System and method for repositioning broken objects

Abstract

The application discloses a kind of systems and methods of broken object splicing reset, it is related to broken object repair field, comprising: reading module, the image of the broken object after segmentation is read and input into statistical shape model;Reset algorithm operation module, the broken object fragment pose matrix is obtained by reset algorithm;Identification module, the active end surface area of complete broken object is identified using statistical shape model;Reset posture operation module, the reset posture cost function of broken object is calculated, and the active end reset pose matrix of broken object is solved;Data processing module, output target broken object and target pose;Mechanical arm, connected with data processing module, carries out splicing reset;The reset system can stably and accurately plan the reset of the fragments in the broken object, can stably and accurately identify the active end surface area of broken object, and realize the recovery of active end dislocation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of broken object repair, and more particularly, to a system for splicing and resetting a broken object and a method thereof. BACKGROUND

[0002] In the process of splicing and resetting a broken object, a technical expert often manually reads data of the broken object and plans, and the planning process is time-consuming and laborious, and the planning result is inaccurate. The existing reset planning algorithm based on the mirror side is highly dependent on the integrity and symmetry of the broken object, and has a small application range and poor effect. SUMMARY

[0003] The present application aims at the deficiencies of the prior art, and provides a system for splicing and resetting a broken object and a method thereof. The automatic resetting device can stably and accurately plan the resetting of the main block and the broken block of the broken object, can stably and accurately identify the active end surface area of the broken object, and can restore the dislocation of the active end according to the active end surface area, thereby providing a new automatic planning scheme for broken object repair and resetting.

[0004] In order to achieve the above-mentioned purpose, the present application provides a system for splicing and resetting a broken object, comprising:

[0005] A reading module reads and inputs the image of the segmented broken object into a statistical shape model;

[0006] A reset algorithm operation module obtains the broken block pose matrix of the broken object by using the mirror image information of the broken object and the statistical shape model through a reset algorithm;

[0007] An identification module identifies the active end surface area of the complete broken object by using the statistical shape model;

[0008] A reset pose operation module calculates the reset pose cost function of the broken object through the landmark points of the active end surface area of the broken object, and solves the reset pose cost function of the broken object by using an optimization algorithm to obtain the active end reset pose matrix of the broken object;

[0009] A data processing module combines the broken block pose matrix of the broken object and the active end reset pose matrix of the broken object to output a target broken object and a target pose;

[0010] A mechanical arm connected with the data processing module can splice and reset the broken object according to the target broken object and the target pose.

[0011] Optionally, the image of the segmented broken object includes broken object main block data and broken object broken block data.

[0012] Optionally, the reset algorithm operation module comprises a first registration unit, which is capable of registering the main block data of the broken object and the mirror image data of the broken object with the statistical shape model of the broken object.

[0013] Optionally, the mirror image data of the broken object is obtained in a self-mirror mode and a symmetric side mirror mode.

[0014] Optionally, the reset algorithm operation module further comprises a parameter updating unit, which is capable of updating the parameters of the statistical shape model of the broken object by calculating the difference between the statistical shape model of the broken object and the mirror image data of the broken object.

[0015] Optionally, the reset algorithm operation module is provided with a first threshold value, and the broken object's block pose matrix can be obtained when the parameter change value of the statistical shape model of the broken object is less than the first threshold value during the updating of the parameters.

[0016] Optionally, the reset pose operation module comprises a second registration unit, which is capable of registering the broken object with the statistical shape model after the initialization of the parameters of the statistical shape model with the marker points.

[0017] Optionally, the reset pose operation module is provided with a second threshold value, and the updating parameters of the statistical shape model can be obtained by calculating the difference between the statistical shape model and the main block data of the broken object, and the active end surface area of the broken object can be obtained when the parameter change value of the statistical shape model is less than the second threshold value during the updating of the parameters.

[0018] Optionally, the reset pose operation module comprises a calculation unit, which is capable of obtaining the reset pose cost function of the broken object by the marker points on the active end surface area.

[0019] The application further provides a method for the reset of the broken object splicing, which utilizes the system for the reset of the broken object splicing.

[0020] The segmented image of the broken object is read and input into the statistical shape model.

[0021] The broken object's block pose matrix is obtained by the reset algorithm using the mirror image information of the broken object and the statistical shape model.

[0022] The active end surface area of the broken object is identified by the statistical shape model.

