WO2020107931A1 - 位姿信息确定方法和装置、视觉点云构建方法和装置 - Google Patents

位姿信息确定方法和装置、视觉点云构建方法和装置 Download PDF

Info

Publication number
WO2020107931A1
WO2020107931A1 PCT/CN2019/099207 CN2019099207W WO2020107931A1 WO 2020107931 A1 WO2020107931 A1 WO 2020107931A1 CN 2019099207 W CN2019099207 W CN 2019099207W WO 2020107931 A1 WO2020107931 A1 WO 2020107931A1
Authority
WO
WIPO (PCT)
Prior art keywords
pose information
translation parameters
relative
acquisition device
acquiring
Prior art date
Application number
PCT/CN2019/099207
Other languages
English (en)
French (fr)
Inventor
颜沁睿
杨帅
Original Assignee
南京人工智能高等研究院有限公司
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 南京人工智能高等研究院有限公司 filed Critical 南京人工智能高等研究院有限公司
Publication of WO2020107931A1 publication Critical patent/WO2020107931A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
    • G01C11/04—Interpretation of pictures
    • G01C11/06—Interpretation of pictures by comparison of two or more pictures of the same area
    • G01C11/12—Interpretation of pictures by comparison of two or more pictures of the same area the pictures being supported in the same relative position as when they were taken
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30244—Camera pose

Definitions

  • the present application relates to the field of computer vision. Specifically, the present application relates to a pose information determination method, a pose information determination apparatus, a visual point cloud construction method and a visual point cloud construction apparatus, an electronic device, and a computer-readable storage medium.
  • the map is the foundation of the unmanned driving field.
  • the monocular camera SLAM due to the uncertainty of the monocular camera's scale, it is impossible to construct a vector map with uniform global scale, and because of the uncertainty of the monocular camera's scale, the monocular SLAM tracks images in multiple frames It is easy to accumulate tracking result errors due to scale drift, which eventually leads to tracking failure.
  • the real three-dimensional scale of the map point is obtained directly at every moment, or the high-precision integrated navigation module (IMU) is integrated, and the real scale linear acceleration integral is directly measured through the IMU module to obtain Pose information of real scale between frames.
  • IMU integrated navigation module
  • the real scale of the inter-frame image can be obtained by using binocular vision technology or the IMU module, due to the high cost of the sensor, high calculation cost, high production work cost, complex calibration, and complicated algorithm, and is affected by the cost of the IMU itself
  • the large size makes the use of visual point clouds a huge obstacle.
  • the embodiments of the present application provide a pose information determination method, a pose information determination device, a visual point cloud construction method and a visual point cloud construction device, an electronic device, and a computer-readable storage medium, with low cost and high accuracy 2. Widely determine the posture information of the image acquisition device.
  • a method for determining pose information includes determining the relative pose information of an image acquisition device when acquiring an image of a current frame relative to when acquiring an image of a previous frame;
  • the image acquisition device acquires a first set of translation parameters during motion between the current frame and the previous frame image; adjusts the relative pose information based on the relative pose information and the first set of translation parameters; and based on the adjusted
  • the relative pose information determines the pose information of the image acquisition device when acquiring the current frame image.
  • a method for constructing a visual point cloud comprising: acquiring the pose information of the image acquisition device through the above pose information determination method; and constructing a vision based on the pose information of the image acquisition device Point cloud.
  • a pose information determination apparatus including a relative pose information determination unit, for determining relative pose information of an image acquisition device when acquiring a current frame image relative to when acquiring a previous frame image ; Relative displacement parameter acquisition unit for determining the first set of translation parameters of the image acquisition device during the acquisition of the current frame and the previous frame image by a sensor with an absolute scale; relative position and posture information adjustment unit for The relative pose information and the first set of translation parameters to adjust the relative pose information; and a pose information determination unit for determining that the image acquisition device is based on the adjusted relative pose information Get the pose information of the current frame image.
  • a visual point cloud construction device which includes a relative pose information determination unit for determining relative pose information of an image acquisition device when acquiring a current frame image relative to when acquiring a previous frame image ; Relative displacement parameter acquisition unit, used to determine the first group of translation parameters of the image acquisition device during the acquisition of the current frame and the previous frame image by the sensor with absolute scale; relative position and posture information adjustment unit, Adjust the relative pose information based on the relative pose information and the first set of translation parameters; a pose information determination unit for determining that the image acquisition device is based on the adjusted relative pose information Pose information when acquiring the current frame image; and a visual point cloud construction unit for constructing a visual point cloud based on the posture information of the image acquisition device when acquiring the current frame image.
  • an electronic device including a processor, and a memory, in which a computer program instruction is stored, the computer program instruction, when executed by the processor, causes the processing
  • the device executes the above-mentioned pose information determination method or the above-mentioned visual point cloud construction method.
  • a computer-readable storage medium on which instructions for executing the above-mentioned pose information determination method or the above-mentioned visual point cloud construction method are stored.
  • the posture information determination method, the posture information determination device, the visual point cloud construction method and the visual point cloud construction device, the electronic device, and the computer-readable storage medium can be adopted Determine the relative pose information, the first set of translation parameters and the adjustment coefficients of the image acquisition device when acquiring the current frame and the previous frame image to obtain more accurate pose information of the image acquisition device when acquiring the current frame image. Therefore, by directly using the scale of the external sensor to scale the translation vector of the posture information of the image acquisition device, more accurate posture information is obtained, which will not affect the algorithm framework due to the change of the sensor configuration, and also reduces the The cost of sensors and the cost of calculation further reduce the difficulty of deploying a monocular vision system.
  • FIG. 1 illustrates a schematic diagram of an application scenario of a method for determining pose information according to an embodiment of the present application
  • FIG. 2 illustrates a flowchart of a method for determining pose information according to an embodiment of the present application.
  • FIG. 3 illustrates a schematic diagram of an apparatus for determining pose information according to an embodiment of the present application.
  • FIG. 4 illustrates a block diagram of an electronic device according to an embodiment of the application.
  • the pose information of the camera and thus the current position of the camera is very important.
