemostatic robot of claim 1, wherein, The robot body includes a carrying platform, four support legs provided at the bottom of the carrying platform to support the carrying platform, travel wheels capable of running on the ground provided at the bottom of each support leg, and a protection device mounted on the carrying platform to limit and protect the wounded; The front end of the carrying platform is provided with an image detection device for detecting obstacles and a distance detection device for detecting the distance between the front end of the carrying platform and the obstacles, and the signal output ends of the image detection device and the distance detection device are connected with a vehicle-mounted controller, which obtains the distance between two vehicles in parallel when the vehicles are blocked; The four support legs are respectively located on the left and right sides of the carrying platform, and the upper ends of the four support legs are rotatably connected with the carrying platform through hinge seats, the support legs on the two sides of the carrying platform can be independently telescopic to adjust the height of the support legs, the variable distance output control end of the vehicle-mounted controller is connected with the telescopic enabling end of the support legs, and by telescoping the support legs, the carrying platform can be tilted along the width direction thereof and the distance between the two side support legs can be reduced, so that the first-aid hemostasis robot can pass through the gap between the two vehicles when the vehicles are blocked.

4. The first aid hemostasis robot according to claim 3, characterized in that, The image detection device is a rotatable image sensor mounted at the front end of the carrying platform, and the distance detection device is a rotatable distance sensor mounted at the front end of the carrying platform, and the image sensor and the distance sensor can be synchronously rotated and oriented in the same direction.

5. The first aid hemostasis robot according to claim 3, characterized in that, The protection device includes a plurality of straps provided on the carrying platform, and further includes a baffle provided on the side of the carrying platform, the baffle is provided at the low end when the carrying platform is tilted, and the baffle is an arc-shaped baffle curved upward and inward from the side of the carrying platform.

6. The first aid hemostasis robot according to claim 1, characterized in that, The robot body further includes a counterweight device, which adopts one of the following structures or a combination of the two structures; Structure one: the counterweight device includes a counterweight block provided on one side of the high end when the carrying platform is tilted, or the counterweight block is provided on the support leg or the travel wheel on one side of the high end when the carrying platform is tilted; Structure two: the counterweight device includes a support wheel provided on the inner side of the travel wheel, the support wheel is a universal wheel capable of supporting on the ground, and the support wheel is telescopically connected with the inner side of the travel wheel through a telescopic rod.

7. A control method of the emergency hemostatic robot according to any one of claims 1 to 6, characterized by, The method comprises the following steps: S1, the first-aid hemostasis robot reaches the rescue site; S2, the first aid hemostasis robot obtains the image of the bleeding part of the wounded through the medical detection device, and judges whether it is laceration bleeding or rupture bleeding. If it is laceration bleeding, step S3 is executed, and if it is rupture bleeding, step S4 is executed. S3, if it is laceration bleeding, it is judged whether it is arterial bleeding or venous bleeding, and the pressing head is pressed at the proximal end of the corresponding arterial bleeding to stop bleeding, and the pressing head is pressed at the distal end of the corresponding venous bleeding to stop bleeding. S4, if it is rupture bleeding, the occlusion member is inserted into the wound part through the continuum robot along the bleeding port, the camera at the front end of the continuum robot collects the image of the bleeding part, judges the bleeding point, finds the large blood vessel of bleeding, and inserts the occlusion member into the inside of the bleeding port, and the occlusion member expands after blocking the bleeding port to close the large blood vessel of bleeding.

8. The control method according to claim 7, characterized by, The control method for stopping the expansion of the occlusion member is: S41, the first aid hemostasis robot detects whether the bleeding stops through the medical detection device. If the bleeding stops, the occlusion member stops expanding and maintains the blocking state. If the bleeding does not stop, step S42 is executed. S42, if the vessel wall thickness is within the safety threshold, the occlusion member continues to expand and continuously detects the vessel wall thickness. If the vessel wall reaches the safety threshold, step S43 is executed. S43, the pressing head is used for joint pressing hemostasis.

9. The control method according to claim 7, characterized by, In step S2, the first aid hemostasis robot obtains the image of the bleeding part of the wounded through the medical detection device, and judges whether it is laceration bleeding or rupture bleeding. The method is: The image of the bleeding part of the wounded obtained by the medical detection device is feature extracted, which is specifically realized by convolution operation and pooling process: ， wherein, is a feature map obtained after processing, is a deep learning model, is an input image, represents feature extraction on the input image by a convolutional neural network; For each input image , the classifier module of the CNN deep learning model outputs a score representing the likelihood of a laceration bleed and a score representing the likelihood of a disruption bleed . ， wherein, a score indicative of laceration bleeding, a score indicative of laceration bleeding; The probabilities of laceration and rupture hemorrhage are obtained by normalizing the Softmax function on ​ wherein, P (rupture | image) denotes the probability that the image belongs to a rupture bleeding, P (rupture | image) denotes the probability that the image belongs to a rupture bleeding, e is the natural constant; wherein is a probability threshold.

10. The control method according to claim 7, characterized by, In step S4, the bleeding point is judged, the large blood vessel of bleeding is found, and the occlusion member is inserted into the inside of the bleeding port through the image of the bleeding part. The method is: S401, find all bleeding points; The mutation point is identified through the gray gradient of the image, specifically: wherein, is the gray value of the bleeding site image at position; is the gray gradient size of the point, indicating the intensity of local gray change; a point having a gray scale gradient higher than a gradient threshold is a suspected bleeding point, i.e. , When the suspected bleeding point is located at ( , ), obtain its activity level, when its activity level is greater than 0: And the gradient change in the continuous multiple frames of images increases or the gray scale gradient continuously increases or is always greater than the second threshold value ( Greater than ), it is determined that the point is an active bleeding point; S402, find the point with the largest bleeding amount in all active bleeding points; calculating a real-time bleeding rate for each active bleeding site (t): wherein, is the surface area of the bleeding region for the i-th active bleeding spot; is the local blood flow velocity field for the i-th active bleeding spot; n is the surface normal; k is a proportionality constant depending on the sensor calibration; Amount of bleeding over a period of time : select the point with the largest amount of bleeding, ; to measure the end of the time period; S403, find the bleeding vessel port of the bleeding point; wherein, are weight coefficients, respectively representing the importance of velocity divergence, volume source term, and flow direction; is the amount of bleeding; is the gradient operator; is the bleeding gradient, used to determine the blood flow direction; is the bleeding velocity divergence, used to determine whether there is a bleeding point; d is the flow direction unit vector, used to determine the main direction of blood flow; is the bleeding vessel orifice position, which is a point that satisfies the maximum divergence and the starting point of the flow direction; is the position argument value when the following function takes the maximum value; S404, insert the occlusion member into the inside of the bleeding port; Target point is the vessel orifice The control equation is as follows: wherein, is the movement trajectory of the occlusion member at time t; is the movement trajectory of the occlusion member at time t+Δt, is the movement trajectory of the occlusion member at time t+Δt, is the time increment; is the step control gain; is the deepening coefficient along the blood vessel direction; is the unit vector of the blood vessel direction, used to guide the occlusion member to deepen along the blood vessel. The final blocking condition is: wherein is the final position of the closure element, is the position error, And the blocking force or expansion amount is applied: wherein, is the difference between the occluding member and the vessel orifice radius, is the occluding stiffness coefficient; S405, if the bleeding amount is less than the threshold, exit, otherwise execute step S401.

Images