[0023] The reset pose cost function of the broken object is calculated by the marker points on the active end surface area of the broken object.

[0024] The active end reset pose matrix of the broken object is obtained by solving the reset pose cost function of the broken object by the optimization algorithm.

[0025] The broken object piece pose matrix and the broken object active end reset pose matrix are combined to output a target broken object and a target pose, and the broken object is spliced and reset.

[0026] The broken object splicing reset system provided by the application has the advantages that the system is a full-automatic reset planning process, the system combines a statistical shape model and mirror data to realize broken object reset, and can identify an active end of the broken object and a main piece of the broken object to obtain a target pose of the broken object.

[0027] Other features and advantages of the present application will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0028] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the several views, and in which:

[0029] Figure 1 A flowchart of a process of performing splicing reset on a pelvic specimen by a broken object splicing reset system according to an embodiment of the present application is shown.

[0030] Figure 2 A flowchart of a pelvic specimen reset algorithm based on a statistical shape model according to an embodiment of the present application is shown.

[0031] Figure 3 A flowchart of joint surface region identification of a pelvic specimen based on a statistical shape model according to an embodiment of the present application is shown.

[0032] Figure 4 A flowchart of pose reset of a pelvic specimen based on identified joint surfaces according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0033] Preferred embodiments of the present application will be described in more detail below. Although the following describes preferred embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present application is more thoroughly and completely conveyed to those skilled in the art, and the scope of the present application is fully conveyed to those skilled in the art.

[0034] The present application provides a broken object splicing reset system, comprising:

[0035] A reading module reads and inputs an image of the segmented broken object into a statistical shape model;

[0036] a reset algorithm operation module, which obtains a broken object fragment pose matrix of the broken object by a reset algorithm using mirror information of the broken object and a statistical shape model;

[0037] a recognition module, which recognizes a moving end surface area of the complete broken object using the statistical shape model;

[0038] a reset pose operation module, which calculates a reset pose cost function of the broken object through a landmark point of the moving end surface area of the broken object, and solves the reset pose cost function of the broken object through an optimization algorithm to obtain a moving end reset pose matrix of the broken object;

[0039] a data processing module, which combines the broken object fragment pose matrix and the moving end reset pose matrix of the broken object to output a target broken object and a target pose;

[0040] a mechanical arm, which is connected with the data processing module, and can splice and reset the broken object according to the target broken object and the target pose.

[0041] Specifically, the reset system first establishes a statistical shape model using a database of broken objects, then inputs an image of a segmented broken object, and plans the broken object using a broken object reset planning algorithm combining mirror information and the statistical shape model, so as to obtain a reset planning scheme for the broken object.

[0042] The following takes a pelvis specimen as a broken object. The reset system recognizes a moving end surface area of the complete broken object using a statistical shape model, then calculates a reset pose cost function of the broken object through a landmark point of the moving end surface area, and uses an optimization algorithm to solve the reset pose cost function of the broken object to reach a minimum value, so as to obtain a final target pose as a final planning scheme for dislocation recovery of the moving end of the broken object.

[0043] Optionally, the image of the segmented broken object includes broken object main block data and broken object fragment data.

[0044] Specifically, the broken object is segmented into a broken object main block and a broken object fragment, and the corresponding statistical shape model and mirror data can be found according to the broken object main block, so as to facilitate subsequent difference calculation.

[0045] Optionally, the reset algorithm operation module includes a first registration unit, which can register the broken object main block data and the mirror data of the broken object with a statistical shape model template of the broken object.

[0046] Optionally, the mirror data of the broken object is obtained in a self-mirror mode and a symmetric side mirror mode.

[0047] Optionally, the reset algorithm operation module further comprises a parameter updating unit, which updates the statistical shape model of the broken object by calculating the difference between the statistical shape model of the broken object and the mirror image data of the broken object.

[0048] Optionally, the reset algorithm operation module is provided with a first threshold value, and the statistical shape model of the broken object is capable of obtaining the fragment pose matrix of the broken object when the parameter change value is less than the first threshold value during the parameter updating.

[0049] Specifically, in the reset algorithm for the broken object, a statistical shape template is randomly extracted from the statistical shape model, and the statistical shape template is first registered with the main block of the broken object and the mirror image data of the broken object, so as to obtain the actual data of the broken object. The parameter of the statistical shape template is deformed according to the difference between the statistical shape template and the data of the main block of the broken object and the broken object. When the parameter deformation of a certain statistical shape template is less than the first threshold value, it is indicated that the statistical shape template is the optimal planning result for resetting the broken object, and the statistical shape template is the fragment pose matrix of the broken object.