  • the cost of obtaining accurate pose information of the camera and the current position is relatively high. Therefore, an improved method for determining pose information is needed to reduce the cost of obtaining accurate pose information of the camera.
  • the basic idea of this application is to propose a pose information determination method, a pose information determination device, a visual point cloud construction method and a visual point cloud construction device, an electronic device, and a computer-readable storage medium, which can By directly using the scale of the external sensor, especially the scalar scale, the translation vector of the image acquisition device is scaled, thereby reducing the cost and making the deployment of the monocular vision system less difficult.
  • the method and device for determining posture information of the present application can obtain more accurate posture information without using high-precision sensors or excessive manual intervention, and then obtain a globally consistent visual point cloud. And thus establish a high-precision vector map, thus reducing the production cost of high-precision maps.
  • FIG. 1 illustrates a schematic diagram of an application scenario of a pose information determination method according to an embodiment of the present application.
  • the vehicle 10 may include an image acquisition device, such as an on-board camera 12, which may be a commonly used monocular camera, binocular camera, or more.
  • an on-board camera 12 may be a commonly used monocular camera, binocular camera, or more.
  • FIG. 1 shows that the in-vehicle camera 12 is installed on the top of the vehicle 10, it should be understood that the in-vehicle camera may also be installed in other positions of the vehicle 10, such as a front portion, a front windshield, and so on.
  • the coordinate system shown in Figure 1 is the local coordinate system of the vehicle camera (X c , Y c , Z c ), where the direction of the Z c axis is the optical axis direction of the vehicle camera, and the direction of the Y c axis is perpendicular to the Z c axis In the lower direction, the X c axis direction is the direction perpendicular to the Y c axis and Z c axis.
  • the vehicle 10 may include a pose information determination device 14 that can communicate with the image acquisition device and be used to perform the pose information determination method provided by the present application.
  • the vehicle-mounted camera 12 continuously captures video images while the vehicle 10 is running
  • the posture information determination device 14 obtains the image captured by the vehicle-mounted camera 12, and determines that the vehicle-mounted camera 12 acquires the current frame and the previous frame images
  • the relative pose information, the first set of translation parameters, and the adjustment coefficients of the onboard camera 12 determine the pose information of the onboard camera 12 when acquiring the current frame image.
  • the posture information determination method proposed by the present application is executed by the posture information determination device 14 to determine the posture relationship of the in-vehicle camera 12, and then locate the in-vehicle camera 12.
  • the method 100 for determining pose information according to the present application includes the following steps:
  • Step S110 Determine the relative pose information of the image acquisition device when acquiring the current frame image relative to when acquiring the previous frame image.
  • the image acquisition device may be a camera, a camera, or the like.
  • the camera may be a commonly used monocular camera, binocular camera, or more.
  • any other type of camera known in the art and likely to appear in the future can be applied to the present application, and the method of capturing images is not particularly limited in this application, as long as a clear image can be obtained.
  • the image data collected by the camera may be, for example, a sequence of continuous image frames (ie, a video stream) or a sequence of discrete image frames (ie, an image data group sampled at a predetermined sampling time point).
  • the previous frame image acquired by the image acquisition device refers to the previous frame image before the current frame image acquired by the image acquisition device, the penultimate frame image before the current frame image, or any frame image before the current frame image, etc. .
  • the previous frame image and the current frame image can be separated by one frame image, two frame images, or any frame image.
  • the previous frame image refers to the previous frame image before the current frame image, and selecting the previous frame image before the current frame image as the previous frame can reduce calculation errors.
  • the posture information of the image acquisition device when acquiring the current frame image refers to the posture information of the image acquisition device when acquiring the current frame image, including the rotation matrix R and the translation vector t, where the translation vector t is 3* 1 vector, which represents the position of the image acquisition device relative to the origin, the rotation matrix R is a 3*3 matrix, which represents the attitude of the image acquisition device at this time, and the rotation matrix R can also be expressed as the Euler angle In the form of ⁇ , where ⁇ represents the yaw angle of rotation around the Y axis, and ⁇ represents the pitch angle of rotation along the X axis, Represents the roll angle of rotation along the Z axis.
  • the relative posture information of the image acquisition device when acquiring the current frame image relative to the previous frame image refers to the image acquisition device's posture information when acquiring the previous frame image, and the image acquisition device is acquiring The relative change amount of the pose information of the current frame image with respect to the pose information of the image acquisition device when acquiring the previous frame image.
  • the relative position and posture information of the image acquisition device when acquiring the current frame image relative to the previous frame image is obtained through a visual odometer or visual SLAM system, or calculated from relative pose information known in the art Calculated by the method, for example, the relative pose information can also be obtained through the IMU.
  • step S120 a first group of translation parameters of the movement of the image acquisition device during acquiring the current frame and the previous frame images is determined by a sensor with an absolute scale.
  • the senor with absolute scale may be, for example, a wheel speed encoder, speedometer, odometer, or the like.
  • the absolute scale is also called absolute position, and the sensor with absolute scale can measure the positional relationship relative to the real physical world.
  • the first set of translation parameters of the image acquisition device during the acquisition of the current frame and the previous frame image is determined by the sensor with an absolute scale is the displacement vector obtained by the sensor between the acquisition of the previous frame image and the current frame image .
  • Step S130 Adjust the relative pose information based on the relative pose information and the first set of translation parameters.
  • the relative pose information includes a second set of translation parameters, and the second set of translation parameters is a translation vector t in the relative pose information.
  • the adjustment of the second set of translation parameters that is, the translation vector in the relative pose information, based on the second set of translation parameters and the first set of translation parameters based on the relative pose information, is the adjustment of the relative pose information.
  • the rotation matrix of the relative pose information can also be adjusted based on the rotation matrix of the relative pose information and the first set of translation parameters.
  • Step S140 based on the adjusted relative pose information, determine the pose information of the image acquisition device when acquiring the current frame image.
