Autonomous scanning method and system of breast ultrasonic robot

Abstract

The invention discloses a mammary gland ultrasonic robot autonomous scanning method and system, and the method comprises the steps: placing a probe above a breast, and obtaining the pose, moment and ultrasonic grayscale image of the probe; sequentially executing downward scanning, bottom end rotation, upward scanning and top end rotation scanning from the top end of the breast to be subjected to ultrasonic treatment; in the scanning process, ultrasonic images are analyzed in real time, the contact force between the probe and the skin is detected, it is ensured that the probe continuously makes contact with the skin, and high-quality images are obtained; after a suspicious focus is detected, a related area is subjected to precise scanning, scanning is carried out along the y axis of a probe, an image with the largest cross section area of the focus is selected, the focus position is positioned, information is recorded, the probe is restored to a global scanning interruption position, the image with the largest cross section area of the focus is obtained, and focus information is recorded. According to the invention, artificial intelligence is utilized to identify, detect and position the suspicious lesion in real time, the suspicious lesion is precisely scanned, the most accurate ultrasonic image is selected, and accurate information including the type and position of the lesion is provided for doctors.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot-assisted medical treatment, and more specifically, relates to a breast ultrasound robot autonomous scanning method and system. Background Art

[0002] Breast cancer has a very high incidence and mortality rate, and is one of the most common malignant tumors in women. According to clinical experience, patient survival rate and treatment effect can be improved through effective screening and early detection. With the continuous advancement of science and technology, ultrasound scanning has become an important means of breast cancer examination. However, traditional manual ultrasound scanning relies on the clinical experience of doctors, and its effect is affected by many factors such as the breast status of different patients, the posture during scanning, and the operation of the scanning doctor. Its reliability and accuracy need to be improved.

[0003] In order to solve the above problems, patent document CN118383801A proposes to use a dual-arm robot, based on the breast point cloud model formed by depth camera shooting, with one mechanical arm as the main scanning arm, to obtain the depth image of the chest area, and perform three-dimensional reconstruction of the depth image to form a three-dimensional reconstruction model; generate the scanning trajectory of the ultrasound probe on the breast according to the three-dimensional reconstruction model, and the other arm symmetrically squeezes the breast to reduce the impact of breast deformation on the scanning result. In addition, patent document CN118285841A discloses an adaptive force control method for a breast ultrasound autonomous scanning robot, which scans the breast with an ultrasound probe and calculates the scanning speed according to the scanning trajectory; establishes a mathematical model for deformation estimation, and online identifies and updates the deformation parameters; determines the initial expected force size according to the deformation estimation model; establishes a contact quality evaluation model between the ultrasound probe and the breast, corrects the expected force size, and dynamically controls the contact force between the ultrasound probe and the breast, thereby improving the comfort and accuracy of the scan.

[0004] However, the scanning method of patent document CN118383801A requires taking breast photos in advance to generate a point cloud model, which is difficult for many patients to accept. In addition, a dual-arm robot is required to clamp breast tissue, which makes the patient's comfort difficult to guarantee. Patent document CN118285841A only proposes a force control method, which is lacking in scanning path integrity and targeted treatment after the lesion area is scanned. Summary of the invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a breast ultrasound robot autonomous scanning method and system, which uses artificial intelligence to identify, detect, and locate suspicious lesions in real time, and performs precise scanning of suspicious lesions, selects the most accurate ultrasound images, and provides doctors with accurate information including the type and location of lesions.

[0006] In order to achieve the above object, according to a first aspect of the present invention, a method for autonomous breast ultrasound scanning robot is provided, comprising:

[0007] S100: placing the probe above the breast and obtaining the probe posture, force and torque, and ultrasonic grayscale image;

[0008] S200: starting from the top of the breast to be ultrasounded, sequentially performing down scan, bottom end rotation, up scan and top end rotation scan;

[0009] S300: Analyzes ultrasound images in real time during scanning, detects the contact force between the probe and the skin, ensures that the probe is in continuous contact with the skin, and obtains high-quality images;

[0010] S400: After a suspicious lesion is found, a precise scan is performed on the relevant area, the lesion is scanned along the y-axis of the probe, the image with the largest cross-sectional area of ​​the lesion is selected and the lesion position is located, the information is recorded and the probe is restored to the point where the global scan is interrupted, the image with the largest cross-sectional area of ​​the lesion is obtained, and the lesion information is recorded.