[0050] Optionally, the reset pose operation module comprises a second registration unit, which registers the broken object with the statistical shape model after initializing the parameter of the statistical shape model with the marker point.

[0051] Optionally, the reset pose operation module is provided with a second threshold value, and the difference between the statistical shape model and the data of the main block of the broken object is calculated to obtain the updated parameter of the statistical shape model. When the parameter change value is less than the second threshold value during the parameter updating of the statistical shape model, the active end surface region of the broken object is obtained.

[0052] Specifically, a statistical shape template is randomly extracted from the statistical shape model with the marker point, and is registered with the broken object. The parameter of the statistical shape template is updated by calculating the difference between the statistical shape template and the main block of the broken object, and the deformation of the statistical shape template is realized. When the parameter change value of a certain statistical shape template is less than the second threshold value, the statistical shape template is capable of outputting the matched active end surface region of the broken object. The marker point of the complete active end surface region of the broken object is taken as the input of the active end registration algorithm, and the active end pose of the optimal broken object is calculated. An optimization algorithm is used to solve the pose matrix that makes the reset cost function minimum, i.e. the active end reset pose matrix of the broken object.

[0053] In one embodiment, the optimization algorithm can be a covariance matrix adaptive evolution strategy.

[0054] Optionally, the reset pose calculation module comprises a calculation unit, which obtains the reset pose cost function of the broken object through the landmark points on the active end surface area.

[0055] The application further provides a method for resetting a broken object, using the system for resetting a broken object, and the method comprises the following steps:

[0056] reading the segmented image of the broken object and inputting the image into the statistical shape model;

[0057] obtaining the fragment pose matrix of the broken object through the reset algorithm and the mirror information of the broken object and the statistical shape model;

[0058] identifying the active end surface area of the complete broken object through the statistical shape model;

[0059] calculating the reset pose cost function of the broken object through the landmark points on the active end surface area of the broken object;

[0060] obtaining the active end reset pose matrix of the broken object through the optimization algorithm and the reset pose cost function of the broken object;

[0061] combining the fragment pose matrix of the broken object and the active end reset pose matrix of the broken object to output the target broken object and the target pose, and resetting the broken object.

[0062] Embodiment

[0063] The application provides a system for resetting a broken object, comprising:

[0064] a reading module, which reads the segmented image of the broken object and inputs the image into the statistical shape model;

[0065] a reset algorithm calculation module, which obtains the fragment pose matrix of the broken object through the reset algorithm and the mirror information of the broken object and the statistical shape model;

[0066] an identification module, which identifies the active end surface area of the complete broken object through the statistical shape model;

[0067] a reset pose calculation module, which calculates the reset pose cost function of the broken object through the landmark points on the active end surface area of the broken object, and obtains the active end reset pose matrix of the broken object through the optimization algorithm and the reset pose cost function of the broken object;

[0068] a data processing module, which combines the fragment pose matrix of the broken object and the active end reset pose matrix of the broken object to output the target broken object and the target pose;

[0069] The mechanical arm is connected with the data processing module, and can splice and reset the fractured object according to the target fractured object and the target pose.

[0070] Optionally, the image of the segmented fractured object comprises fractured object main block data and fractured object fragment data.

[0071] Optionally, the reset algorithm operation module comprises a first registration unit, which can register the fractured object main block data and the mirror image data of the fractured object with the statistical shape template of the fractured object.

[0072] Optionally, the mirror image data of the fractured object is obtained in a self-mirror mode and a symmetric side mirror mode.

[0073] Optionally, the reset algorithm operation module further comprises a parameter updating unit, which updates the statistical shape model of the fractured object by calculating the difference between the statistical shape model of the fractured object and the mirror image data of the fractured object.

[0074] Optionally, the reset algorithm operation module is provided with a first threshold value, and when the parameter change value of the statistical shape model of the fractured object is less than the first threshold value during the parameter updating, the fragment pose matrix of the fractured object can be obtained.

[0075] Optionally, the reset pose operation module comprises a second registration unit, which registers the fractured object with the statistical shape model after initializing the parameters of the statistical shape model with the marker points.

