Intelligent collaborative robot system for radioexamination and control method thereof

Abstract

The invention relates to the field of robots, in particular to an intelligent collaborative robot system for radioscopy and a control method thereof, and the method comprises the steps: obtaining three-dimensional point cloud data of a target object, constructing an object surface three-dimensional model, and extracting surface contour features from the object surface three-dimensional model; forming an initial action path of the robot according to radiation inspection expected target spot distribution and surface contour features of the target object; according to regional organization characteristics, calibrating potential artifact sub-regions distributed by expected target spots of radioexamination, so as to segment the initial action path into a plurality of action sub-paths; and predicting radiation imaging interference according to the dynamic characteristics of the target object when the robot moves along the action sub-path so as to correct the driving parameters of the robot to the radiation inspection end. The action path of the robot can be planned through three-dimensional point cloud modeling, so that the robot cooperatively drives the radiographic inspection equipment to align at the target part for radiation imaging, tissue areas possibly generating artifacts in the target part are accurately avoided, and the imaging definition is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robots, and in particular to an intelligent collaborative robot system for radiological examination and a control method thereof. BACKGROUND

[0002] X-ray chest examination is a conventional radiological examination method, which can effectively and accurately image organs and tissues such as lungs. Existing radiological examination usually radiates the human body with a large area of X-rays to realize the overall radiological imaging of the corresponding organs and tissues. Considering that the implementation object of radiological examination does not necessarily have a large volume, when radiological imaging is performed on small-volume parts such as hands or feet, it is not necessary to radiate a large area of X-rays, thereby reducing unnecessary ionizing radiation and reducing the radiation damage to the human body.

[0003] Small-volume radiological examination devices such as X-ray cameras are widely used in small-range radiological imaging scenarios. In order to ensure that the X-ray camera accurately aims at the target part for imaging during the examination, the robot is used to drive the X-ray camera to run so as to aim at the target part for X-ray radiation. In actual radiological examination, the tissue characteristics such as fat content of the target part may cause problems such as image blurring and unclear boundaries. If the X-ray camera aims at the abnormal range of tissue characteristics during the whole movement, not only clear images cannot be obtained, but also the risk of X-ray exposure is increased. Therefore, how to use the robot to cooperatively drive the radiological examination device to aim at the target part for radiological imaging and accurately avoid the tissue region in the target part that may cause artifacts is of great significance for improving the imaging clarity of radiological examination and reducing the risk of X-ray exposure. SUMMARY

[0004] Considering that the fat tissue inside the human body may interfere with the radiological examination imaging, resulting in problems such as image blurring and unclear boundaries, and the positioning jitter of the target part during the radiological examination imaging also affects the imaging clarity, reduces the radiological examination imaging quality, and increases the risk of X-ray exposure. In order to make the robot cooperatively drive the radiological examination device to aim at the target part for radiological imaging and accurately avoid the tissue region in the target part that may cause artifacts, the present application provides an intelligent collaborative robot system for radiological examination, which comprises the following modules:

[0005] A three-dimensional model construction module is configured to acquire three-dimensional point cloud data of a target object, and construct a three-dimensional model of an object surface according to the three-dimensional point cloud data.

[0006] A surface contour extraction module is configured to extract surface contour features from the three-dimensional model of the object surface.

[0007] An action path forming module is configured to form an initial action path of a robot according to a radiological examination expected target point distribution of the target object and the surface contour features.

[0008] an artifact calibration module configured to acquire region tissue features of the radiation examination desired target point distribution of the target object, and calibrate potential artifact sub-regions of the radiation examination desired target point distribution of the target object;

[0009] a path segmentation module configured to segment the initial motion path into a plurality of motion sub-paths according to the potential artifact sub-regions;

[0010] a robot driving correction module configured to acquire dynamic features of the target object during motion of the robot along the motion sub-paths, predict radiation imaging interference according to the dynamic features, and correct driving parameters of the radiation examination end of the robot according to the radiation imaging interference.

[0011] Preferably, the three-dimensional model construction module is configured to acquire three-dimensional point cloud data of the target object, and construct a three-dimensional model of an object surface according to the three-dimensional point cloud data, specifically:

[0012] acquire original point cloud data formed by shooting the target object by a depth camera, perform all-out rejection of out-of-range point cloud data and partial rejection of concentrated point cloud data on the original point cloud data with respect to a preset limited range, to obtain three-dimensional point cloud data, and perform continuous fitting on the three-dimensional point cloud data to construct a three-dimensional model of an object surface;

[0013] The surface contour extraction module is configured to extract surface contour features from the three-dimensional model of the object surface, specifically:

[0014] perform surface normal vector and surface tangent vector calibration and correction on the three-dimensional model of the object surface to determine a surface contour point cluster of the three-dimensional model of the object surface, and determine surface contour features of the target object according to the surface contour point cluster; wherein the surface contour features include surface contour topography and curvature distribution features.

