WO2017185208A1 - 一种机器人三维模型的建立方法、装置及电子设备 - Google Patents

一种机器人三维模型的建立方法、装置及电子设备 Download PDF

Info

Publication number
WO2017185208A1
WO2017185208A1 PCT/CN2016/080133 CN2016080133W WO2017185208A1 WO 2017185208 A1 WO2017185208 A1 WO 2017185208A1 CN 2016080133 W CN2016080133 W CN 2016080133W WO 2017185208 A1 WO2017185208 A1 WO 2017185208A1
Authority
WO
WIPO (PCT)
Prior art keywords
robot
mirror
image
predetermined
dimensional model
Prior art date
Application number
PCT/CN2016/080133
Other languages
English (en)
French (fr)
Inventor
骆磊
Original Assignee
深圳前海达闼云端智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳前海达闼云端智能科技有限公司 filed Critical 深圳前海达闼云端智能科技有限公司
Priority to CN201680002792.3A priority Critical patent/CN107004298B/zh
Priority to PCT/CN2016/080133 priority patent/WO2017185208A1/zh
Publication of WO2017185208A1 publication Critical patent/WO2017185208A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J19/00—Accessories fitted to manipulators, e.g. for monitoring, for viewing; Safety devices combined with or specially adapted for use in connection with manipulators

