Region detection method, device and equipment based on robot control and medium

Abstract

The invention relates to the technical field of robot control, and discloses an area detection method, device, equipment and medium based on robot control, and the method comprises the steps: controlling a robot arm to drive a probe to move in a three-dimensional space, monitoring a contact force through a force sensor and dynamically adjusting the contact force, and tracking and detecting a target position by adopting a visual servo technology. And generating translation and rotation control instructions, and driving the probe to carry out attitude adjustment. And collecting detection signal data in pre-compression and post-compression states, calculating to generate a three-dimensional strain distribution diagram, performing threshold segmentation to extract a rigid region, and outputting spatial position information. The robot arm is controlled to drive the probe in the three-dimensional space, and the real-time force feedback of the sensor, the target position tracking of the visual servo technology and the rotation control of the teleoperation equipment are combined, so that the high-precision dynamic control of the probe in the three-dimensional space is realized, the imaging precision and the detection accuracy of a tissue rigid region are improved, and the detection accuracy of the tissue rigid region is improved. And the operation error is obviously reduced.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a method, apparatus, device and storage medium for area detection based on robot control. Background Technology

[0002] Ultrasound systems are non-invasive imaging tools widely used in the medical field. They offer advantages such as real-time imaging, ease of operation, and no radiation, making them indispensable imaging equipment in clinical diagnosis. Ultrasound imaging technology provides information on tissue structure and dynamics, and has significant clinical application value, particularly in early disease diagnosis, tumor detection, and tissue elasticity assessment. Ultrasound systems generate images by emitting ultrasound signals and receiving their echoes, allowing doctors to determine the patient's tissue structure and lesion areas. However, traditional ultrasound imaging systems rely on manual operation; probe placement and angle adjustment depend primarily on manual control by the doctor. This manual operation method has significant shortcomings in practical applications.

[0003] First, manually operated ultrasound probes are easily affected by the operator's experience and skill level. Even experienced physicians cannot guarantee that the probe's angle and position will be completely consistent every time, especially when dealing with tissues with large surface areas or complex anatomical structures, where probe stability and accuracy are even more difficult to control. This human factor can lead to fluctuations in image quality and even affect the accuracy of diagnosis. Furthermore, ultrasound imaging has strict requirements on the probe's contact force; excessive or insufficient pressure applied to the tissue can affect image clarity and tissue deformation characteristics, and manual operation makes precise control of this contact force difficult.

[0004] Secondly, the natural movement of tissues during ultrasound imaging is also a significant factor affecting image quality. Physiological movements such as breathing and heartbeat can cause changes in tissue position, thus affecting the probe's continuous alignment with the target area. In traditional manual ultrasound procedures, doctors need to adjust the probe position in real time to track the target tissue. This not only increases the difficulty of the operation but can also lead to frequent probe movements, further affecting imaging stability. Furthermore, when the target tissue is located deep within the body or overlaps with other tissues, the doctor's manual control of the probe's alignment accuracy is limited, making it difficult to maintain continuous and stable imaging results.

[0005] To improve the stability and accuracy of ultrasound imaging, robot-assisted ultrasound imaging systems are increasingly being introduced into the medical field. Robotics technology enables ultrasound imaging to be completed within a relatively small error range by precisely controlling the spatial position and orientation of the probe. However, most existing robot-assisted ultrasound systems only control the probe's movement in a two-dimensional plane, or only achieve partial control of the probe's degrees of freedom, failing to fully cover the complex motion patterns of the probe in three-dimensional space. This limited control mode makes it difficult for the probe to maintain continuous tracking of the target tissue when facing complex three-dimensional anatomical structures or dynamic tissue movements.

[0006] Furthermore, existing robot-assisted ultrasound imaging systems have shortcomings in achieving probe posture control (such as rotation and tilt). Especially when monitoring real-time tissue movement within the patient (such as respiration and heartbeat), changes in probe posture significantly affect image quality. Traditional robotic systems often lack effective posture control strategies, making it difficult to ensure the probe remains aligned with the target tissue in dynamic scenarios. While some existing technologies incorporate visual servo control for automatic target tissue tracking, this type of visual servo control is largely limited to planar features of the image, making stable tracking in three-dimensional space difficult. Moreover, tracking accuracy is often compromised when faced with noise, tissue occlusion, or rapidly moving targets.

[0007] Finally, while some existing technologies attempt to introduce force control mechanisms in robot-assisted ultrasound imaging to ensure stable pressure applied by the probe to the tissue, these force control techniques are mostly simple constant force or constant pressure control, which is difficult to adapt to the elastic properties and dynamic changes of different tissues. Especially in complex elastography scenarios, dynamic adjustment of contact force and multi-mode force control are difficult to achieve. This lack of control leads to significant errors in tissue deformation calculations during elastography, affecting the accuracy and precision of the imaging process. Summary of the Invention

[0008] The main objective of this invention is to provide a robot-controlled area detection method, device, equipment, and storage medium, aiming to solve the technical problems of existing robot-assisted ultrasound systems being unable to achieve coordinated high-precision control of probe position, orientation, and contact force in three-dimensional space, resulting in inaccurate target tracking under dynamic physiological motion, unstable palpation pressure, and distorted elastic imaging.

[0009] To achieve the above objectives, the present invention provides a region detection method based on robot control, comprising:

[0010] Control the robot arm to drive the probe's three-dimensional spatial movement;

[0011] Based on the control of the three-dimensional spatial motion, the contact force applied by the probe to the elastic medium is monitored by a force sensor, and the contact force is adjusted according to a preset force control algorithm;

[0012] During the adjustment of the contact force, the real-time position of the target is tracked by visual servo control technology, and translation control commands are generated based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0013] During the adjustment of translational motion, rotation commands are received through remote control devices, and rotation control commands are generated to adjust the rotation angle of the probe.

[0014] Based on the adjusted contact force, translational motion, and rotation angle, detection signal data in the pre-compression and post-compression states are collected, and a three-dimensional strain distribution map is generated.

[0015] The three-dimensional strain distribution map is subjected to threshold segmentation to extract rigid regions and output the spatial location information of the rigid regions.

[0016] Furthermore, to achieve the above objectives, the present invention provides a robot-controlled area detection device, comprising:

[0017] The robot motion control module is used to control the three-dimensional spatial motion of the robot arm driving the probe.

[0018] A force sensor control module is used for controlling the motion in the three-dimensional space. It monitors the contact force applied by the probe to the elastic medium through a force sensor and adjusts the contact force according to a preset force control algorithm.

[0019] The visual servo control module is used to track the real-time position of the target during the adjustment of the contact force using visual servo control technology, and generate translation control commands based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0020] The rotation control module is used to receive rotation commands through a remote control device and generate rotation control commands to adjust the rotation angle of the probe during the adjustment of translational motion.

[0021] The signal acquisition and strain calculation module is used to acquire detection signal data under pre-compression and post-compression states based on the adjusted contact force, translational motion and rotation angle, and calculate and generate a three-dimensional strain distribution map.

[0022] The threshold segmentation and rigid region extraction module is used to perform threshold segmentation processing on the three-dimensional strain distribution map, extract the rigid region, and output the spatial location information of the rigid region.

[0023] Furthermore, to achieve the above objectives, the present invention also provides a computer device, the computer device including a memory, a processor, and a robot-controlled region detection program stored in the memory and executable on the processor, wherein the robot-controlled region detection program, when executed by the processor, implements the steps of the robot-controlled region detection method as described above.

[0024] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing a robot-controlled region detection program, which, when executed by a processor, implements the steps of the robot-controlled region detection method as described above.

[0025] Beneficial Effects: This invention relates to the field of robot control technology and discloses a method, apparatus, device, and medium for region detection based on robot control. The method includes: controlling a robot arm to drive the three-dimensional spatial motion of a probe; based on the control of the three-dimensional spatial motion, monitoring the contact force applied by the probe to an elastic medium using a force sensor, and adjusting the contact force according to a preset force control algorithm; during the adjustment of the contact force, using visual servo control technology to track the real-time position of the detection target, and generating translation control commands based on the deviation between the real-time position and a preset field of view center to drive the robot arm to adjust the translational motion of the probe; during the adjustment of the translational motion, receiving rotation commands through a teleoperation device, generating rotation control commands to adjust the rotation angle of the probe; based on the adjusted contact force, translational motion, and rotation angle, collecting detection signal data under pre-compression and post-compression states, calculating and generating a three-dimensional strain distribution map, performing threshold segmentation processing on the three-dimensional strain distribution map, extracting rigid regions, and outputting the spatial position information of the rigid regions. This invention achieves high-precision dynamic control of the probe in three-dimensional space by controlling a robotic arm to drive the probe, combined with real-time force feedback from a force sensor, target position tracking using visual servo technology, and rotation control from a teleoperated device. This ensures stable contact between the probe and the target under preset contact force, translation, and rotation conditions. Through the acquisition of detection signals and the calculation of three-dimensional strain distribution maps under pre-compression and post-compression states, the elastic properties of the target tissue can be accurately reflected, improving imaging accuracy and the detection accuracy of rigid tissue regions, and significantly reducing operational errors. Attached Figure Description

[0026] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0027] Figure 1 This is a schematic diagram of an application environment for a robot-controlled region detection method according to an embodiment of the present invention;

[0028] Figure 2This is a flowchart illustrating an embodiment of the robot-controlled region detection method of the present invention;

[0029] Figure 3 This is a schematic diagram of the functional modules of a preferred embodiment of the robot-controlled area detection device of the present invention;

[0030] Figure 4 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention;

[0031] Figure 5 This is another structural schematic diagram of a computer device according to one embodiment of the present invention. Detailed Implementation

[0032] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0033] The robot-controlled region detection method provided in this invention can be applied to, for example... Figure 1 In this application environment, the user terminal communicates with the server via a network. The server can control the robot arm to drive the probe's three-dimensional spatial movement through the user terminal. Based on the control of the three-dimensional spatial movement, the contact force applied by the probe to the elastic medium is monitored by a force sensor, and the contact force is adjusted according to a preset force control algorithm. During the adjustment of the contact force, visual servo control technology is used to track the real-time position of the detection target, and translation control commands are generated based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe. During the adjustment of the translational movement, rotation commands are received through a teleoperation device, and rotation control commands are generated to adjust the rotation angle of the probe. Based on the adjusted contact force, translational movement, and rotation angle, detection signal data in the pre-compression and post-compression states are collected, a three-dimensional strain distribution map is calculated and generated, the three-dimensional strain distribution map is subjected to threshold segmentation processing, rigid regions are extracted, and the spatial position information of the rigid regions is output. This invention achieves high-precision dynamic control of the probe in three-dimensional space by controlling a robotic arm to drive the probe, combined with real-time force feedback from a force sensor, target position tracking using visual servo technology, and rotation control from a teleoperated device. This ensures stable contact between the probe and the target under preset contact force, translation, and rotation conditions. Through the acquisition of detection signals and the calculation of three-dimensional strain distribution maps under pre-compression and post-compression states, the elastic characteristics of the target tissue can be accurately reflected, improving imaging accuracy and the detection accuracy of rigid tissue regions, and significantly reducing operational errors. The user end can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server end can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will be described in detail below through specific embodiments.

[0034] Please see Figure 2 , Figure 2 This is a flowchart illustrating an embodiment of the robot-controlled region detection method provided by the present invention. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0035] like Figure 2 As shown, the region detection method based on robot control proposed in this invention includes the following steps:

[0036] S10 controls the three-dimensional spatial movement of the probe driven by the robot arm;

[0037] In this embodiment, controlling the movement of the probe driven by the robotic arm in three-dimensional space is crucial for precisely achieving stable movement of the probe within the three-dimensional coordinate system and ensuring that the probe can adapt to different spatial positions and postures during operation. Achieving this goal first requires defining the probe's initial position information and the three-dimensional coordinates of the target area. The initial position information, including the probe's three-dimensional coordinates and posture angle, can be determined through a real-time pose sensor of the ultrasound probe or through a pre-calibrated coordinate system. The three-dimensional coordinates of the target area can be obtained through a preset medical imaging region or through external equipment (such as CT or MRI image registration).