[0015] Preferably, the motion path formation module is configured to form an initial motion path of a robot according to the radiation examination desired target point distribution of the target object and the surface contour features, specifically:

[0016] acquire a target point cluster spatial position layout of the radiation examination desired target point distribution of the target object, perform target point cluster spatial sorting and filling processing according to the target point cluster spatial position layout, and obtain a corrected target point cluster spatial position layout;

[0017] perform target point cluster relative surface contour mapping processing according to the corrected target point cluster spatial position layout and the surface contour features, to obtain a target point continuous layout line, and convert the target point continuous layout line into the initial motion path of the robot according to a relative spatial relationship between the robot and the target object; wherein the initial motion path refers to a path along which the robot moves the radiation examination end.

[0018] Preferably, the artifact calibration module is configured to acquire regional tissue features of the radiation examination desired target point distribution, and calibrate potential artifact sub-regions of the radiation examination desired target point distribution based on the regional tissue features.

[0019] The artifact calibration module is configured to acquire regional ultrasound detection images of the radiation examination desired target point distribution, identify and extract fat tissue thickness distribution features from the regional ultrasound detection images, estimate imaging contrast of each sub-region where a radiation examination desired target point is located based on the fat tissue thickness distribution features, and calibrate potential artifact sub-regions of the radiation examination desired target point distribution based on the imaging contrast.

[0020] The path segmentation module is configured to segment the initial motion path into a plurality of motion sub-paths based on the potential artifact sub-regions.

[0021] The path segmentation module is configured to segment the initial motion path into a plurality of motion sub-paths based on projections of boundaries of the potential artifact sub-regions in a motion space of the robot, wherein each motion sub-path does not overlap with the projection of the boundary of the potential artifact sub-region.

[0022] Preferably, the robot driving correction module is configured to acquire dynamic features of the target object during movement of the robot along the motion sub-paths, predict radiation imaging interference based on the dynamic features, and correct driving parameters of the robot for a radiation examination end based on the radiation imaging interference.

[0023] The robot driving correction module is configured to acquire dynamic features of positioning of the target object during movement of the robot along the motion sub-paths, predict an imaging blur range of the radiation examination end during movement of the robot along the motion sub-paths based on the dynamic features, and correct spatial driving orientation parameters of a robot arm of the robot for the radiation examination end based on the imaging blur range.

[0024] The robot driving correction module is configured to acquire dynamic features of positioning of the target object during movement of the robot along the motion sub-paths, predict an imaging blur range of the radiation examination end during movement of the robot along the motion sub-paths based on the dynamic features, and correct spatial driving orientation parameters of a robot arm of the robot for the radiation examination end based on the imaging blur range.

[0025] In another aspect, the present application provides a control method of an intelligent collaborative robot system for radiation examination, the control method comprising the following steps:

[0026] S100: acquiring three-dimensional point cloud data of a target object, constructing an object surface three-dimensional model based on the three-dimensional point cloud data, and extracting surface contour features from the object surface three-dimensional model; forming an initial motion path of a robot based on a radiation examination desired target point distribution of the target object and the surface contour features.

[0027] S200: acquiring regional tissue features of the radiation examination desired target point distribution, and calibrating potential artifact sub-regions of the radiation examination desired target point distribution based on the regional tissue features; segmenting the initial motion path into a plurality of motion sub-paths based on the potential artifact sub-regions.

[0028] S300: acquiring a dynamic feature of the target object during movement of the robot along the action sub-path, to predict a radiological imaging interference; and correcting a driving parameter of the radiological examination end of the robot according to the radiological imaging interference.

[0029] Preferably, in S100, three-dimensional point cloud data of the target object is acquired, a three-dimensional model of the object surface is constructed according to the three-dimensional point cloud data, and a surface contour feature is extracted from the three-dimensional model of the object surface, specifically:

[0030] Raw point cloud data formed by the depth camera shooting the target object is acquired, all out-of-bounds point cloud data and part of the concentrated point cloud data about the preset limited range are removed, and three-dimensional point cloud data is obtained; the three-dimensional point cloud data is continuously fitted to construct a three-dimensional model of the object surface;

[0031] The surface normal vector and the surface tangent vector of the three-dimensional model of the object surface are calibrated and corrected to determine a surface contour point cluster of the three-dimensional model of the object surface; the surface contour feature of the target object is determined according to the surface contour point cluster; wherein the surface contour feature includes surface contour topography and curvature distribution feature.

[0032] Preferably, in S100, an initial action path of the robot is formed according to the radiological examination expected target point distribution of the target object and the surface contour feature, specifically:

[0033] The spatial position layout of the target point cluster of the radiological examination expected target point distribution of the target object is acquired, the spatial sorting and filling processing of the target point cluster is performed according to the spatial position layout of the target point cluster, and the corrected spatial position layout of the target point cluster is obtained;

[0034] The target point cluster relative surface contour mapping processing is performed according to the corrected spatial position layout of the target point cluster and the surface contour feature, and a target point continuous layout line is obtained; the target point continuous layout line is converted into the initial action path of the robot according to the relative spatial relationship between the robot and the target object; wherein the initial working path refers to the path of the robot moving the radiological examination end.