[0011] Furthermore, in step S100, the probe posture p p for:

[0012]

[0013] Among them, [x p ,y p ,z p ] T is the position of the probe in the breast coordinate system {N}; is the Euler angle with the corresponding axis as the rotation axis in this coordinate system, which is used to indicate the direction of the probe.

[0014] Furthermore, in step S100, the force and moment are represented by vector f p It is expressed as:

[0015] f p =[f x ,f y ,f z ,T x ,T y ,T z ] T

[0016] The components of the vector respectively represent the force and torque in each axis direction in the probe coordinate system {P}, which are measured by the force sensor at the end of the robotic arm and represent the torque exerted on the probe when the probe contacts the breast.

[0017] Further, step S200 is controlled by a finite state machine, which is expressed as:

[0018] (S,s0 ,Σ,A,λ,δ)

[0019] Among them, S is a series of set states, including the initial state s 0 , four global scan states (s 1 ,s 2 ,s 3 ,s 4 ), probe adjustment status (s 5 ,s 6 ), accurate scanning status 7 , check the status s 8 and end scans 9 ;Σ=(p p ,f p ,θ c ,Γ) represents the various inputs mentioned above, characterizing the kinematic characteristics of the probe and the characteristics of the ultrasound image; A represents a series of action outputs based on the current state, including the global scanning action (a px ,a la ,a ua ), probe adjustment action (a ia ,a oa ), precise scanning action a fs , check action a ch and end action a e ; λ is the function of the system outputting the corresponding action according to the current state; δ is the function of the system obtaining the system state at the next moment according to the current state and input.

[0020] Furthermore, after the finite state machine outputs a(t), the calculated u is converted into p (t) Control is transmitted to the robot system to change the probe position:

[0021] p p (t+1)=p p (t)+Δt·u p (t)

[0022] Where: p p (t) is the probe posture at time t, Δt is the sampling time, u p (t) = [ N v p (t), N ω p (t)] represents the linear velocity and angular velocity of the probe in the {N} coordinate system.

[0023] Further, in step S200, performing a downward scan starting from the top of the breast to be ultrasounded includes:

[0024] S201: The probe moves along the X-axis of the {P} coordinate system at a speed kx To move, the transformation from coordinate system {P} to coordinate system {N} is expressed as follows:

[0025]

[0026] In the formula, represents the rotation matrix from coordinate system {P} to coordinate system {N}; N v p Represents the linear velocity of the probe in the {N} coordinate system.

[0027] Further, in step S200, the bottom end rotation includes:

[0028] S202: The probe is lifted out of contact with the skin, rotated around the z-axis of the {N} coordinate system by an angle θ, and then contacts the skin again. From the initial and final positions, the movement of the probe can be regarded as a rotation around the z-axis of the {N} coordinate system, so its initial position is To target pose The relationship can be expressed by the following formula:

[0029]

[0030] Where: Represents the rotation matrix for rotating the {N} coordinate system about the Z axis by θ.

[0031] Further, in step S200, the bottom end rotation includes:

[0032] S203: For interpolation on the planned path, assuming that the sampling time of the system control is Δt, k samples are taken from the start of the bottom rotation to a certain moment, then the linear velocity v of the probe at this time is i Can be done by function To calculate, the planning needs to sample K times, and all the v i Put into a set V la middle:

[0033]

[0034] Then the linear velocity of the probe during the movement can be directly obtained from the set V la In the result:

[0035] N v p =v k ,v k ∈V la ,k=0,1,...,K。

[0036] Further, in step S200, scanning from the bottom of the breast upwards includes:

[0037] S204: The probe moves along the X axis of the {P} coordinate system at a speed k x To move, the transformation from coordinate system {P} to coordinate system {N} is expressed as follows:

[0038]

[0039] In the formula, represents the rotation matrix from coordinate system {P} to coordinate system {N}; N v p Represents the linear velocity of the probe in the {N} coordinate system.