  • the determining the pose information of the current frame of the image acquisition device based on the adjusted relative pose information includes: based on the adjusted relative pose information and the image acquisition device in Acquiring posture information of the previous frame image, and determining posture information of the image acquisition device when acquiring the current frame image. For example, after obtaining the adjusted relative pose information, vector addition is performed on the adjusted relative pose information and the pose information of the image acquisition device when acquiring the previous frame image to obtain the image acquisition device when acquiring the current frame image Posture information.
  • the image acquisition device based on the sensor with absolute scale acquires the scale of the translation parameter (that is, the translation distance) of the movement between the current frame image and the previous frame image, the image obtained by using the sensor with absolute scale
  • the scale of the translation parameter of the acquisition device moving between the current frame image and the previous frame image adjusts the scale of the translation parameter of the relative pose information of the image acquisition device to eliminate or at least reduce the scale drift that may be generated by the image acquisition device,
  • the accuracy of the adjusted pose information is improved.
  • a more accurate pose information of the camera can be obtained at low cost.
  • step S130 includes: determining an adjustment coefficient based on the first set of translation parameters and a second set of translation parameters; adjusting the second set of translation parameters based on the adjustment coefficient and the second set of translation parameters .
  • the adjustment coefficient refers to a coefficient for adjusting the second group of translation parameters according to the first group of translation parameters and the second group of translation parameters. That is to say, the adjustment coefficient is a factor related to the first group of translation parameters and the second group of translation parameters, for example: determining the second norm of the first group of translation parameters and the second norm of the second group of translation parameters; The adjustment coefficient is determined based on the ratio of the second norm of the first set of translation parameters and the second norm of the second set of translation parameters.
  • the second set of translation parameters that is, the translation vector t(x, y, z) in the relative pose information of the image acquisition device when acquiring the current frame image, respectively calculate the translation parameters t_s (x_s, y_s, z_s) and translation
  • the second norm of the vector t(x,y,z) yields
  • the scale of the first group of translation parameters can be determined by calculating the ratio of the second norm of the first group of translation parameters and the second norm of the second group of translation parameters The ratio of the scale of the second set of translation parameters.
  • the determination of the adjustment coefficient based on the first set of translation parameters and the second set of translation parameters may further include: increasing the ratio of the second norm of the first set of translation parameters and the second norm of the second set of translation parameters The offset determines the adjustment coefficient, and fine-tuning the comparison value itself to determine a more accurate adjustment coefficient.
  • the adjusting the second set of translation parameters based on the adjustment coefficient and the second set of translation parameters includes adjusting based on a product of the adjustment coefficient and the second set of translation parameters The second set of translation parameters.
  • the adjustment coefficient is
  • adjust the translation vector of the pose information of the image acquisition device when acquiring the current frame image to obtain the adjusted translation vector of the relative pose information: t_updated t*(
  • the second set of translation parameters is adjusted by the product of the ratio of the second norm of the first set of translation parameters and the second set of translation parameters and the second set of translation parameters, so that the scale of the adjusted second set of translation parameters
  • the scale of the first group of translation parameters is consistent, that is, the scale obtained by the absolute sensor is used to adjust the scale of the relative pose.
  • the adjusting the second set of translation parameters based on the adjustment coefficient and the second set of translation parameters may further include: increasing the offset after the product The product itself is fine-tuned so that the second set of translation parameters can be adjusted more accurately.
  • a method for constructing a visual point cloud which includes: acquiring the pose information of an image acquisition device by the method for determining pose information according to the present application; Construct a visual point cloud based on the pose information of the image acquisition device.
  • a sensor with an absolute scale is used to measure the scale of the movement of the image acquisition device during the acquisition of the current frame image and the previous frame image, and the scale is used to shift the translation vector in the relative pose information of the image acquisition device. Carry out correction, so that the image acquisition device can obtain more accurate posture information, and further obtain a more accurate visual point cloud.
  • target detection is performed on an image acquired by an image acquisition device to acquire pixel targets and their attribute information in the image; based on the pose information of the image acquisition device, the world coordinate system of each pixel target in the image is determined Three-dimensional coordinates; and combining the three-dimensional coordinates of each pixel target in the world coordinate system of each pixel target to generate a visual point cloud.
  • FIG. 2 shows a schematic diagram of a specific example posture information determination apparatus according to an embodiment of the present application.
  • the pose information determination apparatus 200 includes a relative pose information determination unit 210 for determining the relative pose of the image acquisition device when acquiring the current frame image relative to when acquiring the previous frame image Information; relative displacement parameter acquisition unit 220, used to determine the first set of translation parameters of the image acquisition device during the acquisition of the current frame and the previous frame image by an absolute scale sensor; absolute distance acquisition unit 220, with For obtaining the current translation distance of the camera; the relative pose information adjustment unit 230, used to adjust the relative pose information based on the relative pose information and the first set of translation parameters; and the pose information determination unit 240, It is used to determine the pose information of the image acquisition device when acquiring the current frame image based on the adjusted relative pose information.
  • the relative pose information includes a second set of translation parameters
  • the relative pose information adjustment unit 230 is further configured to determine an adjustment coefficient based on the first set of translation parameters and the second set of translation parameters; based on the adjustment Coefficients and the second set of translation parameters to adjust the second set of translation parameters.
  • the relative pose information adjustment unit 230 is further used to determine the second norm of the first set of translation parameters and the second norm of the second set of translation parameters; based on the first set of translation parameters The ratio of the second norm to the second norm of the second set of translation parameters determines the adjustment coefficient.
  • the relative pose information adjustment unit 230 is further configured to adjust the second set of translation parameters based on the product of the adjustment coefficient and the second set of translation parameters.
  • the relative pose information determination unit 210 determines the relative pose information of the image acquisition device when acquiring the current frame image relative to when acquiring the previous frame image based on the visual odometer.
  • the pose information determination unit 240 is used to determine that the image acquisition device is acquiring based on the adjusted relative pose information and the pose information of the image acquisition device when acquiring the previous frame image The pose information of the current frame image.
  • a visual point cloud construction device which not only includes all units of the posture information determination device of the present application, but also includes visual point cloud construction
  • the unit is configured to construct a visual point cloud based on the pose information of the image acquisition device when acquiring the current frame image.