[0035] Preferably, in S200, the regional organization feature of the radiological examination expected target point distribution is acquired to calibrate a potential artifact sub-region of the radiological examination expected target point distribution; the initial action path is divided into a plurality of action sub-paths according to the potential artifact sub-region, specifically:

[0036] acquire an ultrasound image of a region where the target point is expected to be located in the radiation detection, and identify a fat tissue thickness distribution feature from the ultrasound image; estimate an imaging contrast of a sub-region where each target point is expected to be located in the radiation detection according to the fat tissue thickness distribution feature, so as to mark a potential artifact sub-region of the target point distribution in the radiation detection;

[0037] divide the initial motion path into a plurality of motion sub-paths according to a projection of a boundary of the potential artifact sub-region on a motion space of the robot, wherein each motion sub-path does not overlap with the projection of the boundary of the potential artifact sub-region.

[0038] Preferably, in S300, a dynamic feature of the target object during motion of the robot along the motion sub-path is acquired, so as to predict a radiation imaging interference; and a driving parameter of the robot to the radiation detection end is corrected according to the radiation imaging interference, in particular:

[0039] a positioning dynamic change feature of the target object during motion of the robot along the motion sub-path is acquired, so as to predict an imaging blur range of the radiation detection end during motion of the robot along the motion sub-path;

[0040] a spatial driving orientation parameter of the robot to the radiation detection end is corrected according to the imaging blur range.

[0041] Compared with the prior art, the present application has the following beneficial effects:

[0042] The intelligent collaborative robot system for radiation detection and the control method thereof acquire three-dimensional point cloud data of a target object, construct a three-dimensional model of an object surface and extract a surface contour feature therefrom; an initial motion path of the robot is formed according to a target point distribution of the target object in radiation detection and the surface contour feature; a potential artifact sub-region of the target point distribution in radiation detection is marked according to a regional tissue feature, so as to divide the initial motion path into a plurality of motion sub-paths; and a radiation imaging interference is predicted according to a dynamic feature of the target object during motion of the robot along the motion sub-path, so as to correct a driving parameter of the robot to the radiation detection end. The three-dimensional point cloud modeling is used to plan a motion path of the robot, so that the robot cooperates with the radiation detection device to align a target part for radiation imaging, accurately avoids a tissue region that may cause artifacts in the target part, and improves imaging clarity and reduces radiation exposure risk. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor. Among them:

[0044] Figure 1 is a structural diagram of an intelligent collaborative robot system for radiological examination provided by the present application.

[0045] Figure 2 is the process of obtaining raw point cloud data by shooting of a depth camera.

[0046] Figure 3 is three-dimensional point cloud data of a target object.

[0047] Figure 4 is the correction process of the spatial position layout of the target point cluster.

[0048] Figure 5 is the initial motion path of the robot.

[0049] Figure 6 is the ultrasound detection image of the fat tissue in region A of the target object.

[0050] Figure 7 is the ultrasound detection image of the fat tissue in region B of the target object.

[0051] Figure 8 is the ultrasound detection image of the fat tissue in region C of the target object.

[0052] Figure 9 is the correction curve of the position of the mechanical arm of the robot along the X axis.

[0053] Figure 10 is the correction curve of the pitch angle of the mechanical arm of the robot.

[0054] Figure 11 is a flowchart of a control method of an intelligent collaborative robot system for radiological examination provided by the present application. DETAILED DESCRIPTION

[0055] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described in detail below with reference to the drawings. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that only the parts related to the present application are shown in the drawings for the convenience of description, not all the structures. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0056] The terms "comprising" and "having" and any variations thereof in the present application are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0057] In this paper, the "embodiment" means that the specific features, structures or characteristics described in conjunction with the embodiment can be included in at least one embodiment of the present application. The phrase appears at various places in the specification does not necessarily refer to the same embodiment, nor is it independent or alternative to other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0058] Please refer to Figure 1 As shown in the drawings, the present application provides an intelligent collaborative robot system for radiological examination, which comprises the following modules:

[0059] A three-dimensional model construction module is configured to acquire three-dimensional point cloud data of a target object, and construct a three-dimensional model of an object surface according to the three-dimensional point cloud data;

[0060] A surface contour extraction module is configured to extract surface contour features from the three-dimensional model of the object surface;

[0061] An action path forming module is configured to form an initial action path of the robot according to a radiological examination desired target point distribution of the target object and the surface contour features;

[0062] An artifact calibration module is configured to acquire regional tissue features of the radiological examination desired target point distribution, and calibrate potential artifact sub-regions of the radiological examination desired target point distribution according to the regional tissue features;

[0063] A path segmentation module is configured to segment the initial action path into a plurality of action sub-paths according to the potential artifact sub-regions;

[0064] The robot driving correction module is configured to acquire dynamic characteristics of the target object during movement of the robot along the action sub-path, to predict a radiological imaging interference, and to correct driving parameters of the robot to the radiological examination end according to the radiological imaging interference.

[0065] Further, the three-dimensional model construction module is configured to acquire three-dimensional point cloud data of the target object, and to construct a three-dimensional model of the object surface according to the three-dimensional point cloud data, specifically as follows.

[0066] The three-dimensional point cloud data is obtained by acquiring original point cloud data formed by the depth camera shooting the target object, and by performing all-out rejection of out-of-range point cloud data and partial rejection of concentrated point cloud data on the original point cloud data with respect to a preset limited range.

[0067] The surface contour extraction module is configured to extract surface contour features from the three-dimensional model of the object surface, specifically as follows.

[0068] The surface contour features include surface contour topography and curvature distribution characteristics.