[0040] Furthermore, in step S200, the top end rotation includes:

[0041] S205: The probe rotates around the z-axis of the {N} coordinate system by an angle θ, and its angular velocity is:

[0042] N ω P =[0,0,ω z ] T

[0043] Among them, ω z Indicates the angular velocity of the probe rotating around the z-axis of the {N} coordinate system. After the rotation is completed, the system determines whether the global scan is completed.

[0044] Furthermore, in step S300, when the force applied to the probe exceeds a set range, or the inclination angle of the center of gravity of the image exceeds a set range, a corresponding probe posture adjustment is performed to ensure that the probe is in continuous contact with the skin, including:

[0045] If the probe is subjected to a force f in the z-axis direction of the {P} coordinate system z Its design reference value f zd The difference exceeds the threshold t fz Or the image center of gravity tilt angle θ c Greater than the threshold t c , that is, the following Boolean operations hold:

[0046] (||f z -f zd ||>t fz )∧(θ c >t c )=1

[0047] The system controls the probe to move along the Z axis of the probe at k z The probe moves at a speed of ω around the x-axis. x The angular velocity of rotation, the corresponding control quantities are:

[0048]

[0049] Among them, the movement along the Z axis ensures the stability of the contact force, while the rotation around the x axis ensures that the inclination angle of the image center of gravity is minimized, thereby ensuring the image quality.

[0050] Furthermore, in step S300, when the force applied to the probe exceeds a set range, a corresponding probe posture adjustment is performed to eliminate the influence of deformation of the probe when the probe contacts and moves with the breast, including:

[0051] When the probe is in the y-axis vertical plane of the {P} coordinate system, the lateral force f l The reference lateral force f ld The angle error between y Greater than its design reference value t ty hour:

[0052] The probe moves at an angular velocity ω y Rotate around the y-axis of the {P} coordinate system, and the corresponding control amount is:

[0053]

[0054] This reduces the observed lateral force f l The designed lateral force f ld The angle error between .

[0055] Further, in step S400, obtaining an image with the largest cross-sectional area of ​​the lesion includes:

[0056]

[0057] Where: i is the number of the potential lesion found, m i and n i Represent the pixel height and width of the lesion, and They represent functions that take two values ​​for Ω.

[0058] Furthermore, in step S400, locating the lesion position includes:

[0059]

[0060] Where: Output the image The corresponding probe pose, The horizontal and vertical coordinates of the pixels of the lesion in the image were obtained respectively, w u The scaling factor between the pixel length in the image and the real-world length.

[0061] According to a second aspect of the present invention, a robot autonomous breast ultrasound scanning system is provided, which is used to implement the robot autonomous breast ultrasound scanning method, comprising:

[0062] An initialization module is used to place the probe above the breast and obtain the probe posture, torque and ultrasonic grayscale image;

[0063] An ultrasound scanning module, used to sequentially perform downward scanning, bottom rotation scanning, upward scanning and top rotation scanning starting from the top of the breast to be ultrasounded;

[0064] Image analysis module, used to analyze ultrasound images in real time during scanning, detect the contact force between the probe and the skin, ensure that the probe is in continuous contact with the skin, and obtain high-quality images;

[0065] The precise scanning module is used to perform precise scanning of the relevant area after a suspicious lesion is discovered. It scans along the y-axis of the probe, selects the image with the largest cross-sectional area of ​​the lesion and locates the lesion, records the information and restores the probe to the point where the global scan was interrupted, obtains the image with the largest cross-sectional area of ​​the lesion, and records the lesion information.

[0066] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0067] 1. The method of the present invention utilizes artificial intelligence to identify, detect, and locate suspicious lesions in real time. After a suspicious lesion is found, a precise scan is performed on the relevant area, a scan is performed along the y-axis of the probe, the image with the largest cross-sectional area of ​​the lesion is selected and the lesion position is located, the information is recorded, and the probe is restored to the point where the global scan is interrupted, the image with the largest cross-sectional area of ​​the lesion is obtained, and the lesion information is recorded.

[0068] 2. The method of the present invention uses a single robotic arm to perform a complete ultrasonic scan through the position of the ultrasonic probe, the contact force with the human body and the ultrasonic image, without the need for prior 3D modeling, thereby reducing the amount of calculation and improving the accuracy of the ultrasonic scan.