[0038] After obtaining the initial position information, the probe's three-dimensional motion path is determined based on this information. The three-dimensional motion path includes not only the probe's translational motion but also its rotational motion around its own axis. The motion path should consider the kinematic constraints of the robot arm, including the joint angles of the six-DOF robot arm, the rotational range of each joint, and the maximum and minimum extension range. The three-dimensional motion path can be planned using a trajectory interpolation algorithm. By setting the starting and target points, a smooth three-dimensional path is generated. This path can be generated using a cubic spline interpolation algorithm, ensuring that the pose of each point along the path is continuous and without abrupt changes.

[0039] Based on a defined 3D motion path, a kinematic model of a six-DOF robotic arm is established. The kinematic model includes forward kinematics and inverse kinematics. Forward kinematics describes the calculation of the probe's position and orientation in 3D space given the angles of each joint, while inverse kinematics calculates the rotation angles of each joint given the target position and orientation. The inverse kinematics model is solved using analytical solutions or numerical iteration methods based on the joint parameters of the six-DOF robotic arm. Analytical solutions are suitable for simple joint structures, while numerical iteration methods are suitable for complex multi-joint robotic arms.

[0040] After calculating the inverse kinematics parameters of each joint, the probe is driven to move along a three-dimensional motion path by the joint controller of the six-DOF robot arm. The joint controller uses a feedback control strategy to adjust the angle and angular velocity of each joint in real time. The feedback control employs a proportional-integral-derivative (PID) control algorithm to ensure smooth movement of the probe along the predetermined path while avoiding mechanical vibration. The controller monitors the deviation between the probe's actual pose and the preset path by reading encoder feedback information from each joint and generates correction commands based on the deviation.

[0041] During the probe's movement along the three-dimensional motion path, its spatial pose data is monitored in real time. This spatial pose data includes the probe's coordinates in three-dimensional space and its rotation angles around each axis. This data can be acquired in real time via the robot arm's built-in encoder, angle sensors, or external optical tracking devices. Real-time monitoring helps to dynamically correct for deviations from the preset path or interference from external forces, ensuring that the probe always remains on the predetermined path.

[0042] Based on preset safety thresholds, the three-dimensional motion trajectory is adjusted to ensure the probe's movement safety. Safety thresholds include maximum speed, maximum acceleration, and movement limits for each joint. When the probe's speed or acceleration exceeds the threshold, the controller automatically slows the probe's movement or, if necessary, stops it. Safety thresholds may also include the minimum safe distance between the probe and the patient or other objects to ensure the probe is not damaged by accidental contact during operation.

[0043] After the probe reaches the target area and completes initial positioning, the system proceeds to the subsequent operation process. The completion of initial positioning means that the position and attitude of the probe in three-dimensional space have been calibrated, and subsequent force control, visual servo control, and image acquisition are all carried out on this basis.

[0044] In one implementation, the six-DOF robotic arm employs a serial structure, comprising three translational joints and three rotational joints. Initial position information is calibrated using a laser tracker, with the calibration results represented by three-dimensional coordinates and quaternions, indicating the probe's spatial position and orientation. The three-dimensional motion path is generated using a cubic spline interpolation algorithm, with the path's starting point being the coordinates recorded in the initial position information and the target point being the coordinates of the tissue region determined through image registration. Inverse kinematics is solved using a numerical iterative method, taking into account the physical constraints of the joint angles.

[0045] In another implementation, the robotic arm employs a parallel structure, with six drive motors controlling six degrees of freedom. The three-dimensional motion path utilizes Dijkstra's shortest path algorithm to prevent collisions with other devices during movement. Initial position information is acquired in real-time via the robotic arm's built-in angle encoder and position sensors, and inverse kinematics is solved analytically to ensure computational speed. The movement of each joint is limited by its maximum speed and maximum acceleration, automatically decelerating when thresholds are exceeded.

[0046] In the third embodiment, the controller of the six-DOF robotic arm employs a fuzzy control algorithm to automatically adjust joint angles based on the real-time pose data of the probe. The controller monitors the spatial position of the probe in real time through a visual servoing system and automatically corrects the trajectory when the probe deviates from the preset path. The visual servoing system uses a binocular camera to achieve precise probe positioning through image registration and target tracking.

[0047] Example Description: In an ultrasound elastography system, a robotic arm drives an ultrasound probe along a three-dimensional path along the patient's abdomen. The system first determines the three-dimensional coordinates of the target area based on the patient's CT scan data, and then uses a six-degree-of-freedom robotic arm to achieve precise probe positioning. The probe moves along a cubic spline interpolation path to avoid positional jitter during movement. The probe's spatial pose is monitored in real time, and its position is further fine-tuned via a visual servo system as it approaches the target area, ensuring the probe remains aligned with the target tissue region. Even during patient breathing or slight body movements, the robotic arm automatically adjusts the probe's position to maintain imaging stability.

[0048] This embodiment achieves high-precision pose control of the probe within a three-dimensional coordinate system by controlling the robotic arm to drive its three-dimensional spatial motion. The six-degree-of-freedom structure of the robotic arm ensures that the probe can move freely in three-dimensional space without manual adjustment, reducing operational difficulty. The inverse kinematics model and feedback control strategy ensure that the probe's motion path is smooth and vibration-free, reducing imaging errors caused by mechanical vibration. Safety threshold monitoring and dynamic correction mechanisms ensure the probe's safety and reliability during movement, avoiding risks caused by excessive speed or abnormal displacement.

[0049] S20, based on the control of the three-dimensional spatial motion, the contact force applied by the probe to the elastic medium is monitored by a force sensor, and the contact force is adjusted according to a preset force control algorithm;

[0050] In this embodiment, under the control of three-dimensional spatial motion, a force sensor monitors the contact force applied by the probe to the elastic medium. The force sensor is used to measure the force generated by the probe during the contact process in real time. This force can be pressure perpendicular to the surface of the elastic medium or frictional force tangential to the surface. The choice of force sensor can be adjusted according to the probe's installation method and application scenario. For example, a six-axis force sensor can simultaneously measure force and torque in three directions, making it suitable for multi-degree-of-freedom robot arms. Single-axis or three-axis force sensors are suitable for applications that only require detection of vertical pressure.

[0051] The contact force data acquired by the force sensor is represented in the sensor coordinate system. This coordinate system may not completely coincide with the probe's contact point coordinate system. Therefore, it is first necessary to transform the force sensor data from the sensor coordinate system to the probe's contact point coordinate system. The coordinate system transformation process involves constructing a coordinate system transformation matrix based on the relative positional relationship between the probe and the force sensor. This transformation matrix can be calculated directly through calibration experiments or by analyzing the probe's geometry.

[0052] After coordinate system transformation, based on the transformed contact force data, the error between the current contact force applied by the probe to the elastic medium and the preset expected contact force is calculated. The preset expected contact force can be set according to specific application scenarios. For example, in tissue elastography, the expected contact force can be within a range that can cause tissue deformation but will not damage the tissue. The formula for calculating the force error is:

[0053] ef = Factual - FExpected

[0054] Where ef represents the force error, Factual represents the currently measured contact force, and Fexpected represents the preset expected contact force. The force error can be a scalar (considering only the vertical direction) or a vector (considering multiple directions simultaneously).

[0055] The force error is dynamically adjusted through a feedback control loop. This loop can employ various control algorithms, including proportional-integral-derivative (PID) control, fuzzy control, or adaptive control. PID control is a classic method where the proportional term is used for rapid response to force error, the integral term eliminates steady-state error, and the derivative term suppresses instantaneous fluctuations. The control output of the feedback control loop directly affects the robot arm's drive signal, ensuring that the contact force applied by the probe to the elastic medium always tends towards the preset desired range.

[0056] In practical implementation, the control parameters of the feedback control loop (such as gain, integral time, and derivative time) can be adjusted according to different tissue types and detection targets. For example, for softer tissues, the proportional gain can be reduced to avoid applying excessive pressure to the probe; while for harder tissues, the proportional gain can be increased to ensure that the force quickly reaches the desired value. The update frequency of the feedback control loop should be consistent with the data acquisition frequency of the force sensor to avoid control lag.

[0057] Building upon the feedback control loop, a control mode that periodically alternates between applying maximum and minimum contact force can be introduced to simulate pressure variations during palpation. In this control mode, the probe alternates between applying maximum and minimum contact force within one control cycle, simulating the light and heavy pressure applied by a physician during palpation. Periodic control can be generated using a sine wave or square wave function and dynamically adjusted based on the probe's contact force error. For example, the values ​​of the maximum and minimum contact force can be adaptively adjusted based on the elastic properties of the tissue.

[0058] In one embodiment, the force sensor is a six-axis force sensor, mounted between the probe and the robot arm, capable of measuring force and torque in three axes. The origin of the sensor coordinate system is located at the center of the sensor, while the origin of the probe's contact point coordinate system is located at the contact point on the probe surface. The coordinate transformation matrix is ​​calculated using geometric calibration and torque balancing methods, and the force sensor data is converted into force data in the probe contact point coordinate system in real time.

[0059] In another implementation, the force control algorithm employs PID control, with a proportional gain set to 0.8, an integral time of 0.1 seconds, and a derivative time of 0.01 seconds. The system acquires force sensor data at a frequency of 1000 times per second and calculates the force error in real time. The output of the feedback control loop directly drives the robot arm's drive motor, stabilizing the probe's contact force within a preset desired range.

[0060] In the third implementation, force control employs a periodic oscillation mode with a period of 1 second, where the maximum contact force is applied for the first 0.5 seconds and the minimum contact force is applied for the next 0.5 seconds. The values ​​of the maximum and minimum contact forces are automatically adjusted according to the elastic properties of the tissue; specifically, the maximum contact force is increased when the tissue is stiffer, and the minimum contact force is decreased when the tissue is softer. This periodic force application mode is suitable for elastography scenarios and can more clearly reflect the elastic properties of the tissue.

[0061] Example Description: In an ultrasound elastography application, a robotic arm drives a probe to contact the patient's abdominal tissue, and a force sensor monitors the contact force applied by the probe in real time. The system's preset desired contact force is 3 to 5 Newtons. Through PID control, the probe's contact force fluctuates within this range and stabilizes within 2 seconds. The physician can observe the probe's contact force curve in real time through the system interface and dynamically adjust the control effect by adjusting the proportional gain. To enhance the elastic contrast of tissues, the system introduces a periodic oscillation mode in the force control, with the probe alternately applying pressures of 4 Newtons and 2 Newtons per second. On the ultrasound image, the elastic differences between different tissues are clearly visible, especially tumor tissue, which exhibits a significantly high-rigidity area. Even with slight movement of the patient during breathing, the system can still automatically adjust the probe's contact force through feedback control, ensuring the stability and reliability of the imaging.

[0062] This embodiment utilizes three-dimensional spatial motion control, combined with force sensor monitoring of the contact force applied by the probe to the elastic medium, and dynamically adjusts the contact force through a preset force control algorithm. This ensures that the probe remains within a safe and effective force range throughout the contact process. The introduction of a feedback control loop enables rapid response and stable control, preventing tissue damage due to excessive force or failure to obtain effective data due to insufficient force. The periodic force application mode simulates the actions of a doctor's palpation, enhancing the contrast of tissue deformation during elastography and improving image quality by alternately applying maximum and minimum contact forces.

[0063] S30, during the process of adjusting the contact force, the real-time position of the target is tracked by visual servo control technology, and a translation control command is generated based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0064] In this embodiment, during the adjustment of the contact force, the position of the target is tracked and detected in real time using visual servo control technology. Visual servo control technology is a closed-loop control method based on image feedback. It acquires and processes image information of the target in real time to generate control commands to adjust the movement of the mechanical system. Visual servo control technology uses ultrasound image data acquired by an ultrasound probe to achieve real-time monitoring of the target biological tissue position. Ultrasound image data is real-time and non-invasive, providing the target tissue's position information without damaging the tissue.