Definitions

  • the present application relates to a robot technology, and in particular, to a method, an apparatus, and an electronic device for establishing a three-dimensional model of a robot.
  • the cloud robot is an intelligent robot that puts the cognitive system in the cloud, the body, the drive and the sensor are placed on the robot body, and connects the two through mobile communication; the cloud robot is the development direction of the intelligent humanoid robot.
  • the prior art adopts the following scheme: the robot provides a setting interface of a 3D (3-Dimensions, three-dimensional) model, and the maintenance personnel sets the 3D model of the robot according to the current shape of the robot, and stores the 3D model to The robot's control module is saved, and then the robot dynamically controls the robot's respective degrees of motion based on the 3D model, so that the robot can avoid obstacles or can avoid the respective degrees of collision.
  • the inventors have found that the method for establishing a three-dimensional model of a robot in the prior art is changed if a certain structural shape of the robot changes, for example, a new functional module is added, a new outer casing is replaced, or the robot has When the load is heavy, the maintenance personnel need to reset the 3D model of the robot through the interface and update to the control module, resulting in cumbersome operation.
  • a method and a device for establishing a three-dimensional model of a robot are provided, which are used to solve the problem of cumbersome operation when the three-dimensional model of the robot is established in the prior art.
  • a method for establishing a three-dimensional model of a robot includes: determining that a mirror surface exists in a visible region when detecting a first predetermined condition that satisfies establishing a model; acquiring the robot at the mirror surface The image in the image; based on the acquired image, a three-dimensional model of the robot is established.
  • satisfying the first predetermined condition for establishing the model specifically includes: initializing the robot; or increasing or decreasing the weight of the robot itself beyond a first predetermined threshold when the robot is not in the state of extracting an object or lifting an object.
  • determining that the visible area exists in the visible area comprises: acquiring one or more images in the visible area of the robot, and separately recording the time of acquiring each image; determining corresponding actions performed by the robot at each moment; and identifying Whether there is content in the image corresponding to the action performed by the robot at the corresponding time in the image at the corresponding time; when there is a number of images of the content consistent with the action performed by the robot at the corresponding time exceeding the second predetermined threshold, determining the visual There is a mirror in the area.
  • determining that the visible area exists in the visible area comprises: acquiring pre-stored position information, where the position information includes a position of the mirror; and triggering a predetermined distance that the robot moves to the position.
  • the method further includes: determining that the mirror meets the second predetermined condition.
  • satisfying the second predetermined condition specifically includes: an angle between the mirror surface and a body upright axis of the robot is less than a first predetermined angle.
  • determining that the mirror meets the second predetermined condition comprises: triggering a predetermined structure of the robot to start rotating with the body upright axis as a central axis; and rotating, when rotating, a second predetermined angle, perpendicular to the body Transmitting a first ray in a direction of the upright axis or emitting a first ray in a direction perpendicular to the upright axis of the body every first predetermined time; wherein an axis of the predetermined structure is perpendicular to the upright axis of the body; receiving the specular reflection a second light corresponding to the first light; when the angle between the first light and the second light is less than a third predetermined angle, determining to determine that the mirror is full The second predetermined condition.
  • acquiring an image of the robot in the mirror surface specifically includes: an image capturing component that triggers the robot is facing the mirror surface, and is relatively stationary with the mirror surface, and the remaining structures are rotated about a body upright axis; Upon rotation, the fourth predetermined angle of rotation triggers the image capture component to capture an image of the robot in the mirror, or triggers the image capture component to capture an image of the robot in the mirror every second predetermined time.
  • the method further includes: receiving an action instruction and an execution time corresponding to the action instruction; and at a predetermined time before the execution time, according to the action instruction and the three-dimensional The model simulates the state of the robot when the motion command is executed; according to the state, the obstacle avoidance operation is performed.
  • performing the obstacle avoidance operation according to the state specifically: determining, according to the state, whether the robot will have a degree of freedom collision when executing the motion instruction; if yes, according to the state, the collision is free. Modify the action instruction by the degree of collision between degrees.
  • performing the obstacle avoidance operation according to the state specifically, determining, according to the state and the spatial relationship of other objects in the visible area of the robot that are acquired in advance, determining whether the robot performs the motion instruction Collision of other objects; if so, modify the action command based on the location of the colliding object.
  • a device for establishing a three-dimensional model of a robot comprising: a mirror determining module, configured to determine that a visible region exists when a first predetermined condition that satisfies the model is detected is met a mirror image; an image acquisition module, configured to acquire an image of the robot in the mirror; and a three-dimensional model building module, configured to establish a three-dimensional model of the robot according to the acquired image.
  • satisfying the first predetermined condition for establishing the model specifically includes: initializing the robot; or increasing or decreasing the weight of the robot itself beyond a first predetermined threshold when the robot is not in the state of extracting an object or lifting an object.
  • the mirror determining module is specifically configured to acquire one or more of the visible areas of the robot Images, and respectively record the time at which each image is acquired; determine the corresponding action performed by the robot at each time; identify whether there is content in the image at the corresponding time that is consistent with the action performed by the robot at the corresponding time; when present and the robot When the number of images of the content consistent with the actions performed at the corresponding time exceeds the second predetermined threshold, it is determined that there is a mirror in the visible region.
  • the mirror determining module is configured to acquire pre-stored location information, where the location information includes a position of the mirror; and trigger a predetermined distance that the robot moves to the location.
  • the apparatus for establishing a three-dimensional model of the robot further includes: a second predetermined condition determining module, configured to determine that the mirror meets the second predetermined condition.
  • satisfying the second predetermined condition specifically includes: an angle between the mirror surface and a body upright axis of the robot is less than a first predetermined angle.
  • the second predetermined condition determining module is specifically configured to trigger a predetermined structure of the robot to start rotating with the body upright axis as a central axis; and to rotate the second predetermined angle, perpendicular to the body upright axis a direction of emitting the first ray or emitting a first ray in a direction perpendicular to the upright axis of the body every first predetermined time; wherein the axis of the predetermined structure is perpendicular to the body upright axis; receiving the specular reflection a second light corresponding to the first light; when the angle between the first light and the second light is less than a third predetermined angle, determining to determine that the mirror meets the second predetermined condition.
  • the image acquisition module is specifically configured to trigger the image capturing component of the robot to face the mirror surface, and remain relatively stationary with the mirror surface, and the remaining structure rotates with the body upright axis as a central axis;
  • the fourth predetermined angle triggers the image capture component to capture an image of the robot in the mirror or to trigger the image capture component to capture an image of the robot in the mirror every second predetermined time.
  • the apparatus for establishing a three-dimensional model of the robot further includes: a receiving module, configured to receive an action instruction and an execution time corresponding to the action instruction; and a state simulation module, configured to: at the predetermined time before the execution time, according to the action instruction and The three-dimensional model simulates a state of the robot when the motion command is executed; the obstacle avoidance module is configured to perform an obstacle avoidance operation according to the state.
  • the obstacle avoidance module is specifically configured to determine, according to the state, whether the robot will have a degree of freedom collision when executing the motion instruction; if yes, according to the degree of collision between the degrees of freedom in which the collision occurs Modify the action instructions.
  • the obstacle avoidance module is specifically configured to determine, according to the state and the spatial relationship of other objects in the visible area of the robot, whether the robot collides with the other object when executing the motion instruction; , modify the action command according to the position of the colliding object.
  • the method, device and electronic device for establishing a three-dimensional model of the robot in the embodiment of the present application further determine that a mirror surface exists in the visible region when the first predetermined condition is met, and obtain an image of the robot in the mirror surface, and then according to The acquired image establishes a three-dimensional model of the robot, so that when the robot detects that the first predetermined condition is satisfied, it triggers an action of establishing its own three-dimensional model through the mirror, thereby eliminating the need for maintenance personnel setting and reducing the operation steps.
  • FIG. 1 is a flow chart showing a method for establishing a three-dimensional model of a robot according to Embodiment 1 of the present application;
  • FIG. 2 is a flow chart showing a method for establishing a three-dimensional model of a robot according to Embodiment 2 of the present application;
  • FIG. 3 is a schematic structural diagram of a device for establishing a three-dimensional model of a robot according to Embodiment 3 of the present application;
  • FIG. 4 is a schematic structural diagram of an electronic device according to Embodiment 4 of the present application.
  • the inventors have found that, using the method for establishing a three-dimensional model of a robot in the prior art, if a certain structural shape of the robot changes, for example, a new functional module is added, and a new one is replaced.
  • the maintenance personnel need to reset the three-dimensional model of the robot through the interface and update to the control module, resulting in cumbersome operation.
  • the inventor also found that although the robot has made great progress in acquiring external information, the robot lacks the ability to understand its own shape change, which leads to the judgment of the original size after the change of its own size. It is not possible to avoid obstacles correctly, for example, when the hand hits the foot again, when it is out, it is stuck at the door, the vase is knocked down, and so on.
  • a method and a device for establishing a three-dimensional model of a robot are provided.
  • the first predetermined condition it is determined that a mirror surface exists in the visible region, and an image of the robot in the mirror surface is obtained.
  • the three-dimensional model of the robot is established, so that when the robot detects that the first predetermined condition is met, the action of establishing a three-dimensional model of the mirror through the mirror is triggered, thereby eliminating the need for maintenance personnel to set the operation step.
  • the state of the robot when the motion command is executed is simulated according to the three-dimensional model and the motion command, and the obstacle avoidance operation is performed according to the state, thereby making the robot Can avoid obstacles correctly.
  • the method for establishing a three-dimensional model of the robot in the present application may be implemented by a 3D modeling module installed on a robot head or other rotatable part, or may be combined by multiple modules in other modules of the robot. Implementation, this application does not limit this.
  • the words "the robot sees" and "in the visible area of the robot” mean that the related content is in the image capturing component of the robot. Within the area of the image.
  • FIG. 1 is a flow chart showing a method for establishing a three-dimensional model of a robot according to Embodiment 1 of the present application. As shown in Figure 1, the following steps are included:
  • the initialization may be an initialization process performed when the robot is assembled, and is performed when the robot is first powered on, or may be an initialization process that is re-executed by returning to the factory setting after the robot is assembled and operated.
  • the first predetermined threshold may be a preset weight value, for example, 0.5 kg, etc.; specifically, the first predetermined threshold may be an average of the functional modules of the robot according to the developer of the 3D modeling module.
  • the value of the weight setting may also be a value set by the maintenance personnel of the robot according to the weight change value of the commonly used replaceable casing, the function module, etc. of the robot, and the like, which is not limited in this application.
  • the first predetermined threshold is 1 kg; at a certain moment, the foot pressure sensor of the robot detects that the weight of the robot is increased by 2 kg, and the data returned by the pressure sensor of the hand or the upper limb of the robot can be further obtained; If the data returned by the hand pressure sensor of the machine indicates that the robot has lifted the object of 2kg, it is judged that the first predetermined condition is not satisfied at this time; if the data returned by the hand pressure sensor of the robot is 0, then It is judged that 2 kg is greater than 1 kg; it is determined that the first predetermined condition is satisfied.
  • the determination step of whether the weight increase or decrease of the robot itself exceeds the first predetermined threshold there is no strict timing relationship between whether the robot is in the state of extracting the object or lifting the object, and the determination step of whether the weight increase or decrease of the robot itself exceeds the first predetermined threshold. That is, it can be first determined whether the robot is in the state of extracting the object or lifting the object, and further determining the weight of the robot itself when not in the state of extracting the object or lifting the object Whether the robot increases or decreases the first predetermined threshold; or determines whether the weight increase or decrease of the robot itself exceeds a first predetermined threshold, and further determines whether the robot is in the extracted object or when the weight increases or decreases beyond the first predetermined threshold.
  • the state of the lifting object it can also be carried out in two steps; this application does not limit this.
  • the visible area of the robot may refer to an image capturing component of the robot, for example, a range in which the camera can acquire an image.
  • the robot can find the mirror by image recognition.
  • the robot may acquire one or more images in the visible area of the robot through an image capturing component such as a camera, and separately record the time at which each image is acquired; determine corresponding actions performed by the robot at each time; identify the corresponding time Whether there is content in the image that is consistent with the action performed by the robot at the corresponding time; when there is a number of images of the content that is consistent with the action performed by the robot at the corresponding time exceeds a second predetermined threshold, then the determination may be There is a mirror in the view area.
  • the second predetermined threshold may be 1, or may be a predetermined threshold such as 2, 3, 5, or the like.
  • a robot shape completely coincident with its own motion when a robot shape completely coincident with its own motion is found in the image, it may first be determined that there may be a mirror surface in the direction; when it is determined that there may be a mirror surface in the direction, the specific motion may be further performed and further acquired.
  • the robot when detecting that the first predetermined condition is met, may interrupt the currently executed running instruction, directly acquire an image in the current visible area to determine whether there is a mirror surface, or may not interrupt the currently executed running instruction. Whether or not a mirror is present is determined based on the acquired image in normal motion or in a normal execution task.
  • the mirror can also be found by reading pre-stored information.
  • the pre-stored location information may be read, the location information including the position of the mirror;
  • the predetermined distance may be a distance that enables the robot to see its own whole body image in the mirror surface, or may be a distance that the robot can see a specific part of itself in the mirror surface; specifically, the specific part may include setting according to experience The body part of the robot's shape changes.
  • the predetermined distance can be calculated by using the mirror imaging principle according to the height of the robot, the height of the mirror, and the like; the present application will not be described herein.
  • the location information may be multiple, the robot may select the mirror closest to itself; or the robot may select the best route between the current location, the mirror, and the destination according to the destination in the next running instruction. .
  • the pre-stored location information may be a location preset by a maintenance personnel, for example, on the left side of the living room door, on the bedroom closet door, or the like; or the position of the mirror automatically recorded by the robot, etc.; No restrictions.
  • the position information of the pre-stored mirror may be read when the mirror is not found in the current visible area or in the normal motion for a predetermined length of time; then the predetermined distance or the like for the robot to move to the position is triggered.
  • the modeling is inaccurate.
  • the first light may be emitted to the mirror surface, and whether the mirror surface is parallel to the vertical axis of the robot body is determined according to the angle between the received reflected light and the first light.