[0069] 3. The method of the present invention uses a finite state machine to control the scanning system, and responds in real time to situations such as poor probe posture and the appearance of suspicious lesions during scanning, thereby improving the quality of the scanning.

[0070] 4. The method of the present invention comprehensively considers the contact force and the inclination angle of the center of gravity of the breast image to adjust the probe posture in real time to reduce the influence of tissue deformation on the scan, thereby ensuring the integrity of the scan and the high quality of the ultrasound image. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a working diagram of the autonomous breast ultrasound scanning robot in an embodiment of the present invention;

[0072] Figure 2 Schematic diagram of a breast coordinate system {N} and a probe coordinate system {P} of a system in an embodiment of the present invention;

[0073] Figure 3 Schematic diagram of the inclination angle of the center of gravity of an image in an embodiment of the present invention;

[0074] Figure 4 4 is a flow chart of the global scanning process in an embodiment of the present invention. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0076] like Figure 1 In an embodiment of the present invention, a method for autonomous breast ultrasound robot scanning is provided, comprising:

[0077] S100: placing the probe above the breast and obtaining the probe posture, torque and ultrasonic grayscale image;

[0078] S200: starting from the top of the breast to be ultrasounded, sequentially performing down scan, bottom end rotation, up scan and top end rotation scan;

[0079] S300: Analyzes ultrasound images in real time during scanning, detects the contact force between the probe and the skin, ensures that the probe is in continuous contact with the skin, and obtains high-quality images;

[0080] S400: After the suspicious lesion is found, the relevant area is scanned precisely, the y-axis of the probe is scanned, the image with the largest cross-sectional area of ​​the lesion is selected and the lesion position is located, the information is recorded and the probe is restored to the point where the global scan is interrupted, the image with the largest cross-sectional area of ​​the lesion is obtained, and the lesion information is recorded. The method of the present invention uses artificial intelligence to identify, detect, and locate suspicious lesions in real time, and performs precise scanning on suspicious lesions, selects the most accurate ultrasound image, and provides doctors with accurate information including the type and location of the lesion.

[0081] Embodiment 1:

[0082] (1) When controlling, the system senses the following information: probe position p p , torque f p , ultrasonic grayscale image. Among them,

[0083] The probe position p p for:

[0084]

[0085] Among them, [x p ,y p,z p ] T is the position of the probe in the breast coordinate system {N}; is the Euler angle with the corresponding axis as the rotation axis in this coordinate system, which is used to indicate the direction of the probe.

[0086] The forces and moments are represented by vectors f p It is expressed as:

[0087] f p =[f x ,f y ,f z ,T x ,T y ,T z ] T

[0088] The components of this vector represent the forces and moments in the directions of the axes in the probe coordinate system {P}, which are measured by the force sensor at the end of the robotic arm and represent the moments exerted on the probe when it contacts the breast. Figure 2 The method of the present invention comprehensively considers the contact force and the inclination angle of the center of gravity of the breast image to adjust the probe posture in real time to reduce the influence of tissue deformation on the scan and ensure the integrity of the scan and the high quality of the ultrasound image.

[0089] The system obtains the image's centroid tilt angle θ through grayscale ultrasound images c For uniform, clear, and complete imaging, the image center of gravity should be close to the perpendicular midline of the image, corresponding to θ c Should be close to 0 degrees. Figure 3 At the same time, the potential lesions are identified, segmented and tracked based on artificial intelligence methods, and the results are represented by a set of arrays Γ. Γ={(i,l i ,[g i ,h i ] T ,m i ,n i )|i∈{1,2,...,M}}, where i represents the number of the lesion found, and a total of M lesions were found. i Indicates the type of lesion, which may be a tumor or a nodule. i ,h i ] T Indicates the pixel position of the lesion in the image. m i and n i Represent the pixel height and width of the lesion, respectively.