[0065] First, the ultrasound probe acquires ultrasound image data of the target biological tissue in real time. This image data can be ultrasound radio frequency signal data or ultrasound B-mode image data. Radio frequency signal data provides high-precision raw waveform information, suitable for high-precision displacement detection; B-mode image data displays the structural information of the target tissue in the form of a two-dimensional grayscale image, suitable for direct visual servo control. During the ultrasound image data acquisition process, the probe's operating frequency and gain can be dynamically adjusted according to the tissue type and target depth to ensure image quality.

[0066] After acquiring ultrasound image data, the system preprocesses the image data. Preprocessing includes edge enhancement, noise filtering, and contrast adjustment. Edge enhancement highlights the contour features of the target tissue, for example, by using edge detection algorithms (such as Sobel or Canny edge detection) to extract tissue boundaries. Noise filtering can employ Gaussian filtering, median filtering, or bilateral filtering methods to reduce the speckle noise inherent in ultrasound images. Contrast adjustment ensures a clear brightness difference between the tissue area and the background area, facilitating subsequent feature extraction.

[0067] In the preprocessed ultrasound image, the system calculates the real-time position of the detected target based on contour features. The target's position can be determined by calculating the centroid coordinates of its contour. The centroid coordinates are calculated based on the position and brightness value of each pixel in the binarized image. Specifically, the horizontal centroid coordinates are obtained by summing the horizontal coordinates of all pixels in the target region and dividing by the total number of pixels; similarly, the vertical coordinates are obtained by summing the vertical coordinates and dividing by the total number of pixels. In this way, the system can track the target's position information in real time during dynamic imaging.

[0068] After obtaining the target's real-time centroid coordinates, the system compares these coordinates with the preset field-of-view center coordinates to calculate the positional deviation. The preset field-of-view center coordinates represent the center of the probe's desired observation area. For example, in tissue elastography, the preset field-of-view center can be set at the center of the target tissue (such as a tumor). The positional deviation is the difference between the target's centroid coordinates and the field-of-view center coordinates; the larger the difference, the farther the target deviates from the center of the field of view.

[0069] The system generates translation control commands based on the position deviation. These commands can be generated using a proportional control algorithm, where the translation speed is proportional to the position deviation. Proportional control ensures a rapid response to changes in the target position, making it suitable for dynamic scenarios. For scenarios with high stability requirements, incremental proportional-integral-derivative control can be used, adjusting the control response speed through control gain while eliminating long-term deviations and suppressing oscillations.

[0070] The generated translational control commands are directly transmitted to the translational joints of the robotic arm, driving the arm to adjust the probe's position so that the target's centroid gradually approaches and remains within the preset field of view center. The robotic arm's translational motion can occur in three-dimensional space, depending on the probe's mounting position and control strategy. In tissue elastography, the robotic arm typically adjusts the probe's position only in a plane, ensuring the probe remains perpendicular to the tissue surface. However, in three-dimensional or multi-angle imaging, the robotic arm can move freely in three-dimensional space to cover target areas at more angles.

[0071] Example Description: In an application of abdominal tissue elastography, a robotic arm drives an ultrasound probe to contact the patient's abdominal region. The system acquires real-time image data of the abdominal tissue via ultrasound imaging. Visual servo control technology extracts the contour of the target tumor using an edge detection algorithm and calculates the tumor's centroid coordinates. Due to the patient's breathing causing vertical movement of the abdominal tissue, the system calculates the changes in the centroid coordinates in real time and generates translation control commands via proportional control to drive the probe to move in the horizontal plane. Even when the patient is breathing deeply, the real-time adjustment of the probe ensures that the tumor remains centered in the field of view. On the ultrasound image, the tumor's edges are clearly visible, the tissue elasticity distribution is stable, and the image shows no significant jitter or blurring. Doctors can observe the tumor's elasticity changes in real time on the interface and adjust visual servo control parameters (such as proportional gain) as needed to obtain more stable image quality.

[0072] This embodiment utilizes visual servo control technology to track the real-time position of the target, ensuring the probe remains aligned with the target area and preventing imaging shifts due to changes in target position or probe movement. Through image preprocessing and centroid coordinate calculation, the system can accurately identify the target position in dynamic scenes and maintain the target centroid at the preset center of the field of view through real-time feedback control. Compared to traditional manual probe control, it offers advantages in high precision and stability, making it particularly suitable for dynamic tissue (such as heart and lung) detection and elastography. Even when the patient is experiencing physiological movements such as breathing and heartbeat, the system can automatically compensate for probe translational movement through feedback control, ensuring image quality.

[0073] S40, during the process of adjusting the translational motion, receives rotation commands through the remote operation device and generates rotation control commands to adjust the rotation angle of the probe;

[0074] In this embodiment, during the adjustment of translational motion, a rotation command is received via a teleoperation device, which generates a rotation control command to adjust the probe's rotation angle. The teleoperation device is an input device capable of generating control commands through user operation; it can be a physical joystick, knob, virtual control interface, or force feedback device with tactile feedback. The teleoperation device is used to control the rotation angle of the ultrasound probe, enabling multi-angle imaging or dynamic adjustment of the target tissue. The teleoperation control method allows the operator to finely adjust the probe's rotation direction without direct contact, ensuring the probe is aligned with the target area.

[0075] First, the system receives rotation angle commands input by the user from the teleoperated device. These commands can be absolute rotation angles (specifying the absolute rotation angle of the probe relative to its initial orientation) or relative rotation angles (specifying the incremental rotation of the probe from the current angle). Input data from the teleoperated device can be transmitted to the control system via analog or digital signals, depending on the device type. For example, the rotation angle of a physical joystick can be converted into a voltage signal using a potentiometer, while a virtual interface generates digital commands through touch coordinates or rotation gestures.

[0076] Upon receiving a rotation angle command, the system first converts it into the desired rotation direction in the probe contact point coordinate system. The probe contact point coordinate system is a local coordinate system established with reference to the plane of the tissue contacted by the probe, with its origin located at the probe's contact point. During the conversion, the system maps the rotation angle command of the teleoperated device to the probe contact point coordinate system based on the probe's mounting position on the robot arm and its current spatial orientation. This coordinate system transformation ensures that the rotation command accurately applies to the probe's actual rotation direction, without errors caused by changes in the probe's position and orientation in space.

[0077] Next, the system calculates the rotation angle error based on the deviation between the desired rotation direction and the probe's current rotation angle. The rotation angle error is the difference between the current rotation angle and the desired rotation direction, usually expressed in degrees (or radians). This error value reflects the degree of deviation between the probe's actual attitude and the desired attitude. The rotation angle error can be positive (indicating clockwise deviation) or negative (indicating counterclockwise deviation), and a larger error value indicates a greater deviation between the probe's attitude and the target attitude.

[0078] To ensure the probe can quickly and smoothly adjust to the desired angle, the system uses a proportional control algorithm to adjust the gain of the rotation angle error. Proportional control is a classic closed-loop control method where the output of the control command is proportional to the error value. The control gain is the core parameter of this algorithm, used to adjust the system's response speed to errors. A larger control gain can improve the response speed but may also cause oscillations; a smaller control gain ensures stability but results in a slower response. The control gain can be dynamically adjusted based on the characteristics of the target tissue (such as hardness or elasticity) and the type of probe (such as a linear array probe or a phased array probe).

[0079] Based on the adjusted rotation angle control command, the system drives the probe to rotate to the target angle via the rotary joints of the robot arm. The rotary joints of the robot arm can be motor-driven (such as servo motors) or hydraulically driven (such as hydraulic motors), depending on the robot arm's design. During rotation, the system continuously monitors the probe's actual rotation angle and compares it in real time with the desired rotation direction to ensure stable and overshoot-free rotation. In dynamic tissue detection (such as heart and lungs), the probe's rotation can be adjusted in real time to ensure it remains aligned with the target area throughout tissue movement.

[0080] Example Description: In one application of cardiac tissue elastography, a physician controls the rotation angle of an ultrasound probe using a teleoperated joystick with haptic feedback. The teleoperated device converts the physician's rotation input signal into the rotation direction in the probe's contact point coordinate system. Due to the periodic movement of tissue caused by the heartbeat, the system adjusts the probe's rotation angle in real time using a proportional control algorithm, ensuring the probe is always aligned with the left ventricle region of the heart. On the imaging interface, the tissue boundaries of the heart remain clear, and the tissue elasticity distribution is stable, allowing the physician to observe the strain changes of the heart from different angles. In another scenario, abdominal tumor detection, the physician inputs the rotation angle through a virtual control interface, and the probe rotates to different angles to acquire multi-angle images of the tumor. The system automatically adjusts the probe's rotation angle according to the user input, ensuring that the image is stable and blur-free after each rotation. Through multi-angle imaging, physicians can comprehensively observe the tumor's tissue structure and elasticity distribution, improving diagnostic accuracy.

[0081] This embodiment receives rotation commands via a teleoperated device, enabling precise control of the probe's rotation angle and avoiding rotational instability caused by hand tremors or viewing angle errors in traditional manual operation. A proportional control algorithm based on rotation angle error ensures the probe responds quickly to rotation commands and remains aligned with the target area during dynamic tissue (such as heart and lung) detection. Whether using a physical joystick, a virtual control interface, or a teleoperated device with haptic feedback, users can adjust the probe's rotation direction in real time to adapt to different tissues and detection scenarios. Furthermore, coordinate system transformation converts the teleoperated input into the rotation direction in the probe contact point coordinate system, avoiding control deviations caused by changes in probe installation position and orientation. Even in complex 3D tissue imaging, the system ensures the probe remains aligned with the target area, achieving high-quality ultrasound imaging.

[0082] S50, based on the adjusted contact force, translational motion and rotation angle, collects detection signal data in the pre-compression state and the post-compression state, and calculates and generates a three-dimensional strain distribution map;

[0083] In this embodiment, based on the adjusted contact force, translational motion, and rotation angle, detection signal data are collected under pre-compression and post-compression states, and a three-dimensional strain distribution map is calculated and generated. In ultrasonic elastography, a three-dimensional strain distribution map is a three-dimensional image that reflects the elastic characteristics of the target tissue. It is calculated by analyzing the displacement changes of the tissue under different compression states. Its core lies in using ultrasonic radio frequency signal data collected by an ultrasonic probe under different compression states, combined with a displacement estimation algorithm, to calculate the strain generated by the tissue during compression.

[0084] First, ultrasound radiofrequency signal data of the target biological tissue are acquired using an ultrasound probe under pre-compression and post-compression conditions. Pre-compression refers to the acquisition of signal data with minimal probe contact force, while post-compression refers to the acquisition of signal data with greater probe contact force. Ultrasound radiofrequency signal data is a high-frequency, uncompressed, raw ultrasound echo data that provides higher resolution and finer tissue structure information than standard B-mode ultrasound images. By acquiring radiofrequency signal data under different compression conditions, the dynamic response of the tissue to pressure changes can be recorded.

[0085] Next, based on the radio frequency signal data acquired under the pre-compressed and post-compressed states, the system employs a motion estimation algorithm to perform three-dimensional displacement matching and obtain tissue displacement field data. The motion estimation algorithm is a method for estimating tissue displacement by analyzing pixel displacements between adjacent frames or adjacent volume data. The system uses three-dimensional optical flow, phase correction, or correlation methods for displacement estimation; the specific algorithm chosen depends on the dynamic characteristics of the tissue and the availability of computational resources. The three-dimensional optical flow method estimates three-dimensional displacement by calculating image gradients and the rate of change over time, suitable for cases of slow tissue movement; the phase correction method determines displacement by comparing the phase difference between pre-compressed and post-compressed radio frequency signals, suitable for cases of rapid tissue movement; and the correlation method estimates displacement by calculating the similarity between image blocks, suitable for cases with distinct tissue texture. The tissue displacement field data obtained through the motion estimation algorithm includes the displacement vector of each pixel or voxel in three-dimensional space, reflecting the deformation of the tissue under pressure.

[0086] After obtaining the tissue displacement field data, the system calculates the local strain values ​​based on the least squares method, generating a sequence of two-dimensional strain maps. The least squares method is an optimization algorithm based on minimizing the sum of squared errors, used to calculate local strain from displacement field data. Strain is a measure of tissue deformation, representing the rate of change of displacement per unit length. The local strain values ​​calculated using the least squares method accurately reflect the deformation characteristics of the tissue during compression. The two-dimensional strain map sequence is a set of images representing the strain distribution on different slices or cross-sections, with the grayscale value of each pixel representing the strain magnitude in that region.