  • the first predetermined angle may be a smaller angle such as 1 degree, 2 degrees, or the like.
  • the predetermined structure of the triggering robot starts with the body upright axis as a center axis; when rotating, each time Rotating a second predetermined angle to emit a first ray in a direction perpendicular to the upright axis of the body, or along a square perpendicular to the upright axis of the body every first predetermined time Transmitting a first light; wherein the axis of the predetermined structure is perpendicular to the body upright axis; receiving the mirror-reflected second light corresponding to the first light; the angle between the first light and the second light is less than At the third predetermined angle, it is determined that the angle between the mirror surface and the body upright axis of the robot is less than the first predetermined angle.
  • the first light may be light such as infrared rays or laser light.
  • the predetermined structure may be the head of the robot or other structure that can be rotated 360 degrees around the axis of the upright axis of the body, which is not limited in this application.
  • the second predetermined angle may be a smaller angle such as 1 degree, 2 degrees, or the like.
  • the third predetermined angle may be a smaller angle such as 1 degree, 2 degrees, or the like.
  • the first predetermined time may be, for example, 0.1 second, 1 second, or the like.
  • the robot can keep itself perpendicular to the vertical movement of the body and the ground; then, with the body upright axis as the central axis, 360 degrees rotate the structure of the 3D modeling module, for example, the head, and each turn 1 degree outward emits infrared rays in a direction perpendicular to the vertical axis of the body; during the rotation, a total of 35 first rays are emitted, and two second rays are received, wherein an angle between the incident rays of one of the rays and the reflected rays Less than 2 degrees; at this time, it is judged that the mirror surface is relatively parallel to the body upright axis of the robot.
  • determining that the angle between the mirror surface and the body upright axis of the robot is less than the first predetermined angle, determining that the mirror surface meets the requirements of three-dimensional modeling, at this time, the position, direction, and the like of the mirror surface may be further recorded, so that the robot can next Used when remodeling.
  • the image of the robot in the mirror surface can be obtained by: the image capturing component of the triggering robot is facing the mirror surface, and is relatively stationary with the mirror surface, and the rest of the structure is rotated about the body vertical axis; when rotating, Each rotation of the fourth predetermined angle triggers the image capturing component to capture an image of the robot in the mirror, or triggers the image capturing component to capture an image of the robot in the mirror every second predetermined time.
  • step S104 it can be received When the angle between the incident light and the reflected light is less than a third predetermined angle, it is determined that the image capturing component is facing the mirror at this time, triggering the predetermined structure to stop rotating, and acquiring an image of the robot in the mirror.
  • the orientation of the image capturing component of the robot may be further adjusted before the image of the robot in the mirror is acquired, so that the image capturing component is directly opposite.
  • Mirror surface For example, the orientation of the image capturing component is 180 degrees from the direction of light emission; and when the light is emitted to the left, receiving a third predetermined angle and reflected light that is less than an angle with the incident light, and stopping the reservation.
  • the structure rotates and will rotate 180 degrees with the robot's image capture component.
  • the mirror surface may be a mirror surface that the robot can see the whole body image, or may be a mirror surface that the robot can see a specific part of the robot; specifically, the specific portion may include a shape change of the robot according to experience. Or a body part that is loaded with weight; this application does not limit this.
  • the robot head if the image capturing component of the 3D modeling module is located at the robot head, the robot head is stationary and rotates along the axis below the neck; during the rotation, the image capturing component can be photographed in real time, that is, every predetermined time period For example, one shot is taken in 0.1 second; or the rest of the structure may be taken at a predetermined angle, for example, 1 degree.
  • the robot can move a specific degree of freedom or trigger a certain degree of freedom to pose a specific shape; in particular, the specific degree of freedom can be a body shape change based on empirical values or a weight-bearing body Partial.
  • a three-dimensional model of the robot may be established according to the acquired two-dimensional image by using the prior art, for example, Image-Based Modeling and Rendering (IBMR) technology.
  • IBMR Image-Based Modeling and Rendering
  • the main purpose of image-based modeling is to restore the three-dimensional geometry of the scene from a two-dimensional image.
  • IBMR technology has many unique advantages, such as faster and more convenient modeling, high drawing speed and high realism.
  • Obtain stereo mode with traditional modeling software or 3D scanner Compared with the type method, the image modeling based method is low in cost, strong in realism and high in automation.
  • Many companies have developed such algorithms.
  • the domestic 3D Cloud runs in the cloud, and as long as the photos are uploaded to the cloud, the 3D model can be fully generated.
  • the Disney Zurich research team has developed a new program that uses hundreds of two-dimensional photographs and a specially designed algorithm to make complex and realistic 3D modeling of movies, TV and games, as well as to produce or print high-precision model.
  • step 105 can be implemented using a variety of algorithms that typically reconstruct a three-dimensional image from a two-dimensional image, which is shown for illustrative purposes only and is not intended to limit the application.
  • the method for establishing a three-dimensional model of the robot in the embodiment of the present application, when it is detected that the first predetermined condition is met, it is determined that a mirror surface exists in the visible region, and an image of the robot in the mirror surface is acquired, and then the image is acquired according to the acquired image.
  • the three-dimensional model of the robot is capable of triggering an action of establishing a three-dimensional model of itself through the mirror when the robot detects that the first predetermined condition is satisfied, thereby eliminating the need for maintenance personnel to set up, reducing operational steps.
  • the implementation of the method for establishing a three-dimensional model of the robot in the embodiment of the present application is described in detail.
  • the robot may also perform a corresponding obstacle avoidance action according to the newly established three-dimensional model; a description will be made below in conjunction with another embodiment.
  • FIG. 2 A flow chart of a method of establishing a three-dimensional model of a robot according to a second implementation of the present application is shown in FIG. As shown in FIG. 2, the method for establishing a three-dimensional model of a robot according to the second embodiment of the present application includes the following steps:
  • the obstacle can be avoided according to the three-dimensional model stored by the robot.
  • the three-dimensional model may be an original three-dimensional model; or may be a three-dimensional model established in steps 201-205.
  • the obstacle avoidance may be adopted by: receiving an execution time corresponding to the action instruction and the action instruction, and executing the simulation robot according to the action instruction and the three-dimensional model at a predetermined time before the execution time a state at the time of the motion command; then determining whether the robot has a degree of freedom collision when the motion command is executed; if so, modifying the motion command according to the degree of collision between the degrees of freedom in which the collision occurs, In order to cause a collision The part becomes just squatted or left with a small gap that does not touch.
  • the predetermined time may be a length of time set according to an empirical value, for example, 2 minutes, 1 minute, 10 seconds, etc. before the execution time.
  • the state of the robot when executing the motion instruction may be simulated according to the motion instruction and the three-dimensional model in a simulation manner.
  • the action command is: "lift the right hand forward 10 cm”; then trigger the simulated three-dimensional model to perform the action of "lifting the right hand forward 10 cm”.
  • the action instruction may be modified according to the degree of the collision, for example, canceling the The action command is changed to "lift the right hand forward 5 cm" and so on.
  • the obstacle avoidance may be adopted by: receiving an execution time corresponding to the action instruction and the action instruction, and executing the simulated robot according to the action instruction and the three-dimensional model at a predetermined time before the execution time a state at the time of the motion command; and then determining, according to the state and the spatial relationship of the other objects in the visible area of the robot that are acquired in advance, whether the robot collides with the other object when executing the motion instruction; Yes, the action command is modified according to the position of the colliding object.
  • the robot may collide with the object in combination with the spatial relationship of other objects in the visible area, and whether there is a problem in the passability. If found, the motion mode or the motion track may be modified to avoid possible collision. .
  • the action command is “moving 20 cm forward”, and 10 cm in front of the robot is the door.
  • the size of the door is 2 meters high and 1 meter wide.
  • the robot changes shape and the width exceeds 1 meter.
  • the action command can be modified, for example, cancel the action command or change to "20 degrees to the right, 20 cm to the right".
  • the method for establishing a three-dimensional model of the robot in the embodiment of the present application after establishing the three-dimensional model of the robot through the mirror surface, after receiving the motion instruction, simulates the state of the robot when executing the motion instruction according to the three-dimensional model and the motion instruction. And performing an obstacle avoidance operation according to the state, thereby This makes the robot correct obstacle avoidance.
  • the embodiment of the present application further provides a device for establishing a three-dimensional model of a robot.
  • the principle of solving the problem is similar to the method provided by the first and second embodiments of the present application. The implementation of the method, the repetition will not be repeated.
  • FIG. 3 is a schematic structural diagram of an apparatus for establishing a three-dimensional model of a robot according to Embodiment 3 of the present application.
  • the multiple modules in the apparatus for establishing a three-dimensional model of the robot shown in the third embodiment of the present application may be integrated into a dedicated 3D modeling module, or may be dispersed in each structure of the robot, or may be independent of The robot provides, and this application does not limit this.
  • the apparatus 3 for establishing a three-dimensional model of the robot includes: a mirror determining module 301, configured to determine that a mirror surface exists in the visible area when the first predetermined condition that satisfies the model is detected.
  • the image acquisition module 302 is configured to acquire an image of the robot in the mirror surface
  • the three-dimensional model establishing module 303 is configured to establish a three-dimensional model of the robot according to the acquired image.
  • the image acquisition module 302 may include an existing image capturing component of the robot, and may also include a new image capturing component, which is not limited in this application.
  • satisfying the first predetermined condition for establishing the model may specifically include: initializing the robot; or increasing or decreasing the weight of the robot itself beyond a first predetermined threshold when the robot is not in the state of extracting an object or lifting an object.
  • the mirror determination module may be specifically configured to acquire one or more images in the visible area of the robot, and separately record the time at which each image is acquired; determine corresponding actions performed by the robot at each moment; and identify corresponding moments Whether there is content in the image that is consistent with the action performed by the robot at the corresponding time; determining the visible area memory when there is a number of images of the content that is consistent with the action performed by the robot at the corresponding time exceeds a second predetermined threshold On the mirror.
  • the mirror determining module is specifically configured to acquire pre-stored position information, where the position information includes a position of the mirror; and trigger a predetermined distance that the robot moves to the position.
  • the apparatus for establishing a three-dimensional model of the robot further includes: a second predetermined condition determining module, configured to determine that the mirror meets the second predetermined condition.
  • satisfying the second predetermined condition may specifically include: an angle between the mirror surface and a body upright axis of the robot is less than a first predetermined angle.
  • the second predetermined condition determining module may be specifically configured to trigger a predetermined structure of the robot to start rotating with the body upright axis as a central axis; and to rotate the second predetermined angle, perpendicular to the body Transmitting a first ray in a direction of the upright axis or emitting a first ray in a direction perpendicular to the upright axis of the body every first predetermined time; wherein an axis of the predetermined structure is perpendicular to the upright axis of the body; receiving the specular reflection And a second light corresponding to the first light; when the angle between the first light and the second light is less than a third predetermined angle, determining to determine that the mirror meets the second predetermined condition.
  • the image acquisition module may be specifically configured to trigger the image capturing component of the robot to face the mirror surface, and remain relatively stationary with the mirror surface, and the remaining structure rotates with the body upright axis as a central axis;
  • Each rotation of the fourth predetermined angle triggers the image capturing component to capture an image of the robot in the mirror, or triggers the image capturing component to capture an image of the robot in the mirror every second predetermined time.
  • the apparatus for establishing a three-dimensional model of the robot may further include: a receiving module, configured to receive an action instruction and an execution time corresponding to the action instruction; and a state simulation module, configured to perform the action according to the action at a predetermined time before the execution time
  • the instruction and the three-dimensional model simulate a state of the robot when the motion instruction is executed;
  • the obstacle avoidance module is configured to perform an obstacle avoidance operation according to the state.
  • the obstacle avoidance module may be specifically configured to determine, according to the state, whether the robot will have a degree of freedom collision when executing the motion instruction; if yes, according to the state, the degree of freedom between the collisions occurs. The degree of collision modifies the action instruction.
  • the obstacle avoidance module may be specifically configured to determine, according to the state and the spatial relationship of other objects in the visible area of the robot, whether the robot collides with the other object when executing the motion instruction; If so, the action command is modified according to the position of the colliding object.
  • the apparatus for establishing a three-dimensional model of the robot in the embodiment of the present application determines that a mirror surface exists in the visible region when the first predetermined condition is met, and acquires an image of the robot in the mirror surface, and then establishes the image according to the acquired image.
  • the three-dimensional model of the robot is capable of triggering an action of establishing a three-dimensional model of itself through the mirror when the robot detects that the first predetermined condition is satisfied, thereby eliminating the need for maintenance personnel to set up, reducing operational steps.
  • the robot after the three-dimensional model of the robot in the embodiment of the present application is used, after the three-dimensional model is established by the mirror, after receiving the motion instruction, the robot is simulated in advance according to the three-dimensional model and the motion instruction. The state, and the obstacle avoidance operation is performed according to the state, so that the robot can correctly avoid obstacles.
  • an electronic device 400 as shown in FIG. 4 is also provided in the embodiment of the present application.
  • the electronic device 400 includes a processor 401, a memory 402, a communication interface 403, and a bus 404.
  • the processor 401, the memory 402, and the communication interface 403 are connected by a bus 404 and complete communication with each other;
  • the memory stores executable program code;
  • the processor runs a program corresponding to the executable program code by reading executable program code stored in the memory for executing a method for establishing a three-dimensional model of the robot, including: detecting that the model is satisfied When a predetermined condition is met, it is determined that there is a mirror surface in the visible area; an image of the robot in the mirror surface is acquired; and the three-dimensional model of the robot is established according to the acquired image.
  • the electronic device can include a robot.
  • the method for solving the problem in the electronic device is similar to the method provided in Embodiment 1 or 2 of the present application. Therefore, the implementation of the electronic device may refer to the implementation of the method, and the repeated description is not repeated.
  • the electronic device in the embodiment of the present application when it is detected that the first predetermined condition is met, it is determined that a mirror surface exists in the visible area, and an image of the electronic device in the mirror surface is acquired, and then according to the acquired The image establishes a three-dimensional model of the electronic device, so that when the first predetermined condition is met, the action of establishing a three-dimensional model of the mirror by the mirror is triggered, thereby eliminating the need for maintenance personnel to set the operation step.
  • the state at the time of executing the motion command is simulated in advance based on the three-dimensional model and the motion command, and the obstacle avoidance operation is performed according to the state, so that the obstacle avoidance operation can be avoided. barrier.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device. Instructions are provided for implementation in the flowchart The steps of a process or a plurality of processes and/or block diagrams of a function specified in a block or blocks.