[0090] (2) The control of the ultrasound probe is completed by a finite state machine (FSM), which can be represented by a tuple (S, s 0,Σ,A,λ,δ). Among them, S is a series of set states, including the initial state s 0 , four global scan states (s 1 ,s 2 ,s 3 ,s 4 ), probe adjustment status (s 5 ,s 6 ), accurate scanning status 7 , check the status s 8 and end scans 9 Σ=(p p ,f p ,θ c ,Γ) represents the various inputs mentioned above, characterizing the kinematic characteristics of the probe and the characteristics of the ultrasound image. A represents a series of action outputs based on the current state, including the global scanning action (a px ,a la ,a ua ), probe adjustment action (a ia ,a oa ), precise scanning action a fs , check action a ch and end action a e . λ is the function of the system outputting the corresponding action according to the current state; δ is the function of the system obtaining the system state at the next moment according to the current state and input. The equation a(t)=λ(s(t)) represents that the system determines the current action according to the current state. The equation s(t+1)=δ(s(t),Σ(t)) represents that the system updates the state at the next moment according to the current state and input. The system sets a series of events E, as shown in the following table:

[0091] Table 1 List of events in the finite state machine

[0092]

[0093]

[0094] In the system 1 ,E 2 ,...,E 10 is a series of Boolean values. When the event in the corresponding description is true, the value is 1, otherwise it is 0. When the system is in a certain state s, the Boolean operation result of the related event E is determined to determine the next state. After the FSM obtains the output a(t), it converts the calculated u into p (t) Control is transmitted to the robot system. The system changes the probe position according to the following formula:

[0095] p p (t+1)=p p(t)+Δt·u p (t) (1)

[0096] Where p p (t) is the probe posture at time t, Δt is the sampling time, u p (t) = [ N v p (t), N ω p (t)] represents the linear velocity and angular velocity of the probe in the {N} coordinate system. The method of the present invention uses a finite state machine to control the scanning system, and responds in real time to situations such as poor probe posture and the appearance of suspicious lesions during scanning, thereby improving the quality of scanning.

[0097] like Figure 4 As shown, the state transfer judgment and probe control methods when the system is in global scanning, fine scanning, and posture adjustment are introduced below.

[0098] (3) At the initial moment, the probe is placed above the nipple and the global scan begins. Figure 4 As shown. The system enters 1 State, that is, scanning from the nipple down to the edge of the breast. At this time, perform a px operation, even if the probe moves along the X-axis of the {P} coordinate system at a speed k x To move, the transformation from coordinate system {P} to coordinate system {N} is expressed as follows:

[0099]

[0100] Among them, represents the rotation matrix from coordinate system {P} to coordinate system {N}, N v p Represents the linear velocity of the probe in the {N} coordinate system.

[0101] At this time, if the Boolean operation E 1 ∧! E 2 ∧(!E 5 ∧! E 6 ! E 7 ) is true, indicating that the downward scan has not been completed, the probe posture does not need to be adjusted, and no new potential lesions have been found, so a is always executed. px Operation. When Boolean operation! E 1 ∨E 2 When it is true, the downward scan ends and the system enters s 2 State, that is, low position adjustment state: the probe is lifted out of contact with the skin, rotated an angle θ around the z-axis of the {N} coordinate system, and then contacts the skin again. At this time, the system executes a la Operation, that is, interpolating points along the planned path from the current pose Move to the target pose

[0102]

[0103] Where: Represents the rotation matrix for rotating the {N} coordinate system about the Z axis by θ.

[0104] For interpolation on the planned path, assuming that the sampling time of the system control is Δt, k samples are taken from the start of the bottom rotation to a certain moment, then the linear velocity v of the probe at this time is i Can be done by function To calculate, the planning needs to sample K times, and all the v i Put into a set V la middle:

[0105]

[0106] Then the linear velocity of the probe during the movement can be directly obtained from the set V la In the result:

[0107] N v p =v k ,v k ∈V la ,k=0,1,...,K (5)

[0108] When E 3 =1, it means the probe is in contact with the skin again, and the system enters s 3 state, in which the probe performs a px Move the probe from the edge of the breast back to the vicinity of the nipple. 4 = 1, the movement is completed and the system enters s 4 State, that is, high-position adjustment state: the probe rotates around the z-axis of the {N} coordinate system by an angle θ, and its angular velocity is expressed by the following formula:

[0109] N ω P =[0,0,ω z ] T (6)

[0110] Among them, ω z Indicates the angular velocity of the probe rotating around the z-axis of the {N} coordinate system. After the rotation is completed, the system determines whether the global scan is completed. If not, re-enter s 1 status and start the next scanning cycle.