[0087] Finally, the system spatially superimposes the two-dimensional strain map sequence along the probe's movement direction to generate a three-dimensional strain distribution map. A three-dimensional strain distribution map is a stereoscopic image where the grayscale value of each voxel represents the magnitude of tissue strain at that location. The three-dimensional strain distribution map can visually display the elastic distribution of target tissue at different locations, and is suitable for clinical assessment of tumor stiffness, identification of hard tissue regions, or analysis of tissue elastic characteristics.

[0088] This embodiment acquires radio frequency signal data under pre-compression and post-compression conditions, and generates a three-dimensional strain distribution map through three-dimensional displacement matching and least squares calculation, which can accurately reflect the elastic distribution characteristics of the target tissue. Compared with traditional two-dimensional strain imaging methods, the three-dimensional strain distribution map can provide strain information of the target tissue in three-dimensional space, enabling doctors to more comprehensively assess the tissue's stiffness and elasticity. Furthermore, by dynamically adjusting the probe's contact force, translational motion, and rotation angle, the stability and consistency of radio frequency signal acquisition can be ensured, effectively reducing errors caused by tissue movement or probe position changes.

[0089] S60, perform threshold segmentation on the three-dimensional strain distribution map, extract the rigid region, and output the spatial location information of the rigid region.

[0090] In this embodiment, threshold segmentation is performed on the three-dimensional strain distribution map to extract rigid regions and output their spatial location information. A three-dimensional strain distribution map is a three-dimensional image reflecting the elastic characteristics of a target tissue at different locations; the grayscale value of each voxel (volume pixel) represents the magnitude of tissue strain at that location. Through threshold segmentation, rigid regions—that is, parts of the tissue with low elasticity and high stiffness—can be extracted from the three-dimensional strain distribution map. This is of great significance for detecting hard lesions (such as tumors).

[0091] First, the system determines the segmentation threshold based on a preset calculation method. The segmentation threshold is a strain value threshold used to distinguish between rigid and flexible regions, typically set based on the elastic properties of the target tissue and clinical experience. The calculation method for the segmentation threshold usually involves a combination of a preset center value and a preset percentage, where the segmentation threshold is the sum of the preset center value and the preset percentage multiplied by the maximum absolute value of the strain in the three-dimensional strain distribution map. The preset center value represents the basic strain level for segmentation, while the preset percentage allows for dynamic adjustment of the threshold based on tissue characteristics, making segmentation more flexible.

[0092] After determining the segmentation threshold, the system binarizes the strain value of each voxel in the 3D strain distribution map. Binarization is a process of converting continuous strain values ​​into a binary state, typically achieved by comparing the strain value with a segmentation threshold. Voxels with strain values ​​below the segmentation threshold are marked as rigid regions, while voxels with strain values ​​above the threshold are marked as non-rigid regions. This binarization ensures clear boundaries for rigid regions in the 3D strain distribution map, which is helpful for subsequent region extraction and spatial location calculation.

[0093] After binarization, the system employs connected component analysis (CBI) to extract the maximum connected volume within the rigid regions. CBI is an image processing technique used to identify and separate spatially connected regions. In a 3D strain distribution map, CBI can automatically detect and label all spatially connected rigid regions and filter out the largest rigid region based on its size. It effectively eliminates small regions caused by noise or local strain fluctuations, ensuring the stability and reliability of the segmentation results.

[0094] Finally, the system calculates the spatial location information of the maximum connected volume. This spatial location information typically includes the three-dimensional boundary coordinates and centroid coordinates of the rigid region. The three-dimensional boundary coordinates describe the spatial extent of the rigid region in the three-dimensional strain distribution map, while the centroid coordinates represent the spatial center of the rigid region. The centroid coordinates are usually calculated using the spatial centroid formula, which involves taking a weighted average of the spatial coordinates of all voxels within the rigid region to obtain the center point of that region. The centroid coordinates provide a clear visual representation of the rigid region's location in three-dimensional space, facilitating target tissue localization and diagnosis in clinical practice.

[0095] In one implementation, the system sets the segmentation threshold to be a preset center value plus a preset percentage multiplied by the sum of the maximum absolute strain values ​​in the 3D strain distribution map. The preset center value is set to 0.2, and the preset percentage is set to 0.5. In the generated 3D strain distribution map, the maximum absolute strain value is 0.8, so the segmentation threshold is calculated as 0.2 + 0.5 × 0.8 = 0.6. The system binarizes the strain value of each voxel in the 3D strain distribution map, and voxels with strain values ​​below 0.6 are marked as rigid regions.

[0096] In another implementation, the system uses an adaptive threshold segmentation algorithm to dynamically adjust the segmentation threshold based on the elastic properties of the target tissue. Specifically, the system first sets a base segmentation threshold based on the average strain value of the target tissue, and then dynamically adjusts the threshold according to the size of the rigid region selected by the clinician. When the rigid region is small and scattered, the system automatically lowers the threshold to expand the coverage of the rigid region; when the rigid region is large and continuous, the system automatically raises the threshold to reduce artifact areas.

[0097] In the third implementation, the connected component analysis algorithm is implemented using a three-dimensional depth-first search (DFS). The system traverses all voxels in the three-dimensional strain distribution map. When a voxel belonging to a rigid region is detected, the system starts from that voxel and recursively searches for adjacent rigid region voxels along the three-dimensional direction, marking and counting the number of all connected voxels. Finally, the system selects the connected region with the largest volume and records its spatial coordinates and centroid position.

[0098] In the fourth implementation, the system improves accuracy during the centroid coordinate calculation process using a weighted average algorithm. Specifically, the system adds the product of the coordinates of each voxel in the largest connected volume and its strain value, and divides the sum by the sum of the strain values ​​of all voxels in the rigid region to obtain the weighted centroid coordinates. This weighted calculation method can more accurately reflect the actual location of the rigid region, and is particularly suitable for tissues with uneven strain distribution.

[0099] This embodiment effectively identifies and locates hard lesions in target tissues by performing threshold segmentation on the three-dimensional strain distribution map and extracting rigid regions. Compared with traditional manual judgment methods, the automated method based on threshold segmentation and connected component analysis has higher accuracy and repeatability, and can eliminate subjective biases caused by human operation. The rigid regions in the three-dimensional strain distribution map are represented by spatial coordinates and centroid coordinates, which can provide three-dimensional location information of the target tissue, facilitating doctors to quickly and accurately locate and diagnose lesions in clinical practice.

[0100] This invention relates to the field of robot control technology and discloses a method, apparatus, device, and medium for region detection based on robot control. The method includes: controlling a robot arm to drive the three-dimensional spatial motion of a probe; based on the control of the three-dimensional spatial motion, monitoring the contact force applied by the probe to an elastic medium using a force sensor, and adjusting the contact force according to a preset force control algorithm; during the adjustment of the contact force, using visual servo control technology to track the real-time position of the detection target, and generating translation control commands based on the deviation between the real-time position and a preset field of view center to drive the robot arm to adjust the translational motion of the probe; during the adjustment of the translational motion, receiving rotation commands through a teleoperation device, generating rotation control commands to adjust the rotation angle of the probe; based on the adjusted contact force, translational motion, and rotation angle, collecting detection signal data under pre-compression and post-compression states, calculating and generating a three-dimensional strain distribution map, performing threshold segmentation processing on the three-dimensional strain distribution map, extracting rigid regions, and outputting the spatial position information of the rigid regions. This invention achieves high-precision dynamic control of the probe in three-dimensional space by controlling a robotic arm to drive the probe, combined with real-time force feedback from a force sensor, target position tracking using visual servo technology, and rotation control from a teleoperated device. This ensures stable contact between the probe and the target under preset contact force, translation, and rotation conditions. Through the acquisition of detection signals and the calculation of three-dimensional strain distribution maps under pre-compression and post-compression states, the elastic properties of the target tissue can be accurately reflected, improving imaging accuracy and the detection accuracy of rigid tissue regions, and significantly reducing operational errors.

[0101] In one embodiment, step S10 above includes:

[0102] S101, Based on the initial position information of the target area, determine the three-dimensional motion path of the ultrasonic probe;

[0103] S102, Based on the three-dimensional motion path, establish a kinematic model of the six-degree-of-freedom robot arm, and calculate the inverse kinematics solution parameters of each joint of the six-degree-of-freedom robot arm;

[0104] S103, Based on the inverse kinematics solution parameters, the joint controller of the six-degree-of-freedom robot arm drives the ultrasonic probe to move along the three-dimensional motion path, generating a continuous three-dimensional motion trajectory of the ultrasonic probe.

[0105] S104, during the movement, monitor the spatial pose data of the ultrasonic probe and adjust the three-dimensional motion trajectory according to the preset safety threshold to complete the initial positioning of the ultrasonic probe.

[0106] In this embodiment, the robotic arm is controlled to drive the probe's three-dimensional spatial motion. This process involves planning the ultrasound probe's three-dimensional motion path based on the initial position information of the target area, ensuring that the probe can achieve precise positioning and stable movement in three-dimensional space. Three-dimensional spatial motion is one of the most fundamental and critical tasks in robotic arm control, especially in ultrasound imaging, where the probe's spatial position and orientation directly affect image quality and diagnostic accuracy.

[0107] First, the system determines the three-dimensional motion path of the ultrasound probe based on the initial position information of the target area. This initial position information is typically selected manually by the physician or obtained automatically, and may include the coordinates of the tumor center, organ surface, or specific anatomical structures. In one implementation, the system uses a three-dimensional coordinate system to represent the position of the target area (e.g., X, Y, Z coordinates). Based on this initial position information, combined with a preset safety zone and the workspace of the robotic arm, the system generates the three-dimensional motion path of the ultrasound probe. The three-dimensional motion path can be represented by a discrete set of points, where each point contains three-dimensional coordinates and attitude information (e.g., Euler angles or rotation angles represented by quaternions).

[0108] Secondly, the system establishes a kinematic model of the six-DOF robot arm based on the three-dimensional motion path. A six-DOF robot arm is a robotic arm capable of independently controlling six degrees of freedom in three-dimensional space, including three translational degrees of freedom (X, Y, and Z directions) and three rotational degrees of freedom (rotation about the X, Y, and Z axes). The kinematic model describes the spatial position and orientation of each joint of the robot arm, as well as the geometric relationships between the joints. The establishment of the kinematic model typically employs forward kinematics and inverse kinematics methods. Forward kinematics describes the spatial position of the probe given a joint angle, while inverse kinematics is used to calculate the joint angles based on the target spatial position.

[0109] In inverse kinematics (IK) calculation, the system calculates the joint angles of a six-DOF robot arm based on each path point in the three-dimensional motion path. IK is a mathematical method for determining the rotational angles and displacements of each joint in a robot arm. The calculation methods can employ numerical iteration, analytical methods, or deep learning-based approximation algorithms. Analytical methods are suitable for robot arms with well-defined geometry, such as six-DOF robotic arms with fixed link lengths and known joint types. Numerical iteration methods are suitable for complex robot structures where there are coupling relationships between joints. The system transmits the joint angle parameters obtained from the IK to the joint controller to drive the motors of each joint.

[0110] After obtaining the inverse kinematics parameters, the system drives the ultrasound probe to move along a three-dimensional motion path through the joint controllers of the six-DOF robot arm. A joint controller is a control device that receives joint angle parameters and drives the joint motors to achieve precise angular rotation or displacement. The system converts the joint angle parameters obtained from the inverse kinematics calculation into joint control signals, which are transmitted to each joint controller via a robot control bus (such as CAN bus or EtherCAT bus). Each joint controller adjusts its joint position according to the received control signals, ensuring that the ultrasound probe moves smoothly along the predetermined three-dimensional motion path.