  • FIG. 4 illustrates a structural block diagram of an electronic device 300 according to an embodiment of the present application.
  • an electronic device 300 according to an embodiment of the present application will be described with reference to FIG. 4.
  • the electronic device 300 may be implemented as the posture information determination device 14 in the vehicle 10 shown in FIG. 1, which may communicate with the vehicle-mounted camera 12. To receive their output signals.
  • the electronic device 300 may include a processor 310 and a memory 320.
  • the processor 310 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 300 to perform desired functions.
  • CPU central processing unit
  • the processor 310 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 300 to perform desired functions.
  • the memory 320 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory.
  • the volatile memory may include, for example, random access memory (RAM) and/or cache memory.
  • the non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, and the like.
  • One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 310 may execute the program instructions to implement the pose information determination method and vision of various embodiments of the present application described above Point cloud construction method and/or other desired functions.
  • Various contents such as camera-related information, sensor-related information, and driver programs can also be stored in the computer-readable storage medium.
  • the electronic device 300 may further include an interface 330, an input device 340, and an output device 350, and these components are interconnected by a bus system and/or other forms of connection mechanisms (not shown).
  • the interface 330 may be used to connect to a camera, such as a video camera.
  • the interface 330 may be a commonly used USB interface of a camera, and of course, it may be other interfaces such as a Type-C interface.
  • the electronic device 300 may include one or more interfaces 330 to connect to corresponding cameras, and receive images captured by the cameras for performing the pose information determination method and visual point cloud construction method described above.
  • the input device 340 may be used to receive external input, such as receiving physical point coordinate values input by a user.
  • the input device 340 may be, for example, a keyboard, a mouse, a tablet, a touch screen, and so on.
  • the output device 350 can output the calculated camera external parameters.
  • the output device 350 may include a display, a speaker, a printer, and a communication network and its connected remote output device.
  • the input device 340 and the output device 350 may be an integrated touch display screen.
  • FIG. 4 only shows some components of the electronic device 300 related to the present application, and omits some related peripheral or auxiliary components.
  • the electronic device 300 may further include any other suitable components.
  • embodiments of the present application may also be computer program products, which include computer program instructions that when executed by a processor cause the processor to perform the above-described "exemplary method" of this specification