[0069] In actual radiological examination, small body parts such as hands and feet of a human body can be target objects of X-ray radiological imaging. Considering the complex physiological structure of the hands and feet, which have multiple fingers and multiple toes, and the fact that radiological examination of the above-mentioned body parts is performed using a small-volume X-ray camera as a radiological examination end, the radiological examination end moves in a scanning manner to perform local radiological imaging on different regions of the above-mentioned body parts one by one, so as to realize complete global radiological imaging of the above-mentioned body parts. In order to ensure that the radiological examination end accurately aims at the X-ray imaging of different regions of the target object, a plurality of radiological examination expected target points are preset on the target object, so that the radiological examination expected target points are distributed accordingly. Each radiological examination expected target point can be, but is not limited to, a plurality of marker points at each joint of the target object. As can be seen from the above description, the complex shape of the target object makes the distribution of all radiological examination expected target points on the target object irregular and scattered. In order to ensure that the radiological examination end aims at all radiological examination expected target points for radiological imaging to the maximum extent, the radiological examination end needs to set a radiological imaging scanning path for the target object in combination with the three-dimensional shape of the surface of the target object and the distribution of the radiological examination expected target points.

[0070] Please refer to Figure 2, in order to accurately determine the surface three-dimensional shape of the target object, a depth camera such as a binocular camera can be used to capture the target object to obtain a depth image of the target object. Specifically, the left camera and the right camera are used to capture the target object on the left camera optical axis and the right camera optical axis, respectively, to form images on the corresponding left image plane and the right image plane. The depth image is obtained by performing disparity on the left camera image and the right camera image formed by the left image plane and the right image plane. The depth features of all pixel points are extracted from the above depth image to obtain the depth information of all pixel points in the depth camera shooting space coordinate system. The above depth information is used to identify and convert all pixel points to obtain the original point cloud data of the target object in the depth camera shooting space coordinate system. Considering that the depth camera may be disturbed by factors such as shaking during shooting, resulting in abnormal point cloud data in the original point cloud data. These abnormal point cloud data include point cloud data outside the target object's attention range and noise superimposed point cloud data generated by shooting shaking. These abnormal point cloud data will affect the accuracy and reliability of the original point cloud data. Therefore, the original point cloud data is processed to remove all out-of-range point cloud data within a predetermined range and partially remove concentrated point cloud data. That is, all point cloud data outside the predetermined range is removed, and point cloud data with an actual distance less than a predetermined distance threshold is partially removed to avoid excessive point cloud data superimposition affecting the distinguishability of the point cloud data. The original point cloud data processed above is transformed from the depth camera shooting space coordinate system to the motion space coordinate system of the robot to obtain three-dimensional point cloud data, as shown in Figure 3 , wherein Figure 3 the three-dimensional point cloud data corresponding to the target object being a hand. The above three-dimensional point cloud data comprehensively and accurately reflects the three-dimensional point cloud density distribution of the hand. Then, a feature learning network such as a PointNet++ model is used to continuously fit the above three-dimensional point cloud data to construct an object surface three-dimensional model. Here, the object surface three-dimensional model accurately represents the surface three-dimensional shape of the target object in a wide range, providing a basis for determining the surface contour of the target object.

[0071] The surface three-dimensional contour of the target object determines the actual spatial layout of the desired target point of the radiation examination, which in turn affects the radiation examination path of the robot-driven radiation examination end to the target object. Therefore, the surface normal vector and the surface tangent vector of all surface feature points on the calibrated object surface three-dimensional model are first obtained, and the direction vector deviation correction is performed on the surface normal vector and the surface tangent vector of all surface feature points to obtain the normal vector set and the tangent vector set corresponding to all surface feature points. Then, the point space position identification and continuous fitting are performed on the above normal vector set and tangent vector set to obtain the surface contour point cluster of the calibrated object surface three-dimensional model, so as to determine the surface contour topography and curvature distribution characteristics of the target object, and provide a basis for determining the initial motion path of the robot.

[0072] Further, the action path forming module is configured to form an initial action path of the robot according to the radiation examination target point distribution of the target object and the surface contour feature, specifically:

[0073] obtain a target point cluster spatial position layout of the radiation examination target point distribution of the target object, perform target point cluster spatial sorting and filling processing according to the target point cluster spatial position layout, and obtain a corrected target point cluster spatial position layout;

[0074] perform target point cluster relative surface contour mapping processing according to the corrected target point cluster spatial position layout and the surface contour feature, and obtain a target point continuous layout line; and convert the target point continuous layout line into the initial action path of the robot according to a relative spatial relationship between the robot and the target object; wherein the initial action path refers to a path of movement of the radiation examination end driven by the robot.