[0111] As the probe moves along its three-dimensional motion path, the system continuously monitors its spatial pose data. This spatial pose data is typically acquired by sensors at the end effector of the robotic arm (such as laser rangefinders, inertial measurement units, and encoders) and represented in the form of three-dimensional coordinates and attitude angles. The system calculates the probe's current spatial position and attitude in real time and compares it with a preset position within the three-dimensional motion path. If the probe's actual position deviates from the preset path, the system determines the degree of deviation based on a preset safety threshold. The safety threshold is a maximum permissible deviation range used to prevent the probe from colliding with surrounding tissues or equipment.

[0112] When the probe's actual pose exceeds a safety threshold, the system automatically adjusts its three-dimensional motion trajectory. Adjustment methods may include replanning the motion path, increasing or decreasing the probe's translational and rotational speeds, or correcting the probe's spatial position through a feedback control loop. Ultimately, the system ensures that the probe always moves along a safe three-dimensional path during the initial positioning process, avoiding collisions with surrounding objects while maintaining probe posture stability.

[0113] This embodiment controls the robot arm to drive the probe's three-dimensional spatial movement, ensuring precise positioning and stable motion of the probe in three-dimensional space. Compared to traditional manual operation, robot control can automatically achieve three-dimensional path planning and precise control, eliminating positional offsets and angular errors caused by human operation. The combination of inverse kinematics calculation and joint controllers ensures smooth movement of the probe along a predetermined path, maintaining high precision even in complex three-dimensional spaces. A safety threshold monitoring mechanism further enhances the system's safety, preventing the probe from colliding with surrounding tissues or equipment.

[0114] In one embodiment, step S20 above includes:

[0115] S201, based on the control of the three-dimensional spatial motion, the contact force data of the probe applied to the biological tissue collected by the force sensor is transformed from the sensor coordinate system to the contact point coordinate system of the probe;

[0116] S202, Calculate the force error between the current contact force and the preset expected contact force based on the converted contact force data;

[0117] S203, the force error is adjusted through a feedback control loop so that the actual contact force enters and is maintained within the preset desired contact force range;

[0118] S204 simulates the changes in palpation pressure on biological tissues based on a control mode that periodically alternates between applying maximum and minimum contact force.

[0119] In this embodiment, based on three-dimensional spatial motion control, a force sensor monitors the contact force applied by the probe to the elastic medium, and adjusts the contact force according to a preset force control algorithm. In ultrasound imaging and tissue elasticity assessment, the probe's contact force is crucial to image quality and the accuracy of tissue deformation assessment. Excessive contact force may damage tissue, while insufficient contact force may result in unclear signals. To ensure that the probe maintains a stable and controllable contact force when in contact with biological tissue, the system uses a force sensor for real-time monitoring and a feedback control algorithm for dynamic adjustment.

[0120] First, a force sensor is mounted at the tip of the probe, enabling real-time monitoring of the contact force applied to the biological tissue. Force sensors are typically multi-axis sensors, capable of simultaneously measuring force data in three directions (X, Y, and Z). This force data is initially based on the sensor's own coordinate system, i.e., the force sensor coordinate system. However, since the relative positions of the sensor and the probe contact point may change, the force data needs to be converted to the probe contact point coordinate system. This coordinate system transformation is crucial for ensuring the accuracy of the force data. The system converts the force data in the sensor coordinate system to the probe contact point coordinate system using a coordinate transformation matrix. The transformation process includes: first, generating a transformation matrix based on the spatial position and orientation of the probe and sensor; then, multiplying the contact force data acquired by the sensor by the transformation matrix to obtain the contact force data in the probe contact point coordinate system. This coordinate transformation eliminates force data deviations caused by probe rotation or tilt, ensuring that the force data always reflects the true contact force between the probe and the biological tissue.

[0121] After the force data undergoes coordinate transformation, the system calculates the force error between the current contact force and the preset expected contact force based on the transformed contact force data. The preset expected contact force is a reference value set according to specific application requirements. For example, in tissue elastography, the contact force should be kept within a certain range to avoid excessive tissue deformation. The system calculates the force error by comparing the current contact force data with the preset expected contact force. The force error represents the deviation between the current contact force and the target contact force, and can be either positive (indicating excessive contact force) or negative (indicating insufficient contact force). The unit of force error is usually Newtons (N), and it can be calculated separately in three directions (X, Y, Z).

[0122] Next, the system dynamically adjusts the force error based on a feedback control loop. The feedback control loop is a classic closed-loop control method that can adjust the probe's contact force in real time, ensuring it enters and remains within a preset desired contact force range. The feedback control loop typically employs a proportional-integral-derivative (PID) control algorithm. The proportional control section immediately adjusts the probe position based on the magnitude of the force error, the integral control section makes minor corrections based on the cumulative force error, and the derivative control section adjusts the probe's speed based on the rate of change of the force error. The PID control algorithm can quickly calculate the control output based on the current force error, generating control commands to drive the probe to make minute displacement adjustments, ensuring the contact force remains within the target range.

[0123] Under this feedback control, if the contact force applied by the probe exceeds a preset upper limit, the control loop will immediately reduce the probe's pressing force; conversely, it will increase the pressing force. The response speed of the control loop can be optimized by adjusting PID parameters (such as proportional gain, integral time constant, and derivative time constant). The system can also dynamically adjust the PID parameters based on fluctuations in force error, making the control more stable.

[0124] Furthermore, the system can employ a control mode that periodically alternates between applying maximum and minimum contact forces to simulate changes in palpable pressure on biological tissues. This periodic force application mode is commonly used in tissue elastography, where the probe periodically applies and releases pressure, causing dynamic deformation of the tissue. Periodic force control achieves automatic palpation by preset maximum and minimum contact force thresholds and a periodic time parameter. The system monitors the actual contact force of the probe using a force sensor and uses a feedback control loop to cyclically vary the contact force between the maximum and minimum thresholds. The maximum contact force is typically used to compress the tissue, while the minimum contact force is used to restore tissue deformation. Periodic force application not only improves the accuracy of tissue elastography but also reduces pressure damage to the tissue from the probe.

[0125] This embodiment utilizes three-dimensional spatial motion control, employing force sensors to monitor the contact force applied by the probe to biological tissue. Based on a preset force control algorithm, dynamic adjustments are made to ensure that the pressure applied by the probe to the tissue surface remains within a safe and stable range. Compared to manual operation or fixed pressure control, the feedback control loop automatically adapts to the mechanical properties of the tissue and the probe's motion, maintaining a stable contact force. The periodic alternating force application mode further improves the accuracy of tissue elasticity assessment and helps enhance the resolution of rigid tissue regions.

[0126] In one embodiment, step S30 above includes:

[0127] S301, acquires ultrasound image data of target biological tissue through an ultrasound probe;

[0128] S302, perform edge enhancement and noise filtering on the ultrasound image data to extract the contour features of the target being detected;

[0129] S303, Calculate the real-time centroid coordinates of the detected target based on the contour features;

[0130] S304, Calculate the positional deviation between the real-time centroid coordinates and the preset field-of-view center coordinates, and generate translational motion error;

[0131] S305, convert the translational motion error into a translational speed control command for the probe based on a preset kinematic relationship;

[0132] S306, the translation speed control command is executed through the translation joint of the robot arm to make the real-time centroid coordinates of the detected target converge to the preset field of view center coordinates.

[0133] In this embodiment, during the adjustment of the contact force, visual servo control technology tracks the real-time position of the target and generates translation control commands based on the deviation between the real-time position and the preset field of view center, driving the robotic arm to adjust the translational movement of the probe. This visual servo control technology, combined with ultrasound image data, enables automatic alignment and position adjustment of the probe, thereby ensuring that the probe is always aligned with the target area and avoiding imaging deviations caused by tissue movement or probe drift.

[0134] First, ultrasound image data of the target biological tissue is acquired using an ultrasound probe. The ultrasound probe maintains stable contact with the biological tissue in three-dimensional space and generates ultrasound image data through ultrasound wave emission and echo reception. An ultrasound image is a grayscale image generated based on sound wave propagation and reflection, where brightness and contrast reflect the acoustic properties of the tissue. The system acquires ultrasound image data at a fixed frame rate (e.g., 30 frames per second) and transmits the image data to the image processing module in real time.

[0135] After acquiring ultrasound image data, the system performs edge enhancement and noise filtering to extract the contour features of the target. Edge enhancement is an image processing technique designed to improve the contrast of edge regions in an image, making the target contour clearer. Commonly used edge enhancement algorithms include the Sobel operator and the Laplacian operator. The system first eliminates high-frequency noise in the ultrasound image using Gaussian filtering, and then uses Sobel edge detection to extract the target contour. The edge-enhanced image retains the clear structure of tissue boundaries, which is helpful for subsequent feature extraction.

[0136] Noise filtering eliminates random noise in ultrasound images caused by signal interference or tissue scattering. The system uses median filtering or Gaussian filtering to remove noise, ensuring that the target contour is clearer and more stable in the image. Filtering parameters (such as the standard deviation of Gaussian filtering or the window size of median filtering) can be adjusted according to the characteristics of the target tissue. For example, in relatively smooth tissues (such as muscle tissue), a smaller filter window can preserve details; in high-noise environments (such as deep tissues), a larger filter window can effectively remove noise.

[0137] After image preprocessing, the system calculates the real-time centroid coordinates of the detected target based on the extracted contour features. The centroid coordinates represent the geometric center of the target region and indicate the target's position in the image. The system determines the centroid coordinates by calculating the pixel centroids of the contour region. The centroid calculation formula is a weighted average of the horizontal and vertical coordinates of all contour pixels, where the weighting values ​​can be pixel grayscale values ​​or edge intensities. The centroid coordinates are represented in the image coordinate system, typically using horizontal and vertical coordinates (X, Y).

[0138] Next, the system calculates the positional deviation between the centroid coordinates and the preset field-of-view center coordinates, generating translational motion error. The preset field-of-view center is the ideal probe alignment position defined in the system, typically located in the center of the image. The system determines the distance and direction of the target's deviation from the center by comparing the centroid coordinates with the preset field-of-view center coordinates. The positional deviation can be expressed as offsets in the horizontal and vertical directions. For example, if the centroid coordinates are located to the lower right of the field-of-view center, the deviation is in the positive X and positive Y directions.

[0139] After obtaining the positional deviation, the system converts the translational motion error into a translational speed control command for the probe based on a preset kinematic relationship. The kinematic relationship is a mapping between the probe's translational motion and translational speed, typically using a proportional control algorithm. The system calculates the translational speed command based on the positional deviation and the proportional control gain (such as a speed proportionality coefficient). The proportional control gain determines the impact of deviation changes on the probe's movement speed. The gain parameter can be adjusted according to the characteristics of the target tissue and the physical properties of the probe. For example, in soft tissues (such as abdominal tissue), a lower gain can prevent drastic probe movement; in harder tissues (such as around bone), a higher gain can be used to ensure a faster response.

[0140] The system executes translation speed control commands through the translation joints of the robotic arm. These translation joints are the actuators that control the probe's movement along the X, Y, and Z directions in three-dimensional space. Upon receiving control commands, the translation joints automatically adjust the probe's position, gradually converging the real-time centroid coordinates of the target to a preset field of view center. The system can monitor the deviation between the centroid coordinates and the field of view center in real time and continuously adjust the translation speed to ensure the probe remains aligned with the target area.

[0141] This embodiment utilizes visual servo control technology, enabling the system to automatically track the real-time position of the detected target and dynamically generate translation control commands based on the deviation between the target position and the preset field of view center. Compared to traditional manual probe alignment methods, visual servo control can monitor the target position in real time and automatically adjust the probe, avoiding imaging deviations caused by inaccurate manual operation or tissue movement. Automated translation control not only improves image stability and clarity but also reduces the workload of operators. Furthermore, the system can adapt to dynamic changes in different tissues (such as tissue displacement caused by breathing or heartbeat), always keeping the probe aligned with the target area and improving imaging results.

[0142] In one embodiment, step S40 above includes:

[0143] S401, Obtain the rotation angle command input by the remote operation device, and convert the rotation angle command into the desired rotation direction in the probe contact point coordinate system;

[0144] S402, Calculate the rotation angle error based on the deviation between the desired rotation direction and the current rotation angle of the probe;

[0145] S403, adjust the gain of the rotation angle error using a proportional control algorithm to generate a rotation angle control command;

[0146] S404, drive the ultrasound probe to rotate to the target angle according to the rotation angle control command.