  • the computer program product may write program codes for performing operations of the embodiments of the present application in any combination of one or more programming languages, and the programming languages include object-oriented programming languages, such as Java, C++, etc. , Also includes conventional procedural programming languages, such as "C" language or similar programming languages.
  • the program code may be executed entirely on the user's computing device, partly on the user's device, as an independent software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server On the implementation.
  • an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon, which when executed by a processor causes the processor to perform the above-mentioned "exemplary method" part of the specification The steps in the method for determining pose information and the method for constructing a visual point cloud according to various embodiments of the present application described in.
  • the computer-readable storage medium may employ any combination of one or more readable media.
  • the readable medium may be a readable signal medium or a readable storage medium.
  • the readable storage medium may include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any combination of the above, for example. More specific examples of readable storage media (non-exhaustive list) include: electrical connections with one or more wires, portable disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the foregoing.
  • each component or each step can be decomposed and/or recombined.
  • decompositions and/or recombinations shall be regarded as equivalent solutions of this application.

Abstract

公开了一种位姿信息确定方法和确定装置、视觉点云构建方法和视觉点云构建装置。一种位姿信息确定方法包括确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；基于所述相对位姿信息和第一组平移参数，调整所述相对位姿信息；以及基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息。采用上述位姿信息确定方法，直接使用外部传感器的尺度对相机的位姿信息的平移向量进行尺度校正，得到更准确的位姿信息。

Description

位姿信息确定方法和装置、视觉点云构建方法和装置 技术领域
本申请涉及计算机视觉领域，具体地，本申请涉及一种位姿信息确定方法、位姿信息确定装置、视觉点云构建方法和视觉点云构建装置、电子设备以及计算机可读的存储介质。
背景技术
地图是无人驾驶领域的基础。然而，在单目相机SLAM中，由于单目相机的尺度不确定性，导致无法构建出全局尺度一致的向量地图，而且由于单目相机的尺度不确定性，单目SLAM在多帧跟踪图像之间容易由于尺度漂移而使跟踪结果误差累积，最终导致跟踪失败。
在现有技术中，通常通过双目视觉技术，直接在每个时刻获得地图点的真实三维尺度，或者融合高精度组合导航模块(IMU)，直接通过IMU模块测量得到真实尺度线加速度积分，获得帧间真实尺度的位姿信息。然而，尽管通过使用双目视觉技术或者IMU模块可以获得帧间图像的真实尺度，但是由于传感器成本高昂、计算成本高、制作工作成本高、标定复杂、算法较为复杂，并且受IMU本身的成本影响较大，使视觉点云的使用遭到了巨大障碍。
因此，需要一种能够低成本、高精度、适用范围广地确定相机位姿参数及视觉点云构建的方法及装置。
发明内容
为了解决上述技术问题，提出了本申请。本申请的实施例提供了一种位姿信息确定方法、位姿信息确定装置、视觉点云构建方法和视觉点云构建装置、电子设备以及计算机可读的存储介质，以低成本地、高精度、适用范围广地确定图像获取设备的位姿信息。
根据本申请的一个方面，提供了一种位姿信息确定方法，包括确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；通 过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；基于所述相对位姿信息和第一组平移参数，调整所述相对位姿信息；以及基于调整后的所述相对位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。
根据本申请的另一方面，提供一种视觉点云构建方法，包括：通过上述位姿信息确定方法获取所述图像获取设备的位姿信息；和基于所述图像获取设备的位姿信息构建视觉点云。
根据本申请的又一方面，提供了一种位姿信息确定装置，包括相对位姿信息确定单元，用于确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；相对位移参数获取单元，用于通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间的第一组平移参数；相对位姿信息调整单元，用于基于所述相对位姿信息和所述第一组平移参数，调整所述相对位姿信息；和位姿信息确定单元，用于基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息。
根据本申请的又一方面，提供了一种视觉点云构建装置，包括相对位姿信息确定单元，用于确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；相对位移参数获取单元，用于通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；相对位姿信息调整单元，用于基于所述相对位姿信息和所述第一组平移参数，调整所述相对位姿信息；位姿信息确定单元，用于基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息；和视觉点云构建单元，用于基于所述图像获取设备在获取当前帧图像时的位姿信息，构建视觉点云。
根据本申请的又一方面，提供了一种电子设备，包括处理器，以及存储器，在所述存储器中存储有计算机程序指令，所述计算机程序指令在被所述处理器运行时使得所述处理器执行上述位姿信息确定方法、或上述视觉点云构建方法。
根据本申请的又一方面，提供了一种计算机可读的存储介质，其上存储有用于执行上述位姿信息确定方法、或上述视觉点云构建方法的指令。