Abstract

一种机器人三维模型的建立方法、装置及电子设备。该方法包括：在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；获取该机器人在该镜面中的图像；根据获取的该图像，建立该机器人的三维模型。能够及时建立机器人的三维模型，减少操作。

Description

一种机器人三维模型的建立方法、装置及电子设备 技术领域
本申请涉及机器人技术，特别涉及一种机器人三维模型的建立方法、装置及电子设备。
背景技术
云端机器人是将认知系统放在云里，身体、驱动、传感器放在机器人本体上，通过移动通信将二者连接起来的智能机器人；云端机器人是智能仿人机器人发展的方向。
在现有技术中，很多机器人已经具备多自由度的运动能力。然而，大多数机器人的每个自由度依然各自分立运动，即，缺乏相互的关联判断；从而导致在运动中，尤其是机器人外壳有变化或者有额外负重后，机器人的各个自由度在执行运动时可能会发生碰撞，比如手打到腿，手打到身体，或者手部攥拳攥不上等等。
为解决上述问题，现有技术中采用以下方案：机器人提供3D(3-Dimensions，三维)模型的设置接口，由维护人员根据机器人的当前外形设置该机器人的3D模型，并将该3D模型存储至机器人的控制模块中保存，然后机器人基于该3D模型动态地对机器人各自由度的动作进行控制，以使得机器人可以避开障碍，或者可以避免各自由度上的碰撞。
但是，发明人发现，采用现有技术中的机器人的三维模型的建立方法，若机器人的某个结构外形发生了变化，例如，增加了新的功能模块，更换了新的外壳，或者是机器人有负重时，需要由维护人员通过接口重新设置机器人的三维模型，并更新至控制模块，导致操作繁琐。
发明内容
本申请实施例中提供了一种机器人三维模型的建立方法及装置，用于解决现有技术中机器人三维模型建立时操作繁琐的问题。
根据本申请实施例中的一个方面，提供了一种机器人三维模型的建立方法，包括：在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；获取该机器人在该镜面中的图像；根据获取的该图像，建立该机器人的三维模型。
可选地，满足建立模型的第一预定条件具体包括：该机器人初始化；或者在该机器人未处于提取物体或托举物体状态时，该机器人自身重量增加或减少超过第一预定阈值。
可选地，确定可视区域内存在镜面，具体包括：获取该机器人可视区域内的一个或多个图像，并分别记录获取各图像的时刻；确定在各时刻该机器人执行的相应动作；识别相应时刻的图像中是否存在与该机器人在该相应时刻执行的动作一致的内容；当存在与该机器人在该相应时刻执行的动作一致的内容的图像数量超过第二预定阈值时，则确定可视区域内存在镜面。
可选地，确定可视区域内存在镜面，具体包括：获取预先存储的位置信息，该位置信息包括镜面的位置；触发该机器人移动至该位置的预定距离。
可选地，在确定可视区域内存在镜面之后，在获取该机器人在该镜面中的图像之前，还包括：确定该镜面满足第二预定条件。
可选地，满足第二预定条件具体包括：该镜面与该机器人的身体直立轴之间的夹角小于第一预定角度。
可选地，确定该镜面满足第二预定条件具体包括：触发该机器人的预定结构以该身体直立轴为中心轴，开始转动；在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线，或者每隔第一预定时间沿垂直于所述身体直立轴的方向发射第一光线；其中，该预定结构的轴线垂直于该身体直立轴；接收经该镜面反射的、与该第一光线对应的第二光线；在该第一光线和该第二光线之间的夹角小于第三预定角度时，判断确定该镜面满 足第二预定条件。
可选地，获取该机器人在该镜面中的图像具体包括：触发所述机器人的图像拍摄组件正对所述镜面，且与所述镜面保持相对静止，其余结构以身体直立轴为中心轴转动；在转动时，每转动第四预定角度触发该图像拍摄组件拍摄该机器人在该镜面中的图像，或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
可选地，在根据获取的该图像，建立该机器人的三维模型之后，还包括：接收动作指令和该动作指令对应的执行时刻；在该执行时刻之前的预定时刻，根据该动作指令和该三维模型，模拟该机器人在执行该动作指令时的状态；根据该状态，执行避障操作。
可选地，根据该状态，执行避障操作，具体包括：根据该状态，判断在执行该动作指令时，该机器人是否会发生自由度碰撞；如果是，则根据该状态中，发生碰撞的自由度之间的碰撞程度修改动作指令。
可选地，根据该状态，执行避障操作，具体包括：根据该状态和预先获取的、该机器人可视区域内的其他物体的空间关系，判断该机器人在执行该动作指令时，是否与该其他物体碰撞；如果是，则根据发生碰撞物体的位置修改动作指令。
根据本申请实施例中的另一个方面，还提供了一种机器人三维模型的建立装置，包括：镜面确定模块，用于在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；图像获取模块，用于获取该机器人在该镜面中的图像；三维模型建立模块，用于根据获取的该图像，建立该机器人的三维模型。
可选地，满足建立模型的第一预定条件具体包括：该机器人初始化；或者在该机器人未处于提取物体或托举物体状态时，该机器人自身重量增加或减少超过第一预定阈值。
可选地，镜面确定模块，具体用于获取该机器人可视区域内的一个或多 个图像，并分别记录获取各图像的时刻；确定在各时刻该机器人执行的相应动作；识别相应时刻的图像中是否存在与该机器人在该相应时刻执行的动作一致的内容；当存在与该机器人在该相应时刻执行的动作一致的内容的图像数量超过第二预定阈值时，则确定可视区域内存在镜面。
可选地，镜面确定模块，具体用于获取预先存储的位置信息，该位置信息包括镜面的位置；触发该机器人移动至该位置的预定距离。
可选地，该机器人三维模型的建立装置还包括：第二预定条件确定模块，用于确定该镜面满足第二预定条件。
可选地，满足第二预定条件具体包括：该镜面与该机器人的身体直立轴之间的夹角小于第一预定角度。
可选地，第二预定条件确定模块具体用于触发该机器人的预定结构以该身体直立轴为中心轴，开始转动；在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线，或者每隔第一预定时间沿垂直于所述身体直立轴的方向发射第一光线；其中，该预定结构的轴线垂直于该身体直立轴；接收经该镜面反射的、与该第一光线对应的第二光线；在该第一光线和该第二光线之间的夹角小于第三预定角度时，判断确定该镜面满足第二预定条件。
可选地，图像获取模块具体用于触发所述机器人的图像拍摄组件正对所述镜面，且与所述镜面保持相对静止，其余结构以身体直立轴为中心轴转动；在转动时，每转动第四预定角度触发该图像拍摄组件拍摄该机器人在该镜面中的图像，或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
可选地，机器人三维模型的建立装置还包括：接收模块，用于接收动作指令和该动作指令对应的执行时刻；状态模拟模块，用于在该执行时刻之前的预定时刻，根据该动作指令和该三维模型，模拟该机器人在执行该动作指令时的状态；避障模块，用于根据该状态，执行避障操作。
可选地，避障模块具体用于根据该状态，判断在执行该动作指令时，该机器人是否会发生自由度碰撞；如果是，则根据该状态中，发生碰撞的自由度之间的碰撞程度修改动作指令。
可选地，避障模块具体用于根据该状态和预先获取的、该机器人可视区域内的其他物体的空间关系，判断该机器人在执行该动作指令时，是否与该其他物体碰撞；如果是，则根据发生碰撞物体的位置修改动作指令。
采用本申请实施例中的机器人三维模型的建立方法、装置及电子设备，在检测到满足第一预定条件时，进一步确定可视区域内存在镜面，并获取该机器人在镜面中的图像，然后根据获取的该图像，建立所述机器人的三维模型，从而能够在机器人在检测到满足第一预定条件时，触发通过镜面建立自身的三维模型的动作，从而无需维护人员设置，减少了操作步骤。
附图说明
此处所说明的附图用来提供对本申请的进一步理解，构成本申请的一部分，本申请的示意性实施例及其说明用于解释本申请，并不构成对本申请的不当限定。在附图中：
图1中示出了根据本申请实施例一的机器人三维模型的建立方法流程图；
图2中示出了根据本申请实施二的机器人三维模型的建立方法的流程图；
图3中示出了根据本申请实施例三所示机器人三维模型的建立装置的结构示意图；
图4中示出了根据本申请实施例四所示的电子设备的结构示意图。
具体实施方式
在实现本申请的过程中，发明人发现，采用现有技术中的机器人的三维模型的建立方法，若机器人的某个结构外形发生了变化，例如，增加了新的功能模块，更换了新的外壳，或者是机器人有负重时，需要由维护人员通过该接口重新设置机器人的三维模型，并更新至控制模块，导致操作繁琐。
另外，发明人还发现，机器人虽然在获取外界信息方面取得了长足的进步，但机器人缺乏了解自身外形变化的能力，也就导致了在自身尺寸发生变化后，仍然按照原有尺寸进行判断，从而不能正确避障，例如，再次发生手打到脚、出门时卡在门口、碰倒花瓶等等。
针对上述问题，本申请实施例中提供了一种机器人三维模型的建立方法及装置，在检测到满足第一预定条件时，确定可视区域内存在镜面，并获取该机器人在镜面中的图像，然后根据获取的该图像，建立所述机器人的三维模型，从而能够在机器人在检测到满足第一预定条件时，触发通过镜面建立自身的三维模型的动作，从而无需维护人员设置，减少了操作步骤。
另外，在通过镜面建立自身的三维模型之后，在接收到动作指令后，根据该三维模型和动作指令提前模拟机器人在执行该动作指令时的状态，并根据该状态执行避障操作，从而使得机器人可以正确避障。。
为了使本申请实施例中的技术方案及优点更加清楚明白，以下结合附图对本申请的示例性实施例进行进一步详细的说明，显然，所描述的实施例仅是本申请的一部分实施例，而不是所有实施例的穷举。需要说明的是，在不冲突的情况下，本申请中的实施例及实施例中的特征可以相互组合。
在具体实施时，本申请中的机器人三维模型的建立方法可以通过安装在机器人头部或其他可旋转部位的3D建模模组中实现，也可以由机器人的其他模组中的多个模块结合实现，本申请对此不作限制。
本领域技术人员应当理解，在本申请实施例中的，“机器人看到……”、“在机器人的可视区域内”等词语，均是指相关的内容处于机器人的图像拍摄组件能够获取到图像的区域内。
实施例一
图1中示出了根据本申请实施例一的机器人三维模型的建立方法流程图。如图1所示，包括以下步骤：