[0147] In this embodiment, during the translational adjustment process, a rotation command is received via a teleoperation device, generating a rotation control command to adjust the probe's rotation angle. This process ensures that the ultrasound probe can not only achieve translational adjustment in three-dimensional space but also precisely adjust its angle, thereby keeping the probe always aligned with the target area and acquiring ultrasound images at the optimal angle. The adjustment of the rotation angle combines user input from the teleoperation device, coordinate system transformation, angle error calculation, and proportional control strategies to ensure the accuracy and stability of the control process.

[0148] First, the system receives rotation angle commands from the teleoperated device. The teleoperated device can be a joystick, steering wheel, touchscreen, or other interactive device capable of acquiring the user's rotation control intentions in real time. This rotation angle command is typically expressed in degrees or radians, representing the target angle the user wants the probe to rotate. The command input data can be an absolute angle (e.g., rotating to a specified angle) or a relative angle (e.g., adding or subtracting a certain angle from the current angle). The rotation angle input by the user through the teleoperated device is transmitted to the control system in real time and recorded as a rotation angle command.

[0149] Next, the system converts the rotation angle command input by the teleoperated device into the desired rotation direction in the probe contact point coordinate system. The probe contact point coordinate system is a three-dimensional coordinate system fixed at the probe tip, ensuring that all rotation control is based on the probe's actual attitude. The system uses a coordinate transformation matrix to convert the teleoperated device's rotation angle from the device coordinate system to the probe contact point coordinate system. This transformation process considers the spatial relationship between the teleoperated device and the probe, including the rotation axis direction and angular offset. For example, if the rotation angle input by the teleoperated device is about a horizontal axis, then in the probe coordinate system, this rotation might correspond to a rotation about the vertical axis of the probe tip coordinate system.

[0150] After coordinate system transformation, the system calculates the deviation between the desired rotation direction and the probe's current rotation angle. The probe's current rotation angle is the angle data detected in real time by the system through a position sensor or gyroscope. The desired rotation direction represents the target angle that the probe should achieve. Deviation calculation involves comparing the desired rotation angle with the current rotation angle and calculating the difference between the two. This deviation value can be positive or negative, indicating whether the probe needs to rotate clockwise or counterclockwise, respectively. Deviation calculation can be based on Euler angles (such as pitch, yaw, and roll angles) or quaternions (for three-dimensional rotation).

[0151] After obtaining the rotation angle deviation, the system adjusts the gain of the rotation angle error using a proportional control algorithm to generate a rotation angle control command. The proportional control algorithm is a closed-loop control strategy that dynamically adjusts the control output based on the error value. The system pre-sets a proportional gain parameter, representing the relationship between rotation speed and error. Specifically, the magnitude of the rotation angle control command is proportional to the rotation angle deviation; that is, a larger deviation results in a higher rotation speed, and a smaller deviation results in a lower rotation speed. The magnitude of the gain parameter can be adjusted according to the probe type and the characteristics of the target tissue. For example, when imaging sensitive tissues (such as the eye), a smaller gain can prevent the probe from rotating violently; while when imaging more stable tissues (such as the liver), a larger gain can accelerate the probe rotation speed.

[0152] Finally, the system drives the ultrasound probe to rotate to the target angle according to the rotation angle control command. The probe's rotation is achieved by the robot arm's rotary joint, which enables precise angle control. The control system sends the rotation angle control command to the robot arm's drive motor, which then drives the probe to rotate according to the command. The system monitors the probe's rotation angle in real time and compares it with the desired rotation angle to ensure that the probe eventually stabilizes at the target angle. The entire rotation control process is a closed-loop control, meaning the system detects the probe angle in real time and dynamically adjusts the rotation speed based on the deviation until the probe reaches the target angle and remains stable.

[0153] This embodiment receives rotation angle commands via a remote control device and generates rotation angle control commands based on a proportional control algorithm, enabling precise rotation control of the probe. Compared to traditional manual probe rotation, automated rotation control not only avoids angle errors caused by manual rotation but also dynamically adjusts the rotation speed to adapt to the rotation requirements of different tissues. The system automatically adjusts the rotation speed based on rotation angle deviations, ensuring the probe quickly and stably reaches the target angle. This technology not only improves imaging stability and accuracy but also reduces the operator's workload and increases imaging efficiency.

[0154] In one embodiment, step S50 above includes:

[0155] S501, in the pre-compression state and the post-compression state, respectively, acquires radio frequency signal data of the target biological tissue through the probe, and generates pre-compression volume data and post-compression volume data;

[0156] S502, Based on the motion estimation algorithm, perform three-dimensional displacement matching on the pre-compressed volume data and the post-compressed volume data to obtain tissue displacement field data;

[0157] S503, Based on the tissue displacement field data, calculate the local strain value using the least squares method to generate a two-dimensional strain map sequence;

[0158] S504, the two-dimensional strain map sequence is spatially superimposed along the probe movement direction to generate a three-dimensional strain distribution map.

[0159] In this embodiment, based on adjusted contact force, translational motion, and rotation angle, the system acquires detection signal data of the target biological tissue in pre-compression and post-compression states, and generates a three-dimensional strain distribution map based on these data. This process involves signal data acquisition, three-dimensional displacement matching, strain calculation, and three-dimensional strain map generation, and is the core technology for realizing tissue elasticity imaging.

[0160] First, in the pre-compression and post-compression states, the system acquires radio frequency (RF) signal data from the target biological tissue via the probe. RF signal data is a high-frequency signal representing the echo information of ultrasound waves propagating in the tissue and reflected back to the probe. The pre-compression state refers to the state where the probe is in contact with the tissue but no compressive force is applied, while the post-compression state refers to the state where the probe applies a certain compressive force to the tissue. This compression causes deformation within the tissue, creating different strain states. RF signal data is the foundation of ultrasound imaging, recording the propagation path, reflection intensity, and phase information of ultrasound waves in the tissue. Pre-compression volume data and post-compression volume data represent the RF signal data of the tissue in these two states, respectively. This data is acquired at high speed by the probe's array detector and stored digitally.

[0161] Next, the system performs three-dimensional displacement matching on the pre-compressed and post-compressed volume data based on a motion estimation algorithm to obtain the tissue displacement field data. The motion estimation algorithm is an image registration technique that calculates the relative displacement of tissue under pre-compressed and post-compressed states. In this technique, three-dimensional displacement matching refers to tracking the movement trajectory of each voxel before and after compression in three-dimensional space to obtain tissue displacement information. Commonly used motion estimation algorithms include optical flow and cross-correlation methods. In the optical flow method, the system estimates the axial and lateral displacement of each voxel during compression by analyzing the phase changes of radio frequency signal data. In the cross-correlation method, the system determines the displacement of each voxel by comparing the similarity between the pre-compressed and post-compressed signal data. The displacement field data is a three-dimensional vector field, where each vector represents the displacement direction and magnitude of the corresponding voxel in the tissue.

[0162] After acquiring displacement field data, the system calculates the local strain values ​​of the tissue using the least squares method, thereby generating a sequence of two-dimensional strain maps. The least squares method is a numerical fitting method that can calculate the strain magnitude based on tissue displacement data. Strain is the degree of deformation that occurs in tissue during compression, usually represented by the displacement gradient in the compression direction. In this technique, the system calculates the strain value by fitting the displacement data around each voxel. This strain value can be axial strain (strain along the compression direction), transverse strain (strain perpendicular to the compression direction), or shear strain (mixed strain in both directions). The two-dimensional strain map sequence consists of multiple planar strain maps along the probe acquisition direction, each planar map representing the tissue strain distribution at different depths.

[0163] Finally, the system spatially superimposes the two-dimensional strain map sequence along the probe's movement direction to generate a three-dimensional strain distribution map. A three-dimensional strain distribution map is a stereoscopic image that shows the strain distribution of tissue during compression. Spatial superposition refers to combining each two-dimensional strain map according to its corresponding spatial position to form a three-dimensional structure. During spatial superposition, the system ensures that each two-dimensional strain map is accurately arranged in three-dimensional space based on the probe's position and angle information. This three-dimensional strain distribution map can clearly display the strain magnitude in different regions of the tissue, especially distinguishing between rigid and elastic regions, providing a basis for tissue elasticity assessment and lesion detection.

[0164] This embodiment acquires radio frequency signal data under pre-compression and post-compression states, and calculates strain based on three-dimensional displacement matching and the least squares method. The system can accurately generate a three-dimensional strain distribution map. It not only clearly displays the strain distribution of the tissue but also accurately distinguishes between rigid and elastic regions within the tissue. Compared to traditional two-dimensional strain imaging methods, the three-dimensional strain distribution map provides three-dimensional structural information of the tissue, allowing doctors to more intuitively observe changes in tissue elasticity and the location of lesions.

[0165] In one embodiment, step S60 above includes:

[0166] S601, calculate the segmentation threshold based on the preset center value and preset percentage, wherein the segmentation threshold is the sum of the preset center value and the preset percentage multiplied by the maximum absolute value of the strain in the three-dimensional strain distribution map;

[0167] S602, the strain values ​​of each voxel in the three-dimensional strain distribution map are binarized, and the regions with strain values ​​lower than the segmentation threshold are marked as rigid regions;

[0168] S603, extract the maximum connected volume in the rigid region using a connected component analysis algorithm;

[0169] S604, calculate the centroid coordinates of the maximum connected volume, and output the centroid coordinates as the spatial location information of the rigid region.

[0170] In this embodiment, the system uses threshold segmentation to process the three-dimensional strain distribution map, extracts the rigid region, and outputs the spatial location information of the region. This process, through steps such as segmentation threshold calculation, binarization, connected component analysis, and centroid coordinate calculation, achieves precise positioning of the rigid region of the tissue.

[0171] First, the system calculates a segmentation threshold based on a preset center value and a preset percentage. The segmentation threshold is the upper limit of the strain value used to distinguish between rigid and elastic regions. The preset center value is a baseline strain value representing the strain characteristics of normal tissue. The preset percentage is a scaling factor representing the proportion of the maximum absolute strain value in the three-dimensional strain distribution map. The system obtains the segmentation threshold by multiplying the preset center value and the preset percentage by the maximum absolute strain value in the three-dimensional strain distribution map. This calculation method ensures that the threshold adapts to the strain characteristics of the tissue, effectively distinguishing rigid regions while avoiding segmentation failure due to excessively high or low thresholds. For example, when the strain in normal tissue is low, the system automatically lowers the segmentation threshold, thereby improving the resolution of rigid regions.

[0172] After determining the segmentation threshold, the system binarizes the strain values ​​of each voxel in the 3D strain distribution map. Binarization is an image segmentation technique that converts the strain values ​​in the 3D strain distribution map into binary labels based on the segmentation threshold. Specifically, when the strain value of a voxel is below the segmentation threshold, the system marks the voxel as a rigid region; otherwise, it marks it as an elastic region. Binarization ensures that the 3D strain distribution map can clearly delineate rigid regions in space, thus providing a foundation for subsequent connected component analysis.

[0173] After binarization, the system extracts the largest connected volume in the rigid regions using a connected component analysis algorithm. Connected component analysis is a spatial clustering technique that identifies and separates adjacent connected regions in a 3D image. In this technique, the system combines all adjacent rigid region voxels into connected volumes. The system iterates through all connected volumes, calculates the spatial size (i.e., volume size) of each connected volume, and selects the connected region with the largest volume as the target rigid region. This selection method ensures that the system can ignore noisy regions and isolated voxels in the image, thereby accurately locating the main rigid structures.

[0174] Finally, the system calculates the centroid coordinates of the largest connected volume and outputs these coordinates as the spatial location information of the rigid region. The centroid coordinates are the geometric center of the rigid region in three-dimensional space, representing its spatial location. The system obtains the centroid coordinates by weighted averaging of the three-dimensional coordinates of all voxels within the largest connected volume. This calculation method accurately reflects the spatial distribution of the rigid region and can output the three-dimensional coordinate information of the centroid according to actual needs, including its specific position in the X, Y, and Z axes.