[0111] (4) If in s 2 Status or 3The state detects a potential lesion, or the probe position is unreasonable, that is, the Boolean operation E 5 ∨E 6 ∨E 7 If the result is true, the system will enter s 8 Check the status. At this time, if E 5 =1, the plane attitude calibration state will start s 5 The purpose is to keep the probe in contact with the skin continuously, obtain high-quality images, and thus accurately scan suspicious lesions. If the probe is subjected to a force f in the z-axis direction of the {P} coordinate system z Its design reference value f zd The difference exceeds the threshold t fz Or the image center of gravity tilt angle θ c Greater than the threshold t c , that is, the following Boolean operations hold:

[0112] (||f z -f zd ||>t fz )∧(θ c >t c )=1

[0113] In this state, the output action a ia , that is, along the Z axis of the probe with k z The probe moves at a speed of ω around the x-axis. x The angular velocity of the rotation, the corresponding control amount is calculated by the following formula:

[0114]

[0115] Movement along the Z axis ensures the stability of the contact force, while rotation around the x axis ensures the minimum inclination of the image center of gravity, thus ensuring image quality. 5 ∧! E 7 ∧E 6 =1, the system will enter the Y-axis vertical plane attitude adjustment state s 6 This adjustment is to eliminate the influence of deformation caused by the probe contacting and moving with the breast. The action output of this state is a oa The angular velocity ω y Rotate around the probe y-axis, and the corresponding control quantity is shown as follows:

[0116]

[0117] This reduces the observed lateral force f l The designed lateral force f ld The angle error between .

[0118] (5) When the system detects a potential lesion and does not require plane posture adjustment, that is!5 ∧E 7 =1, the system enters state s 7 , perform a precise scan of potential lesions. This state includes three sub-states: scanning along the y-axis of the probe, selecting the image with the largest lesion cross-sectional area and locating the lesion, recording information and restoring the probe to the point where the global scan was interrupted.

[0119] The probe first scans the suspected lesion until the YOLOv8 algorithm cannot detect the potential lesion or the scanning angle is greater than the threshold. When the scan is completed, the image with the largest lesion area is selected, as shown in the following formula:

[0120]

[0121] Where: i is the number of the potential lesion found, m i and n i Represent the pixel height and width of the lesion, respectively. and Respectively represent the function of finding two values ​​for Ω. Then the lesion position is calculated according to the following formula:

[0122]

[0123] Where: Output the image The corresponding probe pose, The horizontal and vertical coordinates of the pixels of the lesion in the image were obtained respectively, w u Indicates the ratio factor between the pixel length in the image and the real world length. The lesion information after fine scanning is recorded as an array L i This information represents the i-th element in the total lesion detection set Γ. The results of the fine scan are summarized into a set Υ = {(L i ,p Li ,Ω i * )|i∈(1,2,...,M)}, where M represents the number of lesions detected.

[0124] Embodiment 2:

[0125] In another embodiment of the present invention, a robot autonomous breast ultrasound scanning system is provided, which is used to implement the robot autonomous breast ultrasound scanning method, and includes:

[0126] An initialization module is used to place the probe above the breast and obtain the probe posture, torque and ultrasonic grayscale image;

[0127] An ultrasound scanning module, used to sequentially perform downward scanning, bottom rotation scanning, upward scanning and top rotation scanning starting from the top of the breast to be ultrasounded;

[0128] Image analysis module, used to analyze ultrasound images in real time during scanning, detect the contact force between the probe and the skin, ensure that the probe is in continuous contact with the skin, and obtain high-quality images;

[0129] The precise scanning module is used to perform precise scanning of the relevant area after a suspicious lesion is discovered. It scans along the y-axis of the probe, selects the image with the largest cross-sectional area of ​​the lesion and locates the lesion, records the information and restores the probe to the point where the global scan was interrupted, obtains the image with the largest cross-sectional area of ​​the lesion, and records the lesion information.

[0130] The system of the present invention uses a single robotic arm to perform a complete ultrasonic scan through the position of the ultrasonic probe, the contact force with the human body and the ultrasonic image, without the need for prior 3D modeling, thereby reducing the amount of calculation and improving the accuracy of the ultrasonic scan.