与现有技术相比，采用根据本申请实施例的位姿信息确定方法、位姿信 息确定装置、视觉点云构建方法和视觉点云构建装置、电子设备以及计算机可读的存储介质，可以通过确定图像获取设备在获取当前帧和先前帧图像时的相对位姿信息、第一组平移参数和调整系数，得到所述图像获取设备在获取当前帧图像时的更准确的位姿信息。因此，通过直接使用外部传感器的尺度对图像获取设备的位姿信息的平移向量进行尺度校正，得到更准确的位姿信息，从而不会因为传感器配置的变更而对算法框架造成影响，还降低了传感器成本及计算成本，进而降低了单目视觉系统的部署难度。
附图说明
通过结合附图对本申请实施例进行更详细的描述，本申请的上述以及其他目的、特征和优势将变得更加明显。附图用来提供对本申请实施例的进一步理解，并且构成说明书的一部分，与本申请实施例一起用于解释本申请，并不构成对本申请的限制。在附图中，相同的参考标号通常代表相同部件或步骤。
图1图示了根据本申请实施例的位姿信息确定方法的应用场景示意图；
图2图示了根据本申请一实施例的位姿信息确定方法的流程图。
图3图示了根据本申请一实施例的位姿信息确定装置的示意图。
图4图示了根据本申请一实施例的电子设备的框图。
具体实施方式
下面，将参考附图详细地描述根据本申请的示例实施例。显然，所描述的实施例仅仅是本申请的一部分实施例，而不是本申请的全部实施例，应理解，本申请不受这里描述的示例实施例的限制。
申请概述
如上所述，在无人驾驶中，相机的位姿信息并由此计算得到相机的当前位置是非常重要的。然而，为了获得相机的准确位姿信息和当前位置的成本比较高。因此，需要改进的位姿信息确定方法，使获得相机的准确位姿信息的成本得以降低。
针对该技术问题，本申请的基本构思是提出一种位姿信息确定方法、位姿信息确定装置、视觉点云构建方法和视觉点云构建装置、电子设备以及计 算机可读的存储介质，其可以通过直接使用外部传感器的尺度，尤其是标量尺度，对图像获取设备的平移向量进行尺度校正，由此降低了成本，使单目视觉系统的部署难度降低。
换言之，通过本申请的位姿信息确定方法和确定装置，不需要使用高精度的传感器，也不需要人工过多干预，就能得到更准确的位姿信息，进而得到全局一致的视觉点云，并由此建立高精度矢量地图，因而降低了高精度地图的制作成本。
需要说明的是，本申请的上述基本构思不但可以应用于地图制作，也可以应用于其它领域，例如机器人及无人交通工具导航领域等。
在介绍了本申请的基本原理之后，下面将参考附图来具体介绍本申请的各种非限制性实施例。
示例性场景
图1图示了根据本申请实施例的位姿信息确定方法的应用场景的示意图。如图1所示，车辆10可包括图像获取设备，例如车载相机12，其可以是常用的单目相机、双目相机、或者更多目相机。虽然图1示出了车载相机12安装于车辆10的顶部，但是应理解，车载相机亦可安装于车辆10的其他位置处，例如车头部分处、前挡风玻璃处，等等。
图1中所示的坐标系是车载相机局部坐标系(X c，Y c，Z c)，其中Z c轴的方向为车载相机的光轴方向，Y c轴方向为垂直于Z c轴向下的方向，X c轴方向为垂直于Y c轴和Z c轴的方向。
这里，车辆10可包括位姿信息确定装置14，位姿信息确定装置14可与图像获取设备通信，并用来执行本申请提供的位姿信息确定方法。在一实施例中，车载相机12在车辆10的行驶过程中，连续拍摄视频图像，位姿信息确定装置14获得车载相机12拍摄的图像，并通过确定车载相机12在获取当前帧和先前帧图像时车载相机12的相对位姿信息、第一组平移参数、调整系数，确定车载相机12在获取当前帧图像时的位姿信息。
通过位姿信息确定装置14执行本申请提出的位姿信息确定方法，可以确定车载相机12的位姿关系，进而对车载相机12进行定位。
示例性方法1
图2是根据本申请一示例性实施例的位姿信息确定方法的流程示意图。如图2所示，根据本申请的位姿信息确定方法100包括如下步骤：
步骤S110，确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息。
图像获取设备可以是相机、摄像头等。相机可以是常用的单目相机、双目相机、或者更多目相机。当然，本领域中已知的以及将来可能出现的任何其他类型的相机都可以应用于本申请，本申请对其捕捉图像的方式没有特别限制，只要能够获得清晰的图像即可。相机采集到的图像数据例如可以是连续图像帧序列(即，视频流)或离散图像帧序列(即，在预定采样时间点采样到的图像数据组)等。
在一个示例中，图像获取设备获取的先前帧图像是指图像获取设备获取的当前帧图像之前的上一帧图像、当前帧图像之前的倒数第二帧图像或者当前帧图像之前的任意帧图像等。也就是说，先前帧图像和当前帧图像之间可间隔一帧图像、两帧图像或任意帧图像。在一个示例中，先前帧图像是指当前帧图像之前的上一帧图像，选择当前帧图像之前的上一帧图像作为先前帧，可以减小计算误差。
在一个示例中，图像获取设备获取当前帧图像时的位姿信息是指图像获取设备在采集当前帧图像时的位姿信息，包括旋转矩阵R和平移向量t，其中，平移向量t是3*1向量，表示图像获取设备相对于原点的位置，旋转矩阵R是3*3矩阵，表示图像获取设备此时的姿态，旋转矩阵R也可以表示成欧拉角
Figure PCTCN2019099207-appb-000001
的形式，其中ψ表示绕Y轴旋转的航向角(yaw)，θ表示沿X轴旋转的俯仰角(pitch)，
Figure PCTCN2019099207-appb-000002
表示沿Z轴旋转的滚转角(roll)。
在一个示例中，图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息是指以图像获取设备在获取先前帧图像时的位姿信息为基准，图像获取设备在获取当前帧图像时的位姿信息相对于图像获取设备在获取先前帧图像时的位姿信息的相对变化量。
在一个示例中，图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息是通过视觉里程计或视觉SLAM系统获取的，或者通过本领域已知的相对位姿信息计算方法计算得到的，例如，该相对位姿信息还可以通过IMU等获得。
步骤S120，通过带有绝对尺度的传感器确定所述图像获取设备在获取 所述当前帧与先前帧图像期间运动的第一组平移参数。
在一个示例中，带有绝对尺度的传感器可以是例如轮速编码器、速度计、行程计等。在这里，绝对尺度又称绝对位置，带有绝对尺度的传感器能够测量相对于真实物理世界的位置关系。
通过带有绝对尺度的传感器确定图像获取设备在获取当前帧和先前帧图像期间运动的第一组平移参数是通过传感器得到图像获取设备在采集先前帧图像和采集当前帧图像之间运动的位移向量。
步骤S130，基于所述相对位姿信息和第一组平移参数，调整所述相对位姿信息。
其中，相对位姿信息包括第二组平移参数，所述第二组平移参量即相对位姿信息中的平移向量t。基于所述相对位姿信息的第二组平移参数和第一组平移参数对第二组平移参数即所述相对位姿信息中的平移向量的调整，即是对相对位姿信息的调整。
在一个示例中，还可以基于相对位姿信息的旋转矩阵和第一组平移参数，调整相对位姿信息的旋转矩阵。
步骤S140，基于调整后的所述相对位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。
在一个示例中，所述基于调整后的所述相对位姿信息，确定所述图像获取设备当前帧的位姿信息，包括：基于调整后的所述相对位姿信息以及所述图像获取设备在获取所述先前帧图像时的位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。例如，在得到调整后的相对位姿信息之后，对调整后的相对位姿信息和图像获取设备在采集先前帧图像时的位姿信息进行矢量加和，得到图像获取设备在采集当前帧图像时的位姿信息。
由于基于带有绝对尺度的传感器得到的图像获取设备在采集当前帧图像和先前帧图像之间运动的平移参数的尺度(即平移距离)较为精准，因此，利用带有绝对尺度的传感器得到的图像获取设备在采集当前帧图像和先前帧图像之间运动的平移参数的尺度对图像获取设备的相对位姿信息的平移参数的尺度进行调整，消除或者至少减小图像获取设备可能产生的尺度漂移，使调整后的位姿信息的精确度提高。对于根据本申请实施例的位姿信息确定方法，通过采用这种位姿信息确定方法，可以低成本地得到更准确的相机的位姿信息。