S101，检测是否满足建立模型的第一预定条件；如果满足，则执行S102； 如果不满足，则结束。
在具体实施时，可以是在机器人初始化时判断满足建立模型的第一预定条件。具体地，可以在机器人读取初始化数据时，判断机器人处于初始化状态，从而判断满足建立模型的第一预定条件；也可以在机器人初始化完毕时，判断满足建立模型的第一预定条件。具体地，该初始化可以是在机器人组装完成后，首次上电时执行的初始化流程，也可以是在机器人组装运行后，通过返回出厂设置时重新执行的初始化流程。
在具体实施时，还可以在该机器人未处于提取物体或托举物体状态时，机器人自身重量增加或减少超过第一预定阈值时，判断满足该第一预定条件。具体地，可以通过机器人手部或上肢的压力传感器获取的数据来判断该机器人是否处于提取物体或托举物体状态；同时可以通过机器人脚部的压力传感器获取的数据来判断机器人自身重量增加或减少的数值。具体地，该第一预定阈值可以是预先设置的一个重量值，例如，0.5kg等；具体地，该第一预定阈值可以是由3D建模模组的开发人员根据机器人的各功能模块的平均重量设置的数值，也可以是由机器人的维护人员根据该机器人的常用可替换外壳、功能模块等的重量变化值设置的数值等，本申请对此不作限制。
例如，该第一预定阈值为1kg；在某一时刻，机器人的脚部压力传感器检测到机器人的自身重量增加了2kg，此时可以进一步获取机器人的手部或上肢的压力传感器传回的数据；假如此时机器的手部压力传感器传回的数据显示机器人托举了2kg的物体，则判断此时未满足第一预定条件；假如此时机器人的手部压力传感器传回的数据为0，则判断出2kg大于1kg；确定满足第一预定条件。
在具体实施时，机器人是否处于提取物体或托举物体的状态判断，和机器人自身重量增加或减少是否超过第一预定阈值的判断步骤之间没有严格的时序关系。即，可以先判断机器人是否处于提取物体或托举物体的状态，并且在未处于提取物体或托举物体的状态时，才进一步判断机器人自身重量增 加或减少是否超过第一预定阈值；也可以先判断机器人自身重量增加或减少是否超过第一预定阈值，并且在自身重量增加或减少超过第一预定阈值时，才进一步判断机器人是否处于提取物体或托举物体的状态；还可以两个步骤同时进行；本申请对此均不作限制。
S102，确定可视区域内是否存在镜面；如果存在，则执行S103，如果不存在，则结束。
在具体实施时，机器人的可视区域可以是指机器人的图像拍摄组件，例如，摄像头可以获取图像的范围。
在第一种具体实施方式中，机器人可以通过图像识别的方法发现镜面。具体地，机器人可以通过例如摄像头的图像拍摄组件获取机器人可视区域内的一个或多个图像，并分别记录获取各图像的时刻；确定在各时刻所述机器人执行的相应动作；识别相应时刻的图像中是否存在与所述机器人在所述相应时刻执行的动作一致的内容；当存在与所述机器人在所述相应时刻执行的动作一致的内容的图像数量超过第二预定阈值时，则确定可视区域内存在镜面。具体地，该第二预定阈值可以是1，也可以是如2，3，5等预定阈值。
在具体实施时，当在图像中发现与自身动作完全实时一致的机器人形时，首先可以判断此方向可能存在镜面；在判断出该方向可能存在镜面时，还可以进一步执行特定动作，并进一步获取该方向上的图像，并识别图像中是否存在与自身动作完全实时一致的内容，以进一步判断此方向上拍摄到的完全实时一致的内容为自身的镜像，而非另一个机器人。
在具体实施时，机器人在检测到满足第一预定条件时，可以中断当前执行的运行指令，直接在当前可视区域内获取图像以判断是否存在镜面；也可以不中断当前执行的运行指令，而在正常运动中或者正常执行任务中根据获取的图像来判断是否存在镜面。
在第二种具体实施方式中，还可以通过读取预先存储的信息来发现镜面。具体地，可以读取预先存储的位置信息，该位置信息包括镜面的位置；并触 发机器人移动至该位置的预定距离。具体地，该预定距离可以是使机器人能够在镜面中看到自身的全身像的距离，也可以是机器人能够在镜面中看到自身特定局部的距离；具体地，该特定局部可以包括根据经验设置的、机器人外形变化的身体局部。该预定距离可以根据机器人的身高、镜面的高度等利用镜面成像原理计算得到；本申请在此不赘述。
在具体实施时，该位置信息可以有多个，机器人可以选择离自己最近的镜面；或者机器人可以根据下一运行指令中的目的地，选择当前位置、镜面、和目的地之间的最佳路线。
具体地，该预先存储的位置信息可以是由维护人员预先设置的位置，例如，在客厅大门左侧、卧室衣橱门上等；也可以是机器人自动记录的镜面的位置等；本申请对此不作限制。
在具体实施时，还可以结合上述两种发现镜面的方式。具体地，可以在当前可视区域或者在预定时长的正常运动中，均未发现镜面时，读取预先存储的镜面的位置信息；然后触发机器人移动至该位置的预定距离等。
S103，判断镜面与机器人的身体直立轴之间的夹角是否小于第一预定角度；如果是，则执行S104-S105；如果否，则结束。
为避免机器人在镜中看到的自己存在视角误差，造成建模不准确，在找到镜面之后，可以进一步判断镜面与机器人的身体直立轴之间的夹角是否小于第一预定角度。即，可以挑选与机器人的身体直立轴相对平行的镜面来获取建模图像。具体地，可以向镜面发射第一光线，并根据接收到的反射光线与第一光线之间的夹角来判断该镜面是否与机器人身体直立轴平行。具体地，该第一预定角度可以是1度、2度等较小的角度。
在具体实施时，可以通过以下方式判断镜面与机器人的身体直立轴之间的夹角是否小于第一预定角度：触发机器人的预定结构以身体直立轴为中心轴，开始转动；在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线，或者每隔第一预定时间沿垂直于所述身体直立轴的方 向发射第一光线；其中，预定结构的轴线垂直于身体直立轴；接收经镜面反射的、与第一光线对应的第二光线；在第一光线和所述第二光线之间的夹角小于第三预定角度时，判断镜面与机器人的身体直立轴之间的夹角小于第一预定角度。
具体地，该第一光线可以是红外线、激光等光线。该预定结构可以是机器人的头部或其他可以围绕身体直立轴的轴线360度旋转的结构，本申请对此不作限制。具体地，该第二预定角度可以是1度，2度等较小的角度。具体地，该第三预定角度可以是1度、2度等较小的角度。具体地，该第一预定时间可以是例如0.1秒、1秒等时长。
例如，机器人在发现镜面后，可以使自身保持身体与地面垂直标的准直立动作；然后以身体直立轴为中心轴，360度转动3D建模模组所在的结构，例如，头部，并每转1度向外沿垂直于身体直立轴的方向发射红外线；在转动过程中，共发射出35道第一光线，收到2道第二光线，其中有一道光线的入射光线与反射光线的夹角小于2度；此时，判断镜面与机器人的身体直立轴相对平行。
在判断镜面与机器人的身体直立轴之间的夹角小于第一预定角度后，确定该镜面满足三维建模的要求，此时，可以进一步记录该镜面的位置、方向等信息，以便机器人下次重新建模时使用。
S104，获取机器人在该镜面中的图像。
在具体实施时，可以通过以下方式获取机器人在该镜面中的图像：触发机器人的图像拍摄组件正对镜面，且与镜面保持相对静止，其余结构以身体直立轴为中心轴转动；在转动时，每转动第四预定角度触发图像拍摄组件拍摄机器人在镜面中的图像，或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
在具体实施时，如果机器人的图像拍摄组件的朝向与光线发射的方向一致，例如，均位于头部且呈上下或左右分布。则在S104步骤中，可以在接收 收到入射光线与反射光线的夹角小于的第三预定角度时，判断此时图像拍摄组件正对该镜面，触发该预定结构停止转动，并获取机器人在该镜面中的图像。
在具体实施时，如果机器人的图像拍摄组件的朝向与光线发射的方向不一致，在获取机器人在该镜面中的图像之前，还可以进一步调整机器人的图像拍摄组件的方向，以使图像拍摄组件正对镜面。例如，图像拍摄组件的朝向与光线发射方向夹角为180度；且在光线向左发射时，接收收到与入射光线的夹角小于的第三预定角度与反射光线，此时，可以停止预定结构的转动，并将与机器人的图像拍摄组件旋转180度。
在具体实施时，该镜面可以是机器人能够看到全身像的镜面，也可以是也可以是机器人能够看到自身特定局部的镜面；具体地，该特定局部可以包括根据经验设置的、机器人外形变化或者负重的身体局部；本申请对此不作限制。
在具体实施时，如果3D建模模组的图像拍摄组件位于机器人头部，则机器人头部静止，脖子以下沿轴转动；在转动过程中，图像拍摄组件可以实时拍摄，即，每隔预定时长，如，0.1秒拍摄一张；也可以是其余结构在每转动一预定角度，例如，1度时拍摄一张。根据自身建模算法的不同，机器人可以活动特定的自由度或者触发某一自由度摆出特定的造型；具体地，该特定的自由度可以是根据经验值设置的、机器人外形变化或者负重的身体局部。
S105，根据获取的图像，建立机器人的三维模型。
在具体实施时，可以采用现有技术根据获取的二维图像建立机器人的三维模型，例如，基于图像的建模和绘制(Image-Based Modeling and Rendering，IBMR)技术。
基于图像的建模的主要目的是由二维图像恢复景物的三维几何结构。IBMR技术具有许多独特的优点，比如建模更快更方便，可以获得很高的绘制速度和高度的真实感等。与传统的利用建模软件或者三维扫描仪得到立体模 型的方法相比，基于图像建模的方法成本低廉，真实感强，自动化程度高。很多公司都已经研发出了此类算法，如国内3D Cloud以云端形式运行，只要将照片上传至云端，即可全自动生成三维模型。如迪斯尼苏黎世研究团队开发出了一种新程序，使用数百张二维摄影照片和一种特殊设计的算法，给电影、电视和游戏做复杂且真实的3D建模，也可以制作或者打印高精度的模型。
IBMR技术的具体算法流程可以参见多篇文献，例如，2010年8月发表于《贵州大学学报》第27卷第4期，题为《基于图像的三维重建流程及实现》的期刊论文；或者参见2015年1月发表于《计算机工程与设计》第36卷第1期，题为《基于图像识别的建筑物三维重建》的期刊论文等。
应当理解，可以采用常有多种从二维图像重建三维图像的算法实现步骤105的实施，上述IBMR仅用于示例的目的而示出，并不用于限制本申请。
至此，根据本申请实施例的机器人的三维模型的建立方法已执行完毕。
采用本申请实施例中的机器人三维模型的建立方法，在检测到满足第一预定条件时，确定可视区域内存在镜面，并获取该机器人在镜面中的图像，然后根据获取的该图像，建立所述机器人的三维模型，从而能够在机器人在检测到满足第一预定条件时，触发通过镜面建立自身的三维模型的动作，从而无需维护人员设置，减少了操作步骤。