[0175] For example, in automated palpation, an ultrasound probe contacts the tissue via a six-degree-of-freedom robotic arm and applies compressive force to acquire tissue strain data. The specific steps are as follows:

[0176] To achieve accurate tissue palpation, the force applied to the tissue must be controlled to ensure the stability and accuracy of the compression process. Force control uses an oscillatory force control algorithm to apply forces of varying magnitudes at different time intervals.

[0177]

[0178] τ is the time step, F max F min These represent the maximum and minimum forces applied during the execution process. The execution process is simulated by periodically varying the force to ensure that the force applied by the probe to the tissue meets expectations.

[0179] The ultrasound probe measures the force applied to the tissue in real time using a force / torque sensor (such as the ATI Gamma 65-SI). The force data measured by the force sensor needs to be transmitted from the sensor frame F. s Transition to contact point frame F pc , For F pc The conversion formula for the tensor of the measured force is as follows:

[0180]

[0181] Among them, T Fg and T Fs They are gravity frames Fg To the sensor frame F s and from framework F s To the contact point frame F pc The transition matrix H, H g Let the probe mass m p The resulting gravitational force tensor is [0, 0, 9.81m]. p [,0,0,0]. This is further improved by transferring the force to the contact point frame of the probe.

[0182] To achieve precise force control, it is necessary to calculate and adjust the force control error, e. f Based on the currently measured force characteristics s f With Expectation The difference between them consists of:

[0183]

[0184] The system processes the error through a control system feedback loop, minimizing the force error. Where λ f These are adjustable force control parameters.

[0185] During three-dimensional ultrasound imaging, the ultrasound probe must always be aimed at the target (such as a tumor or hard tissue). To ensure that the target is always centered in the ultrasound probe's field of view, a visual servo control method is used.

[0186] Using a visual servoing algorithm, the centroid position P of the target is determined. g =(x g ,y g ,z g It will continuously perform real-time calculations to keep the target within the probe's field of view. The target's real-time position s t =[x g z g ] T and desired location The error e between t It was calculated.

[0187] The speed v of the control probe can be controlled via visual servo control. x ,v z This allows the target to be automatically centered in the field of vision.

[0188] The motion control of the probe follows basic kinematic formulas, which relate to the relationship between the change of the target's center of mass and the probe's velocity:

[0189]

[0190] This formula shows that changes in the target's position directly affect the adjustment of the probe's speed. The control system calculates the error and adjusts the probe's translation speed to keep the target centered in the ultrasonic probe's field of view.

[0191] To further refine the control of 3D ultrasound imaging, the operator needs to control the rotation angle of the ultrasound probe in order to explore the target area more precisely.

[0192] The orientation of the ultrasound probe is determined by its rotation angle θ x ,θ y ,θ z Control. Define the target rotation direction s θ =[θ x θ y θ z ] T and desired rotation direction The error e between θ And by controlling the gain λ θ To adjust the error:

[0193]

[0194] By adjusting the control gain λ θ The system can precisely adjust the rotation of the probe to achieve the desired angle.

[0195] The remote-controlled device allows manual control of the ultrasonic probe's rotation direction. The operator specifies the probe's angle by adjusting the device's rotational degrees of freedom. The desired angle is given by the following formula:

[0196]

[0197] R u From the tactile device frame F v To the ultrasonic probe contact frame F pc The rotation matrix, Φ c The current rotation angle Φ of the tactile device init It is the initial angle of reference.

[0198] With the remote control device, the operator can dynamically adjust the direction of the probe according to actual needs, helping doctors to better explore the target area.

[0199] By calculating the compressive force applied by the ultrasound probe, a three-dimensional elasticity map, i.e., the strain distribution of the tissue, can be generated. By analyzing the radio frequency signals before and after compression, the strain value of the tissue can be estimated.

[0200] By measuring the pre-compressed volume V r and post-compression volume V cThe radio frequency signal data is processed to estimate the strain values ​​of the tissue under different compression states. This process typically employs motion estimation algorithms (such as optical flow methods) and least squares strain estimation, based on the radio frequency signal V under two compression states. r and V c Calculate the three-dimensional strain volume V s :

[0201] V s (i,j,k)=A sk (i,j)

[0202] Among them, A sk The strain map is a two-dimensional strain map estimated using the least squares method in the k-th frame. The resulting three-dimensional strain map can be used to identify rigid regions of the tissue.

[0203] By thresholding the 3D elasticity map, rigid tissue regions are extracted.

[0204] μ = c s +p s ·max(|V s |)

[0205] Among them, c s It is the center value of the strain value, p s The percentage used for segmentation is used to extract the largest rigid regions (e.g., tumors or hard tissues) after thresholding. This step extracts the regions with the largest strain in the volume using a connected component algorithm. And calculate the centroid of the region.

[0206] Example description: In liver ultrasound elastography, the system uses robot control technology to automatically complete the precise positioning, force control, translation and rotation adjustment, data acquisition and three-dimensional strain analysis of the ultrasound probe, thereby realizing the automatic detection and positioning of rigid areas of liver tissue.

[0207] First, the system determines the three-dimensional motion path of the ultrasound probe based on the initial location information of the liver region preset by the doctor. Based on this path, the system constructs a kinematic model of a six-DOF robotic arm and calculates the inverse kinematic parameters of each joint. The joint controller of the robotic arm drives the ultrasound probe to move along the three-dimensional motion path according to these calculated parameters, and monitors the probe's spatial pose data in real time during the movement. If the probe deviates from the preset path or exceeds a safety threshold, the system automatically adjusts the three-dimensional motion trajectory to ensure that the probe accurately reaches the target area and completes the initial positioning.

[0208] After initial positioning, the system monitors the contact force applied by the probe to the liver tissue using a force sensor. The force data acquired by the sensor is initially in the sensor coordinate system; the system converts this to the probe contact point coordinate system to ensure accuracy. The system calculates the error between the current contact force and the preset desired contact force, and automatically adjusts the contact force through a feedback control loop to ensure the actual contact force falls within and is maintained within the preset desired range. Furthermore, the system employs a control mode that periodically alternates between applying maximum and minimum contact forces to simulate the pressure changes experienced by a physician during palpation of the liver. This force control ensures stable probe contact with the liver surface, contributing to the acquisition of high-quality ultrasound images and strain data.

[0209] After the probe stably contacts the liver tissue, the system uses visual servo control technology to achieve precise positioning of the probe on a plane. The ultrasound probe acquires ultrasound image data of the liver tissue in real time. The system performs edge enhancement and noise filtering on these image data to extract the contour features of the liver tissue. The system calculates the centroid coordinates of these features as the real-time position of the liver tissue. Then, the system calculates the positional deviation between these centroid coordinates and the preset field-of-view center coordinates, and converts this deviation into translational velocity control commands. The robotic arm adjusts the translational motion of the probe according to these control commands, gradually converging the real-time centroid coordinates of the liver tissue to the center of the field of view, thereby ensuring that the probe remains aligned with the target area.

[0210] During translational adjustments, the system also allows doctors to adjust the probe's rotation angle via a teleoperation device. The doctor inputs a rotation angle command through the teleoperation device, and the system converts this command into the desired rotation direction in the probe's contact point coordinate system. The system calculates the deviation between this desired rotation direction and the probe's current rotation angle and automatically generates a rotation angle control command based on a proportional control algorithm. The robotic arm then drives the probe to rotate to the target angle according to this control command. This rotational control ensures that the probe can flexibly adapt to the curved structure of the liver, achieving the optimal ultrasound imaging angle.

[0211] After the probe's translation and rotation adjustments are completed, the system begins acquiring detection signal data from the liver tissue. Under pre-compression and post-compression conditions, the system acquires radio frequency signal data from the liver tissue via the probe, generating pre-compression and post-compression volume data. The system performs three-dimensional displacement matching on these two volumes using a motion estimation algorithm to calculate the tissue's displacement field data. Based on this displacement field data, the system calculates the local strain values ​​of the liver tissue using the least squares method, generating a sequence of two-dimensional strain maps. The system then spatially superimposes these two-dimensional strain maps along the probe's movement direction to generate a three-dimensional strain distribution map. This three-dimensional strain distribution map can visually display the strain distribution of the liver tissue under compression, especially the difference in strain distribution between rigid and elastic regions.

[0212] Finally, the system performs threshold segmentation on the 3D strain distribution map to extract rigid regions of the liver tissue. The system calculates a segmentation threshold based on a preset center value and a preset percentage, and compares the strain values ​​of each voxel in the 3D strain distribution map with this threshold. Regions with strain values ​​below the threshold are marked as rigid regions. The system extracts the maximum connected volume from the rigid regions using a connected component analysis algorithm and calculates the centroid coordinates of this region as its spatial location information. Ultimately, the system outputs these centroid coordinates to help doctors quickly locate potentially diseased rigid regions in the liver.

[0213] This embodiment utilizes segmentation threshold calculation, binarization processing, connected component analysis, and centroid coordinate calculation to accurately extract rigid regions from a 3D strain distribution map and output their spatial location information. It can adapt to the strain characteristics of different tissues and ensure the continuity and stability of the segmentation results through connected component analysis. Compared with traditional manual segmentation methods, it can automatically identify and segment rigid regions in tissues, reducing errors caused by human operation and improving segmentation accuracy and stability. Especially in tissue lesion detection, the precise location of rigid regions can effectively assist doctors in determining the nature and location of lesions, improving diagnostic accuracy.

[0214] In one embodiment, a robot-controlled region detection device is provided, which corresponds one-to-one with the robot-controlled region detection method described in the above embodiments. (Refer to...) Figure 3 , Figure 3 This is a schematic diagram of the functional modules of a preferred embodiment of the robot-controlled region detection device of the present invention. The modules include a robot motion control module 10, a force sensor control module 20, a vision servo control module 30, a rotation control module 40, a signal acquisition and strain calculation module 50, and a threshold segmentation and rigid region extraction module 60. Detailed descriptions of each functional module are as follows:

[0215] The robot motion control module 10 is used to control the three-dimensional spatial motion of the robot arm driving the probe.

[0216] The force sensor control module 20 is used for control based on the three-dimensional spatial motion, monitoring the contact force applied by the probe to the elastic medium through the force sensor, and adjusting the contact force according to the preset force control algorithm;

[0217] The visual servo control module 30 is used to track the real-time position of the target being detected through visual servo control technology during the adjustment of the contact force, and generate translation control commands based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0218] The rotation control module 40 is used to receive rotation commands through a remote operation device and generate rotation control commands to adjust the rotation angle of the probe during the adjustment of translational motion.

[0219] The signal acquisition and strain calculation module 50 is used to acquire detection signal data under pre-compression and post-compression states based on the adjusted contact force, translational motion and rotation angle, and calculate and generate a three-dimensional strain distribution map.

[0220] The threshold segmentation and rigid region extraction module 60 is used to perform threshold segmentation processing on the three-dimensional strain distribution map, extract the rigid region, and output the spatial location information of the rigid region.

[0221] In one embodiment, the robot motion control module 10 is specifically used for:

[0222] Based on the initial position information of the target area, the three-dimensional motion path of the ultrasonic probe is determined.

[0223] Based on the three-dimensional motion path, a kinematic model of a six-degree-of-freedom robot arm is established, and the inverse kinematics parameters of each joint of the six-degree-of-freedom robot arm are calculated.

[0224] Based on the inverse kinematics calculation parameters, the joint controller of the six-degree-of-freedom robot arm drives the ultrasonic probe to move along the three-dimensional motion path, generating a continuous three-dimensional motion trajectory of the ultrasonic probe.

[0225] During the movement, the spatial pose data of the ultrasonic probe is monitored, and the three-dimensional motion trajectory is adjusted according to a preset safety threshold to complete the initial positioning of the ultrasonic probe.

[0226] In one embodiment, the force sensor control module 20 is specifically used for:

[0227] Based on the control of the three-dimensional spatial motion, the contact force data of the probe applied to the biological tissue by the force sensor is transformed from the sensor coordinate system to the contact point coordinate system of the probe.