在一个示例中，步骤S130包括：基于所述第一组平移参数和第二组平移参数，确定调整系数；基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数。
这里，调整系数是指根据第一组平移参数和第二组平移参数，对第二组平移参数进行调整的系数。也就是说，调整系数是与第一组平移参数和第二组平移参数有关的因数，例如：确定所述第一组平移参数的二范数和所述第二组平移参数的二范数；基于所述第一组平移参数的二范数和所述第二组平移参数的二范数的比值，确定所述调整系数。
在进一步的示例中，根据已获取的第一组平移参数，即带有绝对尺度的传感器在图像获取设备采集先前帧图像和当前帧图像之间平移的平移参数t_s(x_s,y_s,z_s)，和第二组平移的参数，即图像获取设备在采集当前帧图像时的相对位姿信息中的平移向量t(x,y,z)，分别计算平移参数t_s(x_s，y_s，z_s)和平移向量t(x,y,z)的二范数，得到||t s(x s，y s，z s)|| 2和||t(x，y，z)|| 2，计算二者的比值，得到调整系数：||t_s(x_s，y_s，z_s)|| 2/||t(x，y，z)|| 2。
由于对平移参数取二范数即计算尺度，因此，通过计算第一组平移参数的二范数和所述第二组平移参数的二范数的比值，可以确定出第一组平移参数的尺度和第二组平移参数的尺度的比值。
此外，所述基于第一组平移参数和第二组平移参数确定调整系数还可以包括：将所述第一组平移参数的二范数和所述第二组平移参数的二范数的比值增加偏置量确定调整系数，对比值本身进行微调，从而确定更加准确的调整系数。
在进一步的示例中，所述基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数，包括：基于所述调整系数与所述第二组平移参数的乘积，调整所述第二组平移参数。例如，根据已获取的调整系数和相对位姿信息的平移向量即第二组平移参数，在调整系数为||t_s(x_s，y_s，z_s)|| 2/||t(x，y，z)|| 2时，调整图像获取设备在采集当前帧图像时的位姿信息的平移向量，得到调整后的相对位姿信息的平移向量：t_updated＝t*(||t_s(x_s，y_s，z_s)|| 2/||t(x，y，z)|| 2)。通过采用本示例，通过第一组平移参数和第二组平移参数的二范数的比值与第二组平移参数的乘积来调整第二组平移参数，使调整后的第二组平移参数的尺度与第一组平 移参数的尺度一致，即利用了绝对传感器获得的尺度对相对位姿的尺度进行了调整。
此外，所述基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数调整所述第二组平移参数还可以包括：通过对所述乘积后增加偏置量，对乘积本身进行微调，使第二组平移参数调整的更加准确。
在一示例中，在根据本申请的位姿信息确定方法的基础上，提出一种视觉点云构建方法，包括：通过根据本申请的位姿信息确定方法获取图像获取设备的位姿信息；和基于所述图像获取设备的位姿信息构建视觉点云。通过采用本实施例，利用带有绝对尺度的传感器测量得到图像获取设备在获取当前帧图像和先前帧图像期间运动的尺度，并利用该尺度移对图像获取设备的相对位姿信息中的平移向量进行校正，进而使图像获取设备获得更加精准的位姿信息，并进一步得到更加精准的视觉点云。
例如，对图像获取设备所获取的图像进行目标检测，获取所述图像中的像素目标及其属性信息；基于图像获取设备的位姿信息确定所述图像中每个像素目标的世界坐标系下的三维坐标；以及结合每个像素目标的每个像素目标世界坐标系下的三维坐标，生成视觉点云。
示例性装置
图2示出了根据本申请一实施例的具体示例的位姿信息确定装置的示意图。
如图所示，根据本申请一实施例的位姿信息确定装置200包括相对位姿信息确定单元210，用于确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；相对位移参数获取单元220，用于通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；绝对距离获取单元220，用于获取相机的当前平移距离；相对位姿信息调整单元230，用于基于所述相对位姿信息和所述第一组平移参数，调整所述相对位姿信息；和位姿信息确定单元240，用于基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息。
在一个示例中，相对位姿信息包括第二组平移参数，相对位姿信息调整单元230还用于基于所述第一组平移参数和所述第二组平移参数确定调整系 数；基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数。
在一个进一步的示例中，相对位姿信息调整单元230还用于确定所述第一组平移参数的二范数和所述第二组平移参数的二范数；基于所述第一组平移参数的二范数和所述第二组平移参数的二范数的比值，确定所述调整系数。
在一个进一步的示例中，相对位姿信息调整单元230还用于基于所述调整系数与所述第二组平移参数的乘积，调整所述第二组平移参数。
在一个示例中，相对位姿信息确定单元210是基于视觉里程计确定所述图像获取设备在获取所述当前帧图像时相对于获取先前帧图像时的相对位姿信息的。
在一个示例中，位姿信息确定单元240用于基于调整后的所述相对位姿信息以及所述图像获取设备在获取所述先前帧图像时的位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。
在一个实施例中，在根据本申请的位姿信息确定装置的基础上，提出一种视觉点云构建装置，其不仅包括本申请的位姿信息确定装置的所有单元，还包括视觉点云构建单元，用于基于所述图像获取设备在获取当前帧图像时的位姿信息，构建视觉点云。
上述基于位姿信息确定装置200和视觉点云构建装置中的各个单元和模块的具体功能和操作已经在上面参考图2描述的位姿信息确定方法中详细介绍，并因此，将省略其重复描述。
示例性电子设备
图4图示了根据本申请实施例的电子设备300的结构框图。下面，参考图4来描述根据本申请一实施例的电子设备300，该电子设备300可以实现为图1所示的车辆10中的位姿信息确定装置14，其可以与车载相机12进行通信，以接收它们的输出信号。
如图4所示，电子设备300可包括处理器310和存储器320。
处理器310可以是中央处理单元(CPU)或者具有数据处理能力和/或指令执行能力的其他形式的处理单元，并且可以控制电子设备300中的其他组件以执行期望的功能。
存储器320可以包括一个或多个计算机程序产品，所述计算机程序产品可以包括各种形式的计算机可读存储介质，例如易失性存储器和/或非易失性 存储器。所述易失性存储器例如可以包括随机存取存储器(RAM)和/或高速缓冲存储器(cache)等。所述非易失性存储器例如可以包括只读存储器(ROM)、硬盘、闪存等。在所述计算机可读存储介质上可以存储一个或多个计算机程序指令，处理器310可以运行所述程序指令，以实现上文所述的本申请的各个实施例的位姿信息确定方法、视觉点云构建方法以及/或者其他期望的功能。在所述计算机可读存储介质中还可以存储诸如相机的相关信息、传感器的相关信息以及驱动程序等各种内容。
在一个示例中，电子设备300还可以包括接口330、输入装置340和输出装置350，这些组件通过总线系统和/或其他形式的连接机构(未示出)互连。
接口330可用于连接到摄像头，例如摄像机。例如，接口330可以是摄像头常用的USB接口，当然也可以是其他接口例如Type-C接口等。电子设备300可包括一个或多个接口330，以连接到相应的摄像机，并且从摄像机接收其所拍摄的图像以用于执行上面描述的位姿信息确定方法和视觉点云构建方法。
输入装置340可用于接收外界输入，例如接收用户输入的物理点坐标值等。在一些实施例中，输入装置340可以是例如键盘、鼠标、手写板、触摸屏等。
输出装置350可以输出所计算的摄像机外参。例如，输出装置350可以包括显示器、扬声器、打印机、以及通信网络及其所连接的远程输出设备等。在一些实施例中，输入装置340和输出装置350可以是集成一体的触摸显示屏。