在上述实施例一中，对本申请实施例中的机器人三维模型的建立方法的实施进行了详细阐述。在具体实施时，在完成机器人三维模型的建立之后，机器人还可以根据新建立的三维模型执行相应的避障动作；下面将结合另一实施例进行描述。
实施例二
图2中示出了根据本申请实施二的机器人三维模型的建立方法的流程图。如图2所示，根据本申请实施二的机器人三维模型的建立方法包括以下步骤：
S201，检测是否满足建立模型的第一预定条件；如果满足，则执行S202；如果不满足，则执行S206。
在具体实施时，S201的实施可以参见本申请实施例一中S101的实施，重复之处不再赘述。
S202，确定可视区域内是否存在镜面；如果存在，则执行S203，如果不存在，则执行S206。
在具体实施时，S202的实施可以参见本申请实施例一中S102的实施，重复之处不再赘述。
S203，判断镜面与机器人的身体直立轴之间的夹角是否小于第一预定角度；如果是，则执行S204-S206；如果否，则执行S206。
在具体实施时，S203的实施可以参见本申请实施例一中S103的实施，重复之处不再赘述。
S204，获取机器人在该镜面中的图像。
在具体实施时，S204的实施可以参见本申请实施例一中S104的实施，重复之处不再赘述。
S205，根据获取的图像，建立机器人的三维模型。
在具体实施时，S205的实施可以参见本申请实施例一中S105的实施，重复之处不再赘述。
S206，根据机器人的三维模型避障。
在具体实施时，可以根据机器人存储的三维模型避障。该三维模型可以是原有的三维模型；也可以是步骤201-205中建立的三维模型。
在S206的第一种具体实施方式中，可以采用以下方式避障：接收动作指令和动作指令对应的执行时刻，并在该执行时刻之前的预定时刻，根据动作指令和三维模型，模拟机器人在执行动作指令时的状态；然后判断在执行所述动作指令时，所述机器人是否会发生自由度碰撞；如果是，则根据所述状态中，发生碰撞的自由度之间的碰撞程度修改动作指令，以使得发生碰撞的 部分变为刚刚挨上或者留有微小的空隙不接触。
具体地，该预定时刻可以为根据经验值设置的一时间长度，例如，在执行时刻前的2分钟、1分钟、10秒等。
具体地，可以采用彷真的方式根据动作指令和三维模型，模拟机器人在执行动作指令时的状态。例如，动作指令为：“将右手向前抬起10厘米”；则触发仿真的三维模型执行“将右手向前抬起10厘米”的动作。
具体地，如果在执行过程中，出现某些自由度碰撞的情况，例如，可能会出现由于腰部外形变化导致右手打到腰部的情况；此时可以根据碰撞的程度修改动作指令，例如，取消该动作指令或者改为“将右手向前抬起5厘米”等。
在S206的第二种具体实施方式中，可以采用以下方式避障：接收动作指令和动作指令对应的执行时刻，并在该执行时刻之前的预定时刻，根据动作指令和三维模型，模拟机器人在执行动作指令时的状态；然后根据该状态和预先获取的、所述机器人可视区域内的其他物体的空间关系，判断所述机器人在执行所述动作指令时，是否与所述其他物体碰撞；如果是，则根据发生碰撞物体的位置修改动作指令。
具体地，可以结合可视区域内的其他物体的空间关系，判定机器人是否可能会和物体发生碰撞，以及通过性上是否有问题，如果发现则可以修改自身运动方式或运动轨迹，避免可能的碰撞。
例如，动作指令是“向前移动20厘米”，而在机器人的前面10厘米为门，门的尺寸为高2米，宽1米，此时机器人由于外形变化，宽度超过1米；经模拟发现会出现卡在门外进不去的情况；此时可以修改动作指令，例如，取消该动作指令或者改为“向右转90度后，向右移动20厘米”等。
采用本申请实施例中的机器人三维模型的建立方法，在通过镜面建立自身的三维模型之后，在接收到动作指令后，根据该三维模型和动作指令提前模拟机器人在执行该动作指令时的状态，并根据该状态执行避障操作，从而 使得机器人可以正确避障。
基于同一发明构思，本申请实施例中还提供了一种机器人三维模型的建立装置，由于该装置解决问题的原理与本申请实施例一和二所提供的方法相似，因此该装置的实施可以参见方法的实施，重复之处不再赘述。
实施例三
图3为本申请实施例三所示机器人三维模型的建立装置的结构示意图。在具体实施时，本申请实施例三所示的机器人三维模型的建立装置中的多个模块可以集成在一专用3D建模模组中，也可以由分散在机器人各结构中，还可以独立于机器人提供，本申请对此不作限制。
如图3所示，本申请实施例三所示机器人三维模型的建立装置300，包括：镜面确定模块301，用于在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；图像获取模块302，用于获取该机器人在该镜面中的图像；三维模型建立模块303，用于根据获取的该图像，建立该机器人的三维模型。
在具体实施时，图像获取模块302可以包括机器人现有的图像拍摄组件，也可以包括新增的图像拍摄组件，本申请对此不作限制。
在具体实施时，满足建立模型的第一预定条件具体可以包括：该机器人初始化；或者在该机器人未处于提取物体或托举物体状态时，该机器人自身重量增加或减少超过第一预定阈值。
在具体实施时，镜面确定模块，具体可以用于获取该机器人可视区域内的一个或多个图像，并分别记录获取各图像的时刻；确定在各时刻该机器人执行的相应动作；识别相应时刻的图像中是否存在与该机器人在该相应时刻执行的动作一致的内容；当存在与该机器人在该相应时刻执行的动作一致的内容的图像数量超过第二预定阈值时，则确定可视区域内存在镜面。
在具体实施时，镜面确定模块，具体用于获取预先存储的位置信息，该位置信息包括镜面的位置；触发该机器人移动至该位置的预定距离。
在具体实施时，该机器人三维模型的建立装置还包括：第二预定条件确定模块，用于确定该镜面满足第二预定条件。
在具体实施时，满足第二预定条件具体可以包括：该镜面与该机器人的身体直立轴之间的夹角小于第一预定角度。
在具体实施时，第二预定条件确定模块具体可以用于触发该机器人的预定结构以该身体直立轴为中心轴，开始转动；在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线，或者每隔第一预定时间沿垂直于所述身体直立轴的方向发射第一光线；其中，该预定结构的轴线垂直于该身体直立轴；接收经该镜面反射的、与该第一光线对应的第二光线；在该第一光线和该第二光线之间的夹角小于第三预定角度时，判断确定该镜面满足第二预定条件。
在具体实施时，图像获取模块具体可以用于触发所述机器人的图像拍摄组件正对所述镜面，且与所述镜面保持相对静止，其余结构以身体直立轴为中心轴转动；在转动时，每转动第四预定角度触发该图像拍摄组件拍摄该机器人在该镜面中的图像，或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
在具体实施时，机器人三维模型的建立装置还可以包括：接收模块，用于接收动作指令和该动作指令对应的执行时刻；状态模拟模块，用于在该执行时刻之前的预定时刻，根据该动作指令和该三维模型，模拟该机器人在执行该动作指令时的状态；避障模块，用于根据该状态，执行避障操作。
在具体实施时，避障模块具体可以用于根据该状态，判断在执行该动作指令时，该机器人是否会发生自由度碰撞；如果是，则根据该状态中，发生碰撞的自由度之间的碰撞程度修改动作指令。
在具体实施时，避障模块具体可以用于根据该状态和预先获取的、该机器人可视区域内的其他物体的空间关系，判断该机器人在执行该动作指令时，是否与该其他物体碰撞；如果是，则根据发生碰撞物体的位置修改动作指令。
采用本申请实施例中的机器人三维模型的建立装置，在检测到满足第一预定条件时，确定可视区域内存在镜面，并获取该机器人在镜面中的图像，然后根据获取的该图像，建立所述机器人的三维模型，从而能够在机器人在检测到满足第一预定条件时，触发通过镜面建立自身的三维模型的动作，从而无需维护人员设置，减少了操作步骤。
另外，采用本申请实施例中的机器人三维模型的建立装置，在通过镜面建立自身的三维模型之后，在接收到动作指令后，根据该三维模型和动作指令提前模拟机器人在执行该动作指令时的状态，并根据该状态执行避障操作，从而使得机器人可以正确避障。
实施例四
基于同一发明构思，本申请实施例中还提供了如图4所示的一种电子设备400。
该电子设备400包括：处理器401、存储器402、通信接口403、总线404；该处理器401、存储器402、通信接口403通过总线404连接并完成相互间的通信；该存储器存储可执行程序代码；该处理器通过读取该存储器中存储的可执行程序代码来运行与该可执行程序代码对应的程序，以用于执行一种机器人三维模型的建立方法，包括：在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；获取该机器人在该镜面中的图像；根据获取的该图像，建立该机器人的三维模型。
在具体实施时，该电子设备可以包括机器人。
由于该电子设备中运行的方法解决问题的原理与本申请实施例一或二所提供的方法相似，因此该电子设备的实施可以参见方法的实施，重复之处不再赘述。
采用本申请实施例中的电子设备，在检测到满足第一预定条件时，确定可视区域内存在镜面，并获取该电子设备在镜面中的图像，然后根据获取的 该图像，建立所述电子设备的三维模型，从而能够在检测到满足第一预定条件时，触发通过镜面建立自身的三维模型的动作，从而无需维护人员设置，减少了操作步骤。另外，在通过镜面建立自身的三维模型之后，在接收到动作指令后，根据该三维模型和动作指令提前模拟在执行该动作指令时的状态，并根据该状态执行避障操作，从而可以正确避障。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图 一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本申请的优选实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例作出另外的变更和修改。所以，所附权利要求意欲解释为包括优选实施例以及落入本申请范围的所有变更和修改。
显然，本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的精神和范围。这样，倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内，则本申请也意图包含这些改动和变型在内。