[0131] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A breast ultrasound robot autonomous scanning method, characterized in that: include: S100: placing the probe above the breast and obtaining the probe posture, force and torque, and ultrasonic grayscale image; S200: starting from the top of the breast to be ultrasounded, sequentially performing down scan, bottom end rotation, up scan and top end rotation scan; S300: Analyzes ultrasound images in real time during scanning, detects the contact force between the probe and the skin, ensures that the probe is in continuous contact with the skin, and obtains high-quality images; S400: After a suspicious lesion is found, a precise scan is performed on the relevant area, the lesion is scanned along the y-axis of the probe, the image with the largest cross-sectional area of ​​the lesion is selected and the lesion position is located, the information is recorded and the probe is restored to the point where the global scan is interrupted, the image with the largest cross-sectional area of ​​the lesion is obtained, and the lesion information is recorded.

2. The method for autonomous breast ultrasound scanning by a robot according to claim 1, characterized in that: In step S100, the probe position p p for: Among them, [x p ,y p ,z p ] T is the position of the probe in the breast coordinate system {N}; is the Euler angle with the corresponding axis as the rotation axis in this coordinate system, which is used to indicate the direction of the probe.

3. The method for autonomous breast ultrasound scanning by a robot according to claim 2, characterized in that: In step S100, the force and moment are represented by vector f p It is expressed as: f p =[f x ,f y ,f z ,T x ,T y ,T z ] T The components of the vector respectively represent the force and torque in each axis direction in the probe coordinate system {P}, which are measured by the force sensor at the end of the robotic arm and represent the torque exerted on the probe when the probe contacts the breast.

4. A breast ultrasound robot autonomous scanning method according to any one of claims 1-3, characterized in that: Step S200 is controlled by a finite state machine and is expressed as: (S,s0,Σ,A,λ,δ) Where S is a set series of states, including the initial state s0, four global scanning states (s1, s2, s3, s4), the probe adjustment state (s5, s6), the precise scanning state s7, the inspection state s8 and the end of the scan s9; Σ=(p p ,f p ,θ c ,Γ) represents the various inputs mentioned above, characterizing the kinematic characteristics of the probe and the characteristics of the ultrasound image; A represents a series of action outputs based on the current state, including the global scanning action (a px ,a la ,a ua ), probe adjustment action (a ia ,a oa ), precise scanning action a fs , check action a ch and end action a e ; λ is the function of the system outputting the corresponding action according to the current state; δ is the function of the system obtaining the system state at the next moment according to the current state and input.

5. The method for autonomous breast ultrasound scanning by a robot according to claim 4, characterized in that: After the finite state machine outputs a(t), the calculated u is converted into p (t) Control is transmitted to the robot system to change the probe position: p p (t+1)=p p (t)+Δt·u p (t) Where: p p (t) is the probe posture at time t, Δt is the sampling time, u p (t) = [ N v p (t), N ω p (t)] represents the linear velocity and angular velocity of the probe in the {N} coordinate system.

6. The method for autonomous breast ultrasound robot scanning according to claim 5, characterized in that: In step S200, performing a downward scan starting from the top of the breast to be ultrasounded includes: S201: The probe moves along the X-axis of the {P} coordinate system at a speed k x To move, the transformation from coordinate system {P} to coordinate system {N} is expressed as follows: In the formula, represents the rotation matrix from coordinate system {P} to coordinate system {N}; N v p Represents the linear velocity of the probe in the {N} coordinate system.

7. The method for autonomous breast ultrasound scanning by a robot according to claim 6, characterized in that: In step S200, the bottom end rotation includes: S202: The probe is lifted out of contact with the skin, rotated around the z-axis of the {N} coordinate system by an angle θ, and then contacts the skin again. From the initial and final positions, the movement of the probe can be regarded as a rotation around the z-axis of the {N} coordinate system, so its initial position is To target pose The relationship can be expressed by the following formula: Where: Represents the rotation matrix for rotating the {N} coordinate system about the Z axis by θ.