[0075] The radiation examination target point distribution refers to a plurality of radiation examination target point distributions pre-set on the target object, which represent a plurality of reference points for implementing radiation examination on the target object. The radiation examination target point distribution is usually set according to a pre-set spatial uniform distribution mode (such as a pre-set spatial equidistance distribution mode), so that the radiation examination target point distribution cannot comprehensively and effectively represent the radiation examination demand on the target object relative to the target object itself, resulting in possible disorder of the target point examination sequence in a local region and / or the target point distribution in a local region not accurately covering the actual detection positions required. In order to ensure accurate radiation examination on the target object in the whole range, it is necessary to correct the target point spatial distribution. Please refer to Figure 4 , first obtain a target point cluster spatial position layout of the radiation examination target point distribution of the target object, then determine the actual detection positions required which are not accurately covered by the target points according to the above target point cluster spatial position layout, perform spatial sorting processing and filling processing on all original target points, thereby increasing a plurality of filling target points and sorting all original target points and all filling target points, to obtain a corrected target point cluster spatial position layout, and also set a first order target point and a target point tracking sequence in all corrected target points composed of all original target points and all filling target points (see Figure 4 for details), to provide accurate discrete point reference for subsequent determination of the initial action path of the robot.

[0076] Please refer to Figure 5 , according to the corrected target point cluster spatial position layout and the surface contour feature, map the corrected target point cluster spatial position layout on the three-dimensional surface of the target object as a whole, and combine discrete point continuous smoothing connection processing to obtain a target point continuous layout line. Then, according to the relative spatial attitude relationship between the robot and the target object, perform spatial attitude angle conversion processing on the target point continuous layout line to obtain, for example, Figure 5The initial motion path of the robot shown can be understood as the robot being able to drive the radiological examination end to move along the initial motion path, so that the target object is imaged by radiating X-rays during the movement of the radiological examination end, providing continuous spatial path navigation for the movement of the radiological examination end by the robot and avoiding the target object being excessively exposed to radiation during the movement of the radiological examination end.

[0077] Further, the artifact calibration module is configured to obtain regional tissue features of the radiological examination target point distribution, and calibrate potential artifact sub-regions of the radiological examination target point distribution based on the regional tissue features, specifically:

[0078] Obtain regional ultrasound detection images of the radiological examination target point distribution, and identify and extract fat tissue thickness distribution features from the regional ultrasound detection images; estimate the imaging contrast of each sub-region where the radiological examination target point is located based on the fat tissue thickness distribution features, and calibrate potential artifact sub-regions of the radiological examination target point distribution based on the regional tissue features;

[0079] The path segmentation module is configured to segment the initial motion path into a plurality of motion sub-paths based on the potential artifact sub-regions, specifically:

[0080] Segment the initial motion path into a plurality of motion sub-paths based on the projection of the boundary of the potential artifact sub-region in the motion space of the robot; wherein each motion sub-path does not overlap with the projection of the boundary of the potential artifact sub-region.

[0081] X-rays have high penetration and low divergence. Fat tissue in the target object has a lower density and a larger contrast difference with the surrounding tissue, resulting in image blur and unclear boundary artifact conditions when the radiological examination end radiates X-rays to image the target object. The range and blur degree of the above-mentioned artifacts are related to the condition of the fat tissue itself; specifically, the greater the thickness of the fat tissue, the greater the range and blur degree of the artifact. In order to suppress the imaging artifact conditions of the radiological examination end, it is necessary to avoid the fat tissue and its adjacent regions in the target object. Specifically, regional ultrasound detection images of the radiological examination target point distribution are obtained, and fat tissue thickness distribution features are identified and extracted from the regional ultrasound detection images. Please refer to Figure 6 、 Figure 7 、 Figure 8The ultrasound detection images of the three different regions A, B and C of the target object are respectively analyzed and recognized to determine the fat tissue thickness h of the three regions. The fat tissue thickness h of different regions is different, which leads to different imaging contrast differences between the fat tissue and the surrounding tissue in the images obtained by the radiation examination end imaging of different regions. Specifically, the fat tissue thickness h of region A is the largest, the imaging contrast difference between the fat tissue and the surrounding tissue in the image of region A is smaller, and problems such as blurring and unclear boundary are prone to occur. The above region A should be considered as a potential artifact sub-region; the fat tissue thickness of region C is the smallest, the imaging contrast difference between the fat tissue and the surrounding tissue in the image of region C is larger, and problems such as blurring and unclear boundary are not prone to occur. The above region C should be considered as not belonging to the potential artifact sub-region. By detecting and analyzing the fat tissue thickness of different regions corresponding to the expected target point distribution of the radiation detection period, the imaging contrast of each radiation examination expected target point sub-region is estimated, which helps to accurately determine the potential artifact sub-region, provides a space calibration reference for subsequent planning of the action sub-path of the robot to avoid the corresponding artifact sub-region. The projection of the boundary of the potential artifact sub-region in the action space of the robot is also obtained. The path part of the above initial action path and the projection of the artifact sub-region are removed, so as to divide the initial action path into several action sub-paths, so that each action sub-path does not overlap with the projection of the boundary of the potential artifact sub-region. In this way, when the robot moves along each action sub-path to drive the radiation examination end, the radiation examination end will not image the potential artifact sub-region in the target object, effectively reducing the probability of blurring and unclear boundary of the radiation examination imaging, and improving the quality of the radiation examination imaging.

[0082] Further, the robot driving correction module is used to obtain the dynamic characteristics of the target object during the movement of the robot along the action sub-path, so as to predict the radiation imaging interference; according to the radiation imaging interference, the driving parameters of the robot to the radiation examination end are corrected, specifically:

[0083] The positioning dynamic change characteristics of the target object during the movement of the robot driving the radiation examination end along the action sub-path are obtained, so as to predict the imaging blur range of the radiation examination end during the movement of the action sub-path;

[0084] According to the imaging blur range, the spatial driving direction parameters of the robot to the radiation examination end are corrected.