Claims (23)

  1. 一种机器人三维模型的建立方法，其特征在于，包括：
    在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；
    获取所述机器人在所述镜面中的图像；
    根据获取的所述图像，建立所述机器人的三维模型。
  2. 根据权利要求1所述的方法，其特征在于，满足建立模型的第一预定条件具体包括：
    所述机器人初始化；或者
    在所述机器人未处于提取物体或托举物体状态时，所述机器人自身重量增加或减少超过第一预定阈值。
  3. 根据权利要求1所述的方法，其特征在于，确定可视区域内存在镜面，具体包括：
    获取所述机器人可视区域内的一个或多个图像，并分别记录获取各图像的时刻；
    确定在各时刻所述机器人执行的相应动作；
    识别相应时刻的图像中是否存在与所述机器人在所述相应时刻执行的动作一致的内容；
    当存在与所述机器人在所述相应时刻执行的动作一致的内容的图像数量超过第二预定阈值时，则确定可视区域内存在镜面。
  4. 根据权利要求1所述的方法，其特征在于，确定可视区域内存在镜面，具体包括：
    获取预先存储的位置信息，所述位置信息包括镜面的位置；
    触发所述机器人移动至所述位置的预定距离。
  5. 根据权利要求1所述的方法，其特征在于，在确定可视区域内存在镜面之后，在获取所述机器人在所述镜面中的图像之前，还包括：
    确定所述镜面满足第二预定条件。
  6. 根据权利要求5所述的方法，其特征在于，满足第二预定条件具体包括：
    所述镜面与所述机器人的身体直立轴之间的夹角小于第一预定角度。
  7. 根据权利要求6所述的方法，其特征在于，确定所述镜面满足第二预定条件具体包括：
    触发所述机器人的预定结构以所述身体直立轴为中心轴，开始转动；
    在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线；或者每隔第一预定时间沿垂直于所述身体直立轴的方向发射第一光线；
    接收经所述镜面反射的、与所述第一光线对应的第二光线；
    在所述第一光线和所述第二光线之间的夹角小于第三预定角度时，判断确定所述镜面满足第二预定条件。
  8. 根据权利要求1所述的方法，其特征在于，获取所述机器人在所述镜面中的图像具体包括：
    触发所述机器人的图像拍摄组件正对所述镜面，且与所述镜面保持相对静止，其余结构以身体直立轴为中心轴转动；
    在转动时，每转动第四预定角度触发所述图像拍摄组件拍摄所述机器人在所述镜面中的图像；或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
  9. 根据权利要求1所述的方法，其特征在于，在根据获取的所述图像，建立所述机器人的三维模型之后，还包括：
    接收动作指令和所述动作指令对应的执行时刻；
    在所述执行时刻之前的预定时刻，根据所述动作指令和所述三维模型，模拟所述机器人在执行所述动作指令时的状态；
    根据所述状态，执行避障操作。
  10. 根据权利要求9所述的方法，其特征在于，根据所述状态，执行避障 操作，具体包括：
    根据所述状态，判断在执行所述动作指令时，所述机器人是否会发生自由度碰撞；
    如果是，则根据所述状态中，发生碰撞的自由度之间的碰撞程度修改动作指令。
  11. 根据权利要求9所述的方法，其特征在于，根据所述状态，执行避障操作，具体包括：
    根据所述状态和预先获取的、所述机器人可视区域内的其他物体的空间关系，判断所述机器人在执行所述动作指令时，是否与所述其他物体碰撞；
    如果是，则根据发生碰撞物体的位置修改动作指令。
  12. 一种机器人三维模型的建立装置，其特征在于，包括：
    镜面确定模块，用于在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；
    图像获取模块，用于获取所述机器人在所述镜面中的图像；
    三维模型建立模块，用于根据获取的所述图像，建立所述机器人的三维模型。
  13. 根据权利要求12所述的装置，其特征在于，所述满足建立模型的第一预定条件具体包括：
    所述机器人初始化；或者
    在所述机器人未处于提取物体或托举物体状态时，所述机器人自身重量增加或减少超过第一预定阈值。
  14. 根据权利要求12所述的装置，其特征在于，镜面确定模块，具体用于获取所述机器人可视区域内的一个或多个图像，并分别记录获取各图像的时刻；确定在各时刻所述机器人执行的相应动作；识别相应时刻的图像中是否存在与所述机器人在所述相应时刻执行的动作一致的内容；当存在与所述机器人在所述相应时刻执行的动作一致的内容的图像数量超过第二预定阈值 时，则确定可视区域内存在镜面。
  15. 根据权利要求12所述的装置，其特征在于，镜面确定模块，具体用于获取预先存储的位置信息，所述位置信息包括镜面的位置；触发所述机器人移动至所述位置的预定距离。
  16. 根据权利要求12所述的装置，其特征在于，还包括：
    第二预定条件确定模块，用于确定所述镜面满足第二预定条件。
  17. 根据权利要求16所述的装置，其特征在于，满足第二预定条件具体包括：
    所述镜面与所述机器人的身体直立轴之间的夹角小于第一预定角度。
  18. 根据权利要求17所述的装置，其特征在于，第二预定条件确定模块具体用于触发所述机器人的预定结构以所述身体直立轴为中心轴，开始转动；在转动时，每转动第二预定角度，沿垂直于所述身体直立轴的方向发射第一光线，或者每隔第一预定时间沿垂直于所述身体直立轴的方向发射第一光线；其中，所述预定结构的轴线垂直于所述身体直立轴；接收经所述镜面反射的、与所述第一光线对应的第二光线；在所述第一光线和所述第二光线之间的夹角小于第三预定角度时，判断确定所述镜面满足第二预定条件。
  19. 根据权利要求12所述的装置，其特征在于，图像获取模块具体用于触发所述机器人的图像拍摄组件正对所述镜面，且与所述镜面保持相对静止，其余结构以身体直立轴为中心轴转动；在转动时，每转动第四预定角度触发所述图像拍摄组件拍摄所述机器人在所述镜面中的图像，或者触发所述图像拍摄组件每隔第二预定时间拍摄所述机器人在所述镜面中的图像。
  20. 根据权利要求12所述的装置，其特征在于，还包括：
    接收模块，用于接收动作指令和所述动作指令对应的执行时刻；
    状态模拟模块，用于在所述执行时刻之前的预定时刻，根据所述动作指令和所述三维模型，模拟所述机器人在执行所述动作指令时的状态；
    避障模块，用于根据所述状态，执行避障操作。
  21. 根据权利要求20所述的装置，其特征在于，避障模块具体用于根据所述状态，判断在执行所述动作指令时，所述机器人是否会发生自由度碰撞；如果是，则根据所述状态中，发生碰撞的自由度之间的碰撞程度修改动作指令。
  22. 根据权利要求20所述的装置，其特征在于，避障模块具体用于根据所述状态和预先获取的、所述机器人可视区域内的其他物体的空间关系，判断所述机器人在执行所述动作指令时，是否与所述其他物体碰撞；如果是，则根据发生碰撞物体的位置修改动作指令。
  23. 一种电子设备，其特征在于，包括：处理器、存储器、通信接口和总线；所述处理器、所述存储器和所述通信接口通过所述总线连接并完成相互间的通信；所述存储器存储可执行程序代码；所述处理器通过读取所述存储器中存储的可执行程序代码来运行与所述可执行程序代码对应的程序，以用于执行一种机器人三维模型的建立方法；其中，所述机器人三维模型的建立方法，包括：
    在检测到满足建立模型的第一预定条件时，确定可视区域内存在镜面；
    获取所述机器人在所述镜面中的图像；
    根据获取的所述图像，建立所述机器人的三维模型。
PCT/CN2016/080133 2016-04-25 2016-04-25 一种机器人三维模型的建立方法、装置及电子设备 WO2017185208A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201680002792.3A CN107004298B (zh) 2016-04-25 2016-04-25 一种机器人三维模型的建立方法、装置及电子设备
PCT/CN2016/080133 WO2017185208A1 (zh) 2016-04-25 2016-04-25 一种机器人三维模型的建立方法、装置及电子设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2016/080133 WO2017185208A1 (zh) 2016-04-25 2016-04-25 一种机器人三维模型的建立方法、装置及电子设备