[0076] Optionally, the reset pose operation module is provided with a second threshold value, and the difference between the statistical shape model and the fractured object main block data is calculated to obtain the updated parameters of the statistical shape model, and when the parameter change value of the statistical shape model is less than the second threshold value during the parameter updating, the active end surface area of the fractured object is obtained.

[0077] Optionally, the reset pose operation module comprises a calculation unit, which obtains the reset pose cost function of the fractured object through the marker points on the active end surface area.

[0078] The application also provides a method for splicing and resetting a fractured object, which utilizes the above-mentioned system for splicing and resetting a fractured object, and the method comprises the following steps:

[0079] The image of the segmented fractured object is read and input into the statistical shape model;

[0080] The mirror image information of the fractured object and the statistical shape model are utilized to obtain the fragment pose matrix of the fractured object through the reset algorithm;

[0081] The active end surface area of the complete fractured object is identified by utilizing the statistical shape model;

[0082] The repositioning pose cost function of the fractured object is calculated by the landmark points of the active end surface area of the fractured object;

[0083] The active end repositioning pose matrix of the fractured object is obtained by solving the repositioning pose cost function of the fractured object through an optimization algorithm;

[0084] The target fractured object and the target pose are output by combining the fragment pose matrix of the fractured object and the active end repositioning pose matrix of the fractured object, and the fractured object is spliced and repositioned.

[0085] Referring to Figure 1 , the specific process of the system for repositioning the fractured object to plan the repositioning of the pelvic specimen is as follows:

[0086] The broken pelvic specimen data is input to the statistical shape model, and the broken pelvic specimen data includes the sacrum, left and right iliac bones, and the data of the broken fragments thereof;

[0087] The broken sacrum and iliac bones that need to be repositioned are repositioned;

[0088] The pubic symphysis and sacroiliac joint surface regions of the complete / planned left and right iliac bones and sacrum are identified using the statistical shape model with joint region annotation;

[0089] The obtained complete left and right iliac bones and joint landmarks and sacrum and joint landmarks are input to the joint registration algorithm to calculate the optimal pelvic joint pose;

[0090] The pose matrix that minimizes the pelvic repositioning cost function is solved using an optimization algorithm;

[0091] The optimal target pose and the optimal target pelvic bone are output by combining the single-bone fracture planning pose matrix and the joint repositioning pose matrix, and the process ends.

[0092] The pelvic repositioning cost function includes the following points: the distance between the left and right sacroiliac joints, the distance between the pubic symphysis, and the collision coefficient between the sacrum and the left and right iliac bones.

[0093] Referring to Figure 2 , the specific process of the pelvic specimen fracture repositioning algorithm of the system for repositioning the fractured object based on the statistical shape model is as follows:

[0094] a. Input the segmented target bone fragment pattern to the statistical shape model;

[0095] b. Reposition the bone specimen that needs to be repositioned, and first register the main fragment and the mirror bone with the statistical shape template;

[0096] c. Calculate the difference between the target bone and the statistical shape template through the point correspondence between the target main fragment and the statistical shape template;

[0097] d. Calculate the statistical shape model iteration parameters and update the deformation of the statistical shape template;

[0098] e. Register the other fracture fragments with the statistical deformation template to obtain an updated statistical deformation model bone, and obtain the current fracture reduction planning result;

[0099] f. Repeat steps b-d, and when the statistical shape model parameter change value is less than a given threshold, end the iterative optimization and output the best fracture reduction planning result, i.e., the fracture fragment pose matrix.

[0100] Referring to Figure 3 , the specific process of the system for describing the fracture object splicing reduction based on joint landmark area recognition of the statistical shape model is as follows:

[0101] a. Input the segmented target bone fragment pattern into the statistical shape model with an initialized landmark point;

[0102] b. Register the target bone with the statistical shape template;

[0103] c. Calculate the difference between the target bone and the statistical shape template through the point correspondence between the target bone and the statistical shape template;

[0104] d. Calculate the statistical shape model iteration parameters and update the deformation of the statistical shape template;

[0105] e. Search for the nearest neighbor points of the statistical shape template joint surface area on the reduction bone to serve as the current landmark area recognition result;

[0106] f. Repeat steps b-d, and when the statistical shape model parameter change value is less than a given threshold, end the iterative optimization and output the matching joint surface area.