[0085] In an ideal case, when the robot drives the radiological examination end to perform radiological examination imaging on the target object, the target object should be stationary. If the target object itself has a shaking position, the radiological examination imaging of the target object will be blurred, and the imaging blur range is related to the shaking condition of the target object itself. The greater the shaking amplitude and the higher the shaking frequency, the larger the imaging blur range. In order to adaptively adjust the shaking condition of the target object during the synchronous imaging process of the radiological examination end, it is necessary to pre-acquire and quantify the shaking condition. Specifically, the dynamic change characteristics of the target object, such as the shaking displacement amplitude and the angle dynamic change of the target object during the movement of the robot driving the radiological examination end along the action sub-path, are acquired to predict the imaging blur range of the radiological examination end during the movement along the action sub-path. If the area value corresponding to the imaging blur range exceeds the preset area threshold, the spatial driving displacement and the attitude angle of the robot arm to the radiological examination end are corrected. Please refer to Figure 9 and Figure 10 correspond to the dynamic correction of the X-axis position and the pitch angle of the robot arm, respectively. By correcting the X-axis position and the pitch angle of the robot arm at a certain time interval during the operation of the robot, the imaging blur caused by the shaking of the target object itself can be effectively offset, thereby improving the clarity and reliability of the radiological examination imaging of the target object.

[0086] In addition, the robot further has an image shooting unit, a positioning guiding unit, a voice guiding unit, and a central control unit. During the implementation of the radiological examination on the target object, the image shooting unit shoots the target object from both eyes to obtain a binocular image; the central control unit performs binocular parallax calculation on the binocular image to obtain an actual three-dimensional image of the target object; the central control unit further compares the actual three-dimensional image of the target object with an expected positioning three-dimensional image of the target object to determine a positioning deviation between the actual positioning and the expected positioning of the target object, wherein the positioning deviation includes a displacement deviation and an attitude angle deviation between the actual positioning and the expected positioning in a three-dimensional space. The positioning guiding unit can generate a positioning deviation guiding image according to the positioning deviation, wherein the positioning deviation guiding image contains at least one of translation value information of the target object in the XYZ three axes in the three-dimensional space and at least one of angle change value information of the target object in the pitch angle, the yaw angle, and the roll angle in the three-dimensional space. In this way, the patient can learn the translation value information and the angle change value information from the positioning deviation guiding image, so as to adaptively change the positioning of the target object and ensure that the target object is accurately adjusted to the expected positioning. The voice guiding unit can further generate corresponding three-dimensional space translation voice guiding signals and three-dimensional space attitude angle change guiding signals according to the translation value information and the angle change value information, so that the patient can adaptively change the positioning of the target object after hearing the three-dimensional space translation voice guiding signals and the three-dimensional space attitude angle change voice guiding signals. Therefore, the robot can guide the positioning of the patient in the radiological examination process from two aspects of image and voice, so as to ensure accurate aiming of the target object in the radiological examination.

[0087] Referring to Figure 11 As shown in the drawings, the present application provides a control method of an intelligent collaborative robot system for radiological examination, which comprises the following steps:

[0088] S100: obtaining three-dimensional point cloud data of a target object, constructing an object surface three-dimensional model according to the three-dimensional point cloud data, and extracting surface contour features from the object surface three-dimensional model; forming an initial action path of a robot according to an expected target point distribution of the target object in a radiological examination and the surface contour features.

[0089] Further, in S100, the three-dimensional point cloud data of the target object is obtained, the object surface three-dimensional model is constructed according to the three-dimensional point cloud data, and the surface contour features are extracted from the object surface three-dimensional model, specifically as follows:

[0090] obtaining raw point cloud data formed by shooting the target object by a depth camera, performing all-out rejection of out-of-bounds point cloud data and partial rejection of concentrated point cloud data on the raw point cloud data with respect to a preset limited range to obtain the three-dimensional point cloud data; performing continuous fitting on the three-dimensional point cloud data to construct the object surface three-dimensional model;

[0091] The surface normal vector and the surface tangent vector of the object surface three-dimensional model are calibrated and corrected, and the surface contour point cluster of the object surface three-dimensional model is determined; the surface contour feature of the target object is determined according to the surface contour point cluster; wherein the surface contour feature includes the surface contour topography and the curvature distribution feature.

[0092] Further, in S100, according to the target object's radiation examination expected target point distribution and the surface contour feature, the initial action path of the robot is formed, specifically:

[0093] The target point cluster spatial position layout of the target object's radiation examination expected target point distribution is obtained, and the target point cluster spatial sorting and filling processing is performed according to the target point cluster spatial position layout, to obtain the corrected target point cluster spatial position layout;

[0094] The target point cluster relative surface contour mapping processing is performed according to the corrected target point cluster spatial position layout and the surface contour feature, to obtain the target point continuous layout line; the target point continuous layout line is converted into the initial action path of the robot according to the relative spatial relationship between the robot and the target object; wherein the initial action path refers to the path of the radiation examination end moved by the robot.