Publications (1)

Publication Number Publication Date
WO2017185208A1 true WO2017185208A1 (zh) 2017-11-02

Family

ID=59431655

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/080133 WO2017185208A1 (zh) 2016-04-25 2016-04-25 一种机器人三维模型的建立方法、装置及电子设备

Country Status (2)

Country Link
CN (1) CN107004298B (zh)
WO (1) WO2017185208A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410322A (zh) * 2018-10-23 2019-03-01 北京旷视科技有限公司 三维对象建模方法、装置及电子设备
CN110956580A (zh) * 2019-11-28 2020-04-03 广州华多网络科技有限公司 图像换脸的方法、装置、计算机设备以及存储介质

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108873014A (zh) * 2018-06-28 2018-11-23 北方工业大学 一种基于激光雷达的镜面检测方法及装置
CN111015650A (zh) * 2019-11-18 2020-04-17 安徽机电职业技术学院 一种多点确定目标位置的工业机器人智能视觉系统及方法
CN113879921B (zh) * 2021-10-13 2023-10-10 苏州塔米机器人有限公司 一种机器人进入电梯的控制方法及装置
CN114265399A (zh) * 2021-11-25 2022-04-01 北京顺造科技有限公司 障碍物识别方法及自动清洁设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1840298A (zh) * 2005-04-01 2006-10-04 北京理工大学 基于多信息融合的仿人型机器人作业场景的重构
CN101587329A (zh) * 2009-06-18 2009-11-25 北京理工大学 机器人预测的方法和系统
CN104057453A (zh) * 2013-03-18 2014-09-24 株式会社安川电机 机器人装置以及被加工物的制造方法
CN104944168A (zh) * 2015-05-19 2015-09-30 电子科技大学 一种基于图像三维重构的码垛机器人安全控制方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8432435B2 (en) * 2011-08-10 2013-04-30 Seiko Epson Corporation Ray image modeling for fast catadioptric light field rendering
CN102650886B (zh) * 2012-04-28 2014-03-26 浙江工业大学 基于主动全景视觉传感器的机器人视觉系统
CN102980513B (zh) * 2012-11-02 2016-01-20 浙江工业大学 以物为中心的单目全景立体视觉传感器
CN104567818B (zh) * 2014-12-31 2016-09-28 浙江工业大学 一种便携式全天候主动全景视觉传感器

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1840298A (zh) * 2005-04-01 2006-10-04 北京理工大学 基于多信息融合的仿人型机器人作业场景的重构
CN101587329A (zh) * 2009-06-18 2009-11-25 北京理工大学 机器人预测的方法和系统
CN104057453A (zh) * 2013-03-18 2014-09-24 株式会社安川电机 机器人装置以及被加工物的制造方法
CN104944168A (zh) * 2015-05-19 2015-09-30 电子科技大学 一种基于图像三维重构的码垛机器人安全控制方法

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410322A (zh) * 2018-10-23 2019-03-01 北京旷视科技有限公司 三维对象建模方法、装置及电子设备
CN110956580A (zh) * 2019-11-28 2020-04-03 广州华多网络科技有限公司 图像换脸的方法、装置、计算机设备以及存储介质
CN110956580B (zh) * 2019-11-28 2024-04-16 广州方硅信息技术有限公司 图像换脸的方法、装置、计算机设备以及存储介质

Also Published As

Publication number Publication date
CN107004298B (zh) 2020-11-10
CN107004298A (zh) 2017-08-01

Similar Documents

Publication Publication Date Title
WO2017185208A1 (zh) 一种机器人三维模型的建立方法、装置及电子设备
KR102235003B1 (ko) 충돌 검출, 추정 및 회피
US20230016490A1 (en) Systems and methods for virtual and augmented reality
TWI620627B (zh) 機器人及用於機器人之定位之方法
TWI467494B (zh) 使用深度圖進行移動式攝影機定位
TWI442311B (zh) 在遊戲中使用三維環境模型
CA3073900C (en) Methods and systems for generating detailed datasets of an environment via gameplay
CN102385237B (zh) 基于结构化光和立体视觉的深度相机
JP2021192250A (ja) 単眼カメラを用いたリアルタイム３ｄ捕捉およびライブフィードバックのための方法およびシステム
CN108170137A (zh) 移动机器人及其控制方法和控制系统
KR101810415B1 (ko) 정보 처리 장치, 정보 처리 시스템, 블록 시스템, 및 정보 처리 방법
Peasley et al. Real-time obstacle detection and avoidance in the presence of specular surfaces using an active 3D sensor
JP2018136632A (ja) 形状推定装置
JP6054831B2 (ja) 画像処理装置、画像処理方法及び画像処理プログラム
Diakité et al. First experiments with the tango tablet for indoor scanning
Senthilvel et al. Comparison of handheld devices for 3D reconstruction in construction
CN107462153A (zh) 一种快速建立三维空间模型的方法
TWI413018B (zh) 體積識別方法及系統
Albrektsen Using the Kinect Sensor for Social Robotics
Guo et al. A control system of human-computer interaction based on Kinect somatosensory equipment
Budiharto et al. Ball distance estimation and tracking system of humanoid soccer robot
Callemein et al. Multi-view real-time 3D occupancy map for machine-patient collision avoidance
JP6429240B2 (ja) 形状検出装置、形状検出方法、及びプログラム
RATSAMEE et al. A Comparison of Adaptive View Techniques for Exploratory 3D Drone Teleoperation

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16899722

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 16899722

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 08.01.2019)

122 Ep: pct application non-entry in european phase

Ref document number: 16899722

Country of ref document: EP

Kind code of ref document: A1