[0228] Based on the converted contact force data, calculate the force error between the current contact force and the preset expected contact force;

[0229] The force error is adjusted by a feedback control loop so that the actual contact force enters and is maintained within the preset desired contact force range.

[0230] The control mode of periodically alternating between applying maximum and minimum contact force simulates the changes in palpation pressure on biological tissues.

[0231] In one embodiment, the visual servo control module 30 is specifically used for:

[0232] Ultrasound image data of the target biological tissue is acquired using an ultrasound probe;

[0233] The ultrasound image data is subjected to edge enhancement and noise filtering to extract the contour features of the target being detected;

[0234] The real-time centroid coordinates of the detected target are calculated based on the contour features.

[0235] Calculate the positional deviation between the real-time centroid coordinates and the preset field-of-view center coordinates to generate translational motion error;

[0236] The translational motion error is converted into a translational speed control command for the probe based on a preset kinematic relationship;

[0237] The translation speed control command is executed by the translation joint of the robot arm, so that the real-time centroid coordinates of the detected target converge to the preset field of view center coordinates.

[0238] In one embodiment, the rotation control module 40 is specifically used for:

[0239] Obtain the rotation angle command input by the remote control device, and convert the rotation angle command into the desired rotation direction in the probe contact point coordinate system;

[0240] The rotation angle error is calculated based on the deviation between the desired rotation direction and the current rotation angle of the probe.

[0241] The rotation angle error is adjusted by a proportional control algorithm to generate a rotation angle control command.

[0242] The ultrasonic probe is driven to rotate to the target angle according to the rotation angle control command.

[0243] In one embodiment, the signal acquisition and strain calculation module 50 is specifically used for:

[0244] In the pre-compression and post-compression states, radio frequency signal data of the target biological tissue are acquired by the probe to generate pre-compression volume data and post-compression volume data, respectively.

[0245] Based on the motion estimation algorithm, three-dimensional displacement matching is performed on the pre-compressed volume data and the post-compressed volume data to obtain tissue displacement field data;

[0246] Based on the tissue displacement field data, the local strain values ​​are calculated using the least squares method to generate a two-dimensional strain map sequence.

[0247] The two-dimensional strain map sequence is spatially superimposed along the probe movement direction to generate a three-dimensional strain distribution map.

[0248] In one embodiment, the threshold segmentation and rigid region extraction module 60 is specifically used for:

[0249] Based on a preset center value and a preset percentage, a segmentation threshold is calculated. The segmentation threshold is the sum of the preset center value and the preset percentage multiplied by the maximum absolute value of the strain in the three-dimensional strain distribution map.

[0250] The strain values ​​of each voxel in the three-dimensional strain distribution map are binarized, and the regions with strain values ​​lower than the segmentation threshold are marked as rigid regions.

[0251] The maximum connected volume in the rigid region is extracted using a connected component analysis algorithm.

[0252] Calculate the centroid coordinates of the maximum connected volume and output the centroid coordinates as the spatial location information of the rigid region.

[0253] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external user terminals via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a robot-controlled area detection method on the server side.

[0254] In one embodiment, a computer device is provided, which may be a user terminal, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a robot-controlled area detection method on the user side.

[0255] In one embodiment, a computer device is provided, 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 perform the following steps:

[0256] Control the robot arm to drive the probe's three-dimensional spatial movement;

[0257] Based on the control of the three-dimensional spatial motion, the contact force applied by the probe to the elastic medium is monitored by a force sensor, and the contact force is adjusted according to a preset force control algorithm;

[0258] During the adjustment of the contact force, the real-time position of the target is tracked by visual servo control technology, and translation control commands are generated based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0259] During the adjustment of translational motion, rotation commands are received through remote control devices, and rotation control commands are generated to adjust the rotation angle of the probe.

[0260] Based on the adjusted contact force, translational motion, and rotation angle, detection signal data in the pre-compression and post-compression states are collected, and a three-dimensional strain distribution map is generated.

[0261] The three-dimensional strain distribution map is subjected to threshold segmentation to extract rigid regions and output the spatial location information of the rigid regions.

[0262] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0263] Control the robot arm to drive the probe's three-dimensional spatial movement;

[0264] Based on the control of the three-dimensional spatial motion, the contact force applied by the probe to the elastic medium is monitored by a force sensor, and the contact force is adjusted according to a preset force control algorithm;

[0265] During the adjustment of the contact force, the real-time position of the target is tracked by visual servo control technology, and translation control commands are generated based on the deviation between the real-time position and the preset field of view center to drive the robot arm to adjust the translational movement of the probe.

[0266] During the adjustment of translational motion, rotation commands are received through remote control devices, and rotation control commands are generated to adjust the rotation angle of the probe.

[0267] Based on the adjusted contact force, translational motion, and rotation angle, detection signal data in the pre-compression and post-compression states are collected, and a three-dimensional strain distribution map is generated.

[0268] The three-dimensional strain distribution map is subjected to threshold segmentation to extract rigid regions and output the spatial location information of the rigid regions.

[0269] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and user side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0270] Those skilled in the art will understand that all or part of the processes in the methods of 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, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0271] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0272] It should be noted that if any software tools or components not belonging to this company appear in the embodiments of this application, they are merely illustrative examples and do not represent actual use. The embodiments described above are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for detecting a region based on robot control, characterized by, The method comprises the following steps: controlling the three-dimensional spatial motion of the robot arm to drive the probe; monitoring the contact force of the probe on the elastic medium by the force sensor based on the control of the three-dimensional spatial motion, and adjusting the contact force according to a preset force control algorithm; in the process of adjusting the contact force, tracking the real-time position of the detection target by visual servo control technology, and generating a translation control instruction according to the deviation of the real-time position from the preset field center to drive the robot arm to adjust the translation motion of the probe; in the process of adjusting the translation motion, receiving a rotation instruction through the teleoperation device, generating a rotation control instruction to adjust the rotation angle of the probe; based on the adjusted contact force, translation motion and rotation angle, collecting detection signal data in the pre-compression state and the post-compression state, and calculating to generate a three-dimensional strain distribution map; threshold segmentation processing is performed on the three-dimensional strain distribution map, and the spatial position information of the rigid region is extracted and output.

2. The robot control-based area detection method according to claim 1, wherein, controlling the three-dimensional spatial motion of the robot arm to drive the probe, comprising: determining the three-dimensional motion path of the ultrasonic probe based on the initial position information of the target region according to the initial position information of the target region; based on the three-dimensional motion path, establishing a kinematics model of the six-degree-of-freedom robot arm, and calculating the inverse kinematics solution parameters of each joint of the six-degree-of-freedom robot arm; based on the inverse kinematics solution parameters, driving the ultrasonic probe to move along the three-dimensional motion path through the joint controller of the six-degree-of-freedom robot arm, and generating a continuous three-dimensional motion trajectory of the ultrasonic probe; monitoring the spatial pose data of the ultrasonic probe during the movement, and adjusting the three-dimensional motion trajectory according to a preset safety threshold to complete the initial positioning of the ultrasonic probe.

3. The robot control-based area detection method according to claim 1, wherein based on the control of the three-dimensional spatial motion, monitoring the contact force of the probe on the elastic medium by the force sensor, and adjusting the contact force according to a preset force control algorithm, comprising: based on the control of the three-dimensional spatial motion, converting the contact force data of the probe on the biological tissue collected by the force sensor from the sensor coordinate system to the contact point coordinate system of the probe; calculating the force error between the current contact force and the preset expected contact force according to the converted contact force data; adjusting the force error through a feedback control loop to make the actual contact force enter and maintain within the preset expected contact force range; based on the control mode of periodically alternating maximum contact force and minimum contact force, simulating the palpation pressure change on the biological tissue.

4. The robot control-based area detection method according to claim 1, wherein, in the process of adjusting the contact force, tracking the real-time position of the detection target by visual servo control technology, and generating a translation control instruction according to the deviation of the real-time position from the preset field center to drive the robot arm to adjust the translation motion of the probe, comprising: acquiring ultrasonic image data of the target biological tissue by the ultrasonic probe; performing edge enhancement and noise filtering processing on the ultrasonic image data to extract the contour features of the detection target; calculating the real-time centroid coordinates of the detection target based on the contour features; calculating the position deviation between the real-time centroid coordinates and the preset field center coordinates to generate a translation motion error; convert the translational motion error into a translational velocity control instruction of the probe according to a preset kinematic relationship; execute the translational velocity control instruction through a translational joint of the robot arm, so that a real-time centroid coordinate of the detection target converges to a preset field center coordinate.

5. The robot control-based area detection method according to claim 1, wherein, In the process of adjusting the translational motion, a rotation instruction is received through the teleoperation device, and a rotation control instruction is generated to adjust the rotation angle of the probe, including: obtaining a rotation angle instruction input by the teleoperation device, and converting the rotation angle instruction into an expected rotation direction in a probe contact point coordinate system; calculating a rotation angle error according to a deviation between the expected rotation direction and a current rotation angle of the probe; generating a rotation angle control instruction through gain adjustment on the rotation angle error by a proportional control algorithm; driving the ultrasonic probe to rotate to a target angle according to the rotation angle control instruction.

6. The robot control-based area detection method according to claim 1, wherein, Based on the adjusted contact force, translational motion and rotation angle, detection signal data in a pre-compression state and a post-compression state are collected, and a three-dimensional strain distribution map is calculated and generated, including: In the pre-compression state and the post-compression state, radio frequency signal data of the target biological tissue is collected through the probe to generate pre-compression volume data and post-compression volume data; based on a motion estimation algorithm, three-dimensional displacement matching is performed on the pre-compression volume data and the post-compression volume data to obtain tissue displacement field data; according to the tissue displacement field data, local strain values are calculated by a least squares method to generate a two-dimensional strain map sequence; spatially superimposing the two-dimensional strain map sequence along the probe motion direction to generate a three-dimensional strain distribution map.

7. The robot control-based area detection method according to claim 1, wherein, performing threshold segmentation processing on the three-dimensional strain distribution map to extract a rigid region and output spatial position information of the rigid region, including: calculating a segmentation threshold according to a preset center value and a preset percentage, the segmentation threshold being the sum of the preset center value and the preset percentage multiplied by the maximum strain absolute value in the three-dimensional strain distribution map; performing binaryzation processing on the strain values of each voxel in the three-dimensional strain distribution map, and marking the region with a strain value lower than the segmentation threshold as a rigid region; extracting the maximum connected volume in the rigid region through a connected component analysis algorithm; calculating the centroid coordinate of the maximum connected volume and outputting the centroid coordinate as the spatial position information of the rigid region.

8. A robot control-based area detection apparatus characterized by comprising: The region detection device based on robot control includes: a robot motion control module for controlling the three-dimensional spatial motion of the robot arm to drive the probe; a force sensor control module for monitoring the contact force applied by the probe on the elastic medium through the force sensor based on the control of the three-dimensional spatial motion, and adjusting the contact force according to a preset force control algorithm; a visual servo control module for tracking the real-time position of the detection target through visual servo control technology in the process of adjusting the contact force, and generating a translational control instruction according to the deviation between the real-time position and the preset field center to drive the robot arm to adjust the translational motion of the probe; a rotation control module for receiving a rotation instruction through the teleoperation device in the process of adjusting the translational motion, and generating a rotation control instruction to adjust the rotation angle of the probe. The signal acquisition and strain calculation module is configured to acquire detection signal data in the pre-compression state and the post-compression state based on the adjusted contact force, the translational motion and the rotation angle, and to calculate and generate a three-dimensional strain distribution map; The threshold segmentation and rigid region extraction module is configured to perform threshold segmentation on the three-dimensional strain distribution map, extract a rigid region, and output spatial position information of the rigid region.

9. A computer device, comprising: The computer device comprises a memory, a processor, and a robot control-based region detection program stored in the memory and capable of running on the processor. When the robot control-based region detection program is executed by the processor, the steps of the robot control-based region detection method according to any one of claims 1-7 are implemented.

10. A computer-readable storage medium, characterized in that, The storage medium stores a robot control-based region detection program. When the robot control-based region detection program is executed by the processor, the steps of the robot control-based region detection method according to any one of claims 1-7 are implemented.

Images