8. The method for autonomous breast ultrasound robot scanning according to claim 7, characterized in that: In step S200, the bottom end rotation includes: S203: For interpolation on the planned path, assuming that the sampling time of the system control is Δt, k samples are taken from the start of the bottom rotation to a certain moment, then the linear velocity v of the probe at this time is i By function To calculate, the planning needs to sample K times, and all the v i Put into a set V la middle: Then the linear velocity of the probe during the movement can be directly obtained from the set V la In the result: N v p =v k ,v k ∈V la ,k=0,1,...,K。 9. The method for autonomous breast ultrasound robot scanning according to claim 8, characterized in that: In step S200, scanning from the bottom of the breast upwards includes: S204: The probe moves along the X axis of the {P} coordinate system at a speed k x To move, the transformation from coordinate system {P} to coordinate system {N} is expressed as follows: In the formula, represents the rotation matrix from coordinate system {P} to coordinate system {N}; N v p Represents the linear velocity of the probe in the {N} coordinate system.

10. The method for autonomous breast ultrasound scanning by a robot according to claim 9, characterized in that: In step S200, the top rotation includes: S205: The probe rotates around the z-axis of the {N} coordinate system by an angle θ, and its angular velocity is: N oh P =[0,0,ω z ] T Among them, ω z Indicates the angular velocity of the probe rotating around the z-axis of the {N} coordinate system. After the rotation is completed, the system determines whether the global scan is completed.

11. The method for autonomous breast ultrasound robot scanning according to claim 10, characterized in that: In step S300, when the force applied to the probe exceeds a set range, or the inclination angle of the center of gravity of the image exceeds a set range, a corresponding probe posture adjustment is performed to ensure that the probe is in continuous contact with the skin, including: If the probe is subjected to a force f in the z-axis direction of the {P} coordinate system z Its design reference value f zd The difference exceeds the threshold t fz Or the image center of gravity tilt angle θ c Greater than the threshold t c , that is, the following Boolean operations hold: (||f z -f zd ||>t fz )∧(θ c >t c )=1 The system controls the probe to move along the Z axis of the probe at k z The probe moves at a speed of ω around the x-axis. x The angular velocity of rotation, the corresponding control quantities are: Among them, the movement along the Z axis ensures the stability of the contact force, while the rotation around the x axis ensures that the inclination angle of the image center of gravity is minimized, thereby ensuring the image quality.

12. The method for autonomous breast ultrasound robot scanning according to claim 11, characterized in that: In step S300, when the force applied to the probe exceeds a set range, a corresponding probe posture adjustment is performed to eliminate the influence of deformation caused by the probe contacting and moving with the breast, including: When the probe is in the y-axis vertical plane of the {P} coordinate system, the lateral force f l The reference lateral force f ld The angle error between y Greater than its design reference value t ty hour: The probe moves at an angular velocity ω y Rotate around the y-axis of the {P} coordinate system, and the corresponding control amount is: This reduces the observed lateral force f l The designed lateral force f ld The angle error between .

13. A breast ultrasound robot autonomous scanning method according to any one of claims 1-3, characterized in that: In step S400, obtaining an image with the largest cross-sectional area of ​​the lesion includes: Where: i is the number of the potential lesion found, m i and n i Represent the pixel height and width of the lesion, and They represent functions that take two values ​​for Ω.

14. The method for autonomous breast ultrasound robot scanning according to claim 13, characterized in that: In step S400, locating the lesion position includes: Where: Output the image The corresponding probe pose, The horizontal and vertical coordinates of the pixels of the lesion in the image were obtained respectively, w u The scaling factor between the pixel length in the image and the real-world length.

15. A robotic autonomous breast ultrasound scanning system, characterized in that: The method for implementing the robot autonomous breast ultrasound scanning method according to any one of claims 1 to 14 comprises: An initialization module is used to place the probe above the breast and obtain the probe posture, torque and ultrasonic grayscale image; An ultrasound scanning module, used to sequentially perform downward scanning, bottom rotation scanning, upward scanning and top rotation scanning starting from the top of the breast to be ultrasounded; Image analysis module, used to analyze ultrasound images in real time during scanning, detect the contact force between the probe and the skin, ensure that the probe is in continuous contact with the skin, and obtain high-quality images; The precise scanning module is used to perform precise scanning of the relevant area after a suspicious lesion is discovered. It scans along the y-axis of the probe, selects the image with the largest cross-sectional area of ​​the lesion and locates the lesion, records the information and restores the probe to the point where the global scan was interrupted, obtains the image with the largest cross-sectional area of ​​the lesion, and records the lesion information.

Images