[0107] Referring to Figure 4 , the specific process of the system for describing the fracture object splicing reduction based on joint surface recognition of the pelvic specimen pose reduction is as follows:

[0108] Input the complete left and right iliac bones and sacral bone joint surface areas and their joint landmark points in the joint registration algorithm, and calculate the best pelvic joint pose;

[0109] In the process of converting the initialization pose parameters to the misaligned iliac bone pose, it is necessary to calculate the cost function and calculate the pose parameter gradient to update the pose parameters;

[0110] The evolutionary strategy with covariance self-adaptive adjustment is used to make the pose parameter change value less than a given threshold, so as to obtain the optimal pose parameter, i.e. the joint reset pose matrix, and end.

[0111] Embodiments of the application have been described above, with the understanding that these embodiments are merely examples, are not exhaustive, and are not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art, without departing from the scope and spirit of the described embodiments.

Claims

1. A system for reassembling and restoring a broken object, characterized in that, The system includes: The reading module reads the image of the segmented fractured object and inputs it into the statistical shape model; The reset algorithm calculation module uses the mirror image information and statistical shape model of the fractured object to obtain the fragment pose matrix of the fractured object through the reset algorithm; The identification module uses a statistical shape model to identify the active end surface region of a complete fractured object; The reset posture calculation module calculates the reset posture cost function of the fractured object by using the marker points on the surface region of the moving end of the fractured object, and obtains the reset pose matrix of the moving end of the fractured object by solving the reset posture cost function of the fractured object through an optimization algorithm. The data processing module combines the fragment pose matrix of the fractured object and the reset pose matrix of the movable end of the fractured object to output the target fractured object and the target pose. A robotic arm, connected to a data processing module, is capable of reassembling and repositioning a broken object based on its target fracture and pose.

2. The system for reassembling and repositioning a broken object according to claim 1, characterized in that, The image of the segmented fractured object includes the main block data of the fractured object and the fragment data of the fractured object.

3. The system for reassembling and restoring a broken object according to claim 2, characterized in that, The reset algorithm operation module includes a first registration unit, which can register the main block data of the fractured object and the mirror data of the fractured object with the statistical shape template of the fractured object.

4. The system for splicing and repositioning a broken object according to claim 3, characterized in that, The mirror data of the fractured object is obtained by using both self-mirror and symmetrical side mirror methods.

5. The system for reassembling and repositioning a broken object according to claim 4, characterized in that, The reset algorithm operation module also includes a parameter update unit, which updates the parameters of the statistical shape model of the fractured object by calculating the difference between the statistical shape model of the fractured object and the mirror image data of the fractured object.

6. The system for reassembling and repositioning a broken object according to claim 5, characterized in that, The reset algorithm calculation module is set with a first threshold. When the parameter change value of the statistical shape model of the fractured object is less than the first threshold when updating the parameters, the fragment pose matrix of the fractured object can be obtained.

7. The system for splicing and repositioning a broken object according to claim 2, characterized in that, The reset attitude calculation module includes a second registration unit. After initializing the parameters of the statistical shape model with markers, the second registration unit registers the fractured object with the statistical shape model.

8. The system for reassembling and repositioning a broken object according to claim 7, characterized in that, The reset posture calculation module is set with a second threshold. It calculates the difference between the statistical shape model and the main block data of the fractured object to obtain the updated parameters of the statistical shape model. When the parameter change value of the statistical shape model is less than the second threshold when updating the parameters, the active end surface area of ​​the fractured object is obtained.

9. The system for reassembling and repositioning a broken object according to claim 8, characterized in that, The reset attitude calculation module includes a calculation unit, which obtains the reset attitude cost function of the broken object through the marker points on the active end surface area.

10. A method for reassembling and repositioning a broken object, utilizing the system for reassembling and repositioning a broken object according to any one of claims 1-9, characterized in that, The method includes: The image of the segmented fractured object is read and input into the statistical shape model; The fragment pose matrix of the fractured object is obtained by using the mirror image information and statistical shape model of the fractured object through a reset algorithm. Statistical shape models are used to identify the active end surface regions of complete fractured objects. The reset posture cost function of the broken object is calculated by using the marker points on the active end surface region of the broken object. The reset pose matrix of the active end of the fractured object is obtained by solving the reset pose cost function of the fractured object using an optimization algorithm. The fragment pose matrix of the fractured object and the reset pose matrix of the movable end of the fractured object are combined to output the target fractured object and the target pose, and then spliced ​​and reset.

Images