[0095] S200: Obtain the regional organization feature of the radiation examination expected target point distribution, and mark the potential artifact sub-area of the radiation examination expected target point distribution according to the regional organization feature; according to the potential artifact sub-area, the initial action path is divided into a plurality of action sub-paths.

[0096] Further, in S200, the regional organization feature of the radiation examination expected target point distribution is obtained, and the potential artifact sub-area of the radiation examination expected target point distribution is marked according to the regional organization feature; according to the potential artifact sub-area, the initial action path is divided into a plurality of action sub-paths, specifically:

[0097] The regional ultrasound detection image of the radiation detection expected target point distribution is obtained, and the fat tissue thickness distribution feature is identified and extracted from the regional ultrasound detection image; according to the fat tissue thickness distribution feature, the imaging contrast of each sub-area where the radiation examination expected target point is located is estimated, so as to mark the potential artifact sub-area of the radiation examination expected target point distribution;

[0098] According to the projection of the boundary of the potential artifact sub-area in the action space of the robot, the initial action path is divided into a plurality of action sub-paths; wherein each action sub-path does not overlap with the projection of the boundary of the potential artifact sub-area.

[0099] S300: Obtain the dynamic feature of the target object during the movement of the robot along the action sub-path, and predict the radiation imaging interference according to the dynamic feature; according to the radiation imaging interference, the driving parameter of the robot to the radiation examination end is corrected.

[0100] Further, in S300, the dynamic characteristics of the target object during movement of the robot along the action sub-path are acquired to predict the radiological imaging interference; and according to the radiological imaging interference, the driving parameter of the robot to the radiological examination end is corrected, specifically:

[0101] The positioning dynamic change characteristics of the target object during movement of the robot with the radiological examination end along the action sub-path are acquired to predict the imaging blur range of the radiological examination end during movement along the action sub-path;

[0102] According to the imaging blur range, the spatial driving direction parameter of the robot to the radiological examination end is corrected.

[0103] The control method of the intelligent collaborative robot system for radiological examination of the present application is consistent with the operation and effect of the intelligent collaborative robot system for radiological examination described above, and the control method of the intelligent collaborative robot system for radiological examination will not be repeated here.

[0104] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of a general hardware platform as necessary, and of course can also be implemented by means of a combination of hardware and software. Based on such understanding, the above technical solutions can be embodied in the form of a computer program product, and the present application can be implemented in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0105] Finally, it should be noted that: the above examples are used to illustrate the technical solutions of the present application, and are not limited thereto, and other embodiments can also be used; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent collaborative robot system for radiological examinations, characterized in that, The system includes the following modules: The 3D model construction module is used to acquire the 3D point cloud data of the target object and construct a 3D model of the object surface based on the 3D point cloud data. A surface contour extraction module is used to extract surface contour features from the three-dimensional model of the object surface; The motion path formation module is used to form the robot's initial motion path based on the expected target point distribution for radiographic inspection of the target object and the surface contour features; The artifact calibration module is used to obtain the regional tissue characteristics of the expected target distribution in the radiological examination, thereby calibrating the potential artifact region of the expected target distribution in the radiological examination. The path segmentation module is used to segment the initial action path into several action sub-paths based on the potential pseudo-shadow region. The robot drive correction module is used to acquire the dynamic characteristics of the target object during the robot's movement along the motion sub-path, thereby predicting radiographic interference; and to correct the robot's drive parameters for the radiographic examination end based on the radiographic interference.

2. The intelligent collaborative robot system according to claim 1, characterized in that, The 3D model construction module is used to acquire 3D point cloud data of the target object, and construct a 3D model of the object surface based on the 3D point cloud data, specifically as follows: The original point cloud data formed by the depth camera capturing the target object is obtained. The original point cloud data is then subjected to a process of removing all out-of-bounds point cloud data within a preset limit and partially removing concentrated point cloud data to obtain three-dimensional point cloud data. The three-dimensional point cloud data is continuously fitted to construct a three-dimensional model of the object surface; The surface contour extraction module is used to extract surface contour features from the three-dimensional model of the object surface, specifically: The surface normal vector and surface tangent vector of the object surface three-dimensional model are calibrated and corrected to determine the surface contour point cluster of the object surface three-dimensional model; the surface contour features of the target object are determined based on the surface contour point cluster; wherein, the surface contour features include surface contour morphology and curvature distribution features.

3. The intelligent collaborative robot system according to claim 2, characterized in that, The motion path formation module is used to form the robot's initial motion path based on the expected target point distribution for radiological examination of the target object and the surface contour features, specifically: The spatial layout of the target clusters is obtained as the expected distribution of the target points for radiological examination of the target object. Based on the spatial layout of the target clusters, the target clusters are sorted and filled to obtain the corrected spatial layout of the target clusters. Based on the spatial layout of the modified target cluster and the surface contour features, the target cluster is mapped relative to the surface contour to obtain a continuous target layout line; based on the relative spatial relationship between the robot and the target object, the continuous target layout line is converted into the robot's initial motion path; wherein, the initial working path refers to the path by which the robot drives the radiological examination end to move.

4. The intelligent collaborative robot system according to claim 1, characterized in that, The artifact calibration module is used to obtain the regional tissue characteristics of the expected target distribution in the radiological examination, thereby calibrating the potential artifact regions of the expected target distribution in the radiological examination. Specifically: Acquire ultrasound images of the region where the expected target point for the radiological examination is distributed, and extract the adipose tissue thickness distribution features from the ultrasound images of the region; based on the adipose tissue thickness distribution features, estimate the imaging contrast of the sub-region where each expected target point for the radiological examination is located, thereby identifying the potential artifact regions of the expected target point distribution for the radiological examination. The path segmentation module is used to segment the initial action path into several action sub-paths based on the potential artifact regions, specifically: Based on the projection of the boundary of the potential artifact region onto the robot's motion space, the initial motion path is divided into several motion sub-paths; wherein the projection of each motion sub-path onto the boundary of the potential artifact region does not overlap.

5. The intelligent collaborative robot system according to claim 1, characterized in that, The robot drive correction module is used to acquire the dynamic characteristics of the target object during the robot's movement along the motion sub-path, thereby predicting radiographic interference; based on the radiographic interference, it corrects the robot's drive parameters for the radiographic examination end, specifically as follows: The dynamic changes in the positioning of the target object are obtained during the movement of the radiology examination end by the robot along the action sub-path, so as to predict the imaging blur range of the radiology examination end during the movement of the action sub-path; Based on the range of image blur, the spatial drive orientation parameters of the robot's robotic arm toward the radiological examination end are corrected.

6. A control method for an intelligent collaborative robot system for radiological examination, characterized in that, The control method includes the following steps: S100: Acquire the three-dimensional point cloud data of the target object, construct a three-dimensional model of the object surface based on the three-dimensional point cloud data, and extract surface contour features from the three-dimensional model of the object surface; form the robot's initial motion path based on the expected target point distribution for radiological examination of the target object and the surface contour features; S200: Obtain the regional tissue characteristics of the expected target distribution of the radiological examination, thereby identifying the potential artifact regions of the expected target distribution of the radiological examination; based on the potential artifact regions, divide the initial action path into several action sub-paths; S300: Acquire the dynamic characteristics of the target object during the robot's movement along the action sub-path to predict radiographic interference; and correct the robot's drive parameters for the radiographic examination end based on the radiographic interference.

7. The control method according to claim 6, characterized in that, In S100, the three-dimensional point cloud data of the target object is acquired, a three-dimensional model of the object surface is constructed based on the three-dimensional point cloud data, and surface contour features are extracted from the three-dimensional model of the object surface, specifically: The original point cloud data formed by the depth camera capturing the target object is obtained. The original point cloud data is then subjected to a process of removing all out-of-bounds point cloud data within a preset limit and partially removing concentrated point cloud data to obtain three-dimensional point cloud data. The three-dimensional point cloud data is continuously fitted to construct a three-dimensional model of the object surface; The surface normal vector and surface tangent vector of the object surface three-dimensional model are calibrated and corrected to determine the surface contour point cluster of the object surface three-dimensional model; the surface contour features of the target object are determined based on the surface contour point cluster; wherein, the surface contour features include surface contour morphology and curvature distribution features.

8. The control method according to claim 7, characterized in that, In S100, based on the expected target point distribution for radiographic inspection of the target object and the surface contour features, the initial motion path of the robot is formed, specifically as follows: The spatial layout of the target clusters is obtained as the expected distribution of the target points for radiological examination of the target object. Based on the spatial layout of the target clusters, the target clusters are sorted and filled to obtain the corrected spatial layout of the target clusters. Based on the spatial layout of the modified target cluster and the surface contour features, the target cluster is mapped relative to the surface contour to obtain a continuous target layout line; based on the relative spatial relationship between the robot and the target object, the continuous target layout line is converted into the robot's initial motion path; wherein, the initial working path refers to the path by which the robot drives the radiological examination end to move.

9. The control method according to claim 6, characterized in that, In S200, the regional tissue features of the expected target point distribution for the radiological examination are obtained to identify the potential artifact regions of the expected target point distribution for the radiological examination; based on the potential artifact regions, the initial action path is divided into several action sub-paths, specifically: Acquire ultrasound images of the region where the expected target point for the radiological examination is distributed, and extract the adipose tissue thickness distribution features from the ultrasound images of the region; based on the adipose tissue thickness distribution features, estimate the imaging contrast of the sub-region where each expected target point for the radiological examination is located, thereby identifying the potential artifact regions of the expected target point distribution for the radiological examination. Based on the projection of the boundary of the potential artifact region onto the robot's motion space, the initial motion path is divided into several motion sub-paths; wherein the projection of each motion sub-path onto the boundary of the potential artifact region does not overlap.

10. The control method according to claim 6, characterized in that, In S300, the dynamic characteristics of the target object are acquired during the robot's movement along the motion sub-path to predict radiographic interference; based on the radiographic interference, the robot's drive parameters for the radiographic examination end are corrected, specifically as follows: The dynamic changes in the positioning of the target object are obtained during the movement of the radiology examination end by the robot along the action sub-path, so as to predict the imaging blur range of the radiology examination end during the movement of the action sub-path; Based on the range of image blur, the spatial drive orientation parameters of the robot's robotic arm toward the radiological examination end are corrected.

Images