为了简化，图4中仅示出了电子设备300中与本申请有关的一些组件，而省略了一些相关外围或辅助组件。除此之外，根据具体应用情况，电子设备300还可以包括任何其他适当的组件。
示例性计算机程序产品和计算机可读存储介质
除了上述方法和设备以外，本申请的实施例还可以是计算机程序产品，其包括计算机程序指令，所述计算机程序指令在被处理器运行时使得所述处理器执行本说明书上述“示例性方法”部分中描述的根据本申请各种实施例的位姿信息确定方法和视觉点云构建方法中的步骤。
所述计算机程序产品可以以一种或多种程序设计语言的任意组合来编写用于执行本申请实施例操作的程序代码，所述程序设计语言包括面向对象的程序设计语言，诸如Java、C++等，还包括常规的过程式程序设计语言，诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户计算设备上部分在远程计算设备上执行、或者完全在远程计算设备或服务器上执行。
此外，本申请的实施例还可以是计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令在被处理器运行时使得所述处理器执行本说明书上述“示例性方法”部分中描述的根据本申请各种实施例的位姿信息确定方法和视觉点云构建方法中的步骤。
所述计算机可读存储介质可以采用一个或多个可读介质的任意组合。可读介质可以是可读信号介质或者可读存储介质。可读存储介质例如可以包括但不限于电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。可读存储介质的更具体的例子(非穷举的列表)包括：具有一个或多个导线的电连接、便携式盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。
以上结合具体实施例描述了本申请的基本原理，但是，需要指出的是，在本申请中提及的优点、优势、效果等仅是示例而非限制，不能认为这些优点、优势、效果等是本申请的各个实施例必须具备的。另外，上述公开的具体细节仅是为了示例的作用和便于理解的作用，而非限制，上述细节并不限制本申请为必须采用上述具体的细节来实现。
本申请中涉及的器件、装置、设备、系统的方框图仅作为例示性的例子并且不意图要求或暗示必须按照方框图示出的方式进行连接、布置、配置。如本领域技术人员将认识到的，可以按任意方式连接、布置、配置这些器件、装置、设备、系统。诸如“包括”、“包含”、“具有”等等的词语是开放性词汇，指“包括但不限于”，且可与其互换使用。这里所使用的词汇“或”和“和”指词汇“和/或”，且可与其互换使用，除非上下文明确指示不是如此。这里所使用 的词汇“诸如”指词组“诸如但不限于”，且可与其互换使用。
还需要指出的是，在本申请的装置、设备和方法中，各部件或各步骤是可以分解和/或重新组合的。这些分解和/或重新组合应视为本申请的等效方案。
提供所公开的方面的以上描述以使本领域的任何技术人员能够做出或者使用本申请。对这些方面的各种修改对于本领域技术人员而言是非常显而易见的，并且在此定义的一般原理可以应用于其他方面而不脱离本申请的范围。因此，本申请不意图被限制到在此示出的方面，而是按照与在此公开的原理和新颖的特征一致的最宽范围。
为了例示和描述的目的已经给出了以上描述。此外，此描述不意图将本申请的实施例限制到在此公开的形式。尽管以上已经讨论了多个示例方面和实施例，但是本领域技术人员将认识到其某些变型、修改、改变、添加和子组合。

Claims (11)

  1. 一种位姿信息确定方法，包括：
    确定图像获取设备在获取当前帧图像相对于获取先前帧图像时的相对位姿信息；
    通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；
    基于所述相对位姿信息和第一组平移参数，调整所述相对位姿信息；以及
    基于调整后的所述相对位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。
  2. 根据权利要求1所述的位姿信息确定方法，其中，所述相对位姿信息包括第二组平移参数；
    所述基于所述相对位姿信息和第一组平移参数，调整所述相对位姿信息包括：
    基于所述第一组平移参数和所述第二组平移参数确定调整系数；
    基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数。
  3. 根据权利要求2所述的位姿信息确定方法，其中所述基于第一组平移参数和第二组平移参数确定调整系数，包括：
    确定所述第一组平移参数的二范数和所述第二组平移参数的二范数；
    基于所述第一组平移参数的二范数和所述第二组平移参数的二范数的比值，确定所述调整系数。
  4. 根据权利要求3所述的位姿信息确定方法，所述基于所述调整系数和所述第二组平移参数，调整所述第二组平移参数，包括：
    基于所述调整系数与所述第二组平移参数的乘积，调整所述第二组平移参数。
  5. 根据权利要求1所述的位姿信息确定方法，其中，所述确定图像获 取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息，包括：
    基于视觉里程计确定所述图像获取设备在获取所述当前帧图像时相对于获取先前帧图像时的相对位姿信息。
  6. 根据权利要求1所述的位姿信息确定方法，其中，所述基于调整后的所述相对位姿信息，确定所述图像获取设备当前帧的位姿信息，包括：
    基于调整后的所述相对位姿信息以及所述图像获取设备在获取所述先前帧图像时的位姿信息，确定所述图像获取设备在获取所述当前帧图像时的位姿信息。
  7. 一种视觉点云构建方法，包括：
    通过权利要求1-6任一所述的方法获取所述图像获取设备的位姿信息；
    基于所述图像获取设备的位姿信息构建视觉点云。
  8. 一种位姿信息确定装置，包括：
    相对位姿信息确定单元，用于确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；
    相对位移参数获取单元，用于通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；
    相对位姿信息调整单元，用于基于所述相对位姿信息和所述第一组平移参数，调整所述相对位姿信息；和
    位姿信息确定单元，用于基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息。
  9. 一种视觉点云构建装置，包括：
    相对位姿信息确定单元，用于确定图像获取设备在获取当前帧图像时相对于获取先前帧图像时的相对位姿信息；
    相对位移参数获取单元，用于通过带有绝对尺度的传感器确定所述图像获取设备在获取所述当前帧与先前帧图像期间运动的第一组平移参数；
    相对位姿信息调整单元，用于基于所述相对位姿信息和所述第一组平移参数，调整所述相对位姿信息；
    位姿信息确定单元，用于基于调整后的所述相对位姿信息，确定所述图像获取设备在获取当前帧图像时的位姿信息；和
    视觉点云构建单元，用于基于所述图像获取设备在获取当前帧图像时的位姿信息，构建视觉点云。
  10. 一种电子设备，包括：
    处理器；以及
    存储器，在所述存储器中存储有计算机程序指令，所述计算机程序指令在被所述处理器运行时使得所述处理器执行如权利要求1-6中任一项所述的位姿信息确定方法、或如权利要求7所述的视觉点云构建方法。
  11. 一种计算机可读的存储介质，其上存储有用于执行权利要求1-6中任一项所述的位姿信息确定方法、或如权利要求7所述的视觉点云构建方法的指令。
PCT/CN2019/099207 2018-11-30 2019-08-05 位姿信息确定方法和装置、视觉点云构建方法和装置 WO2020107931A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811459199.9A CN109544630B (zh) 2018-11-30 2018-11-30 位姿信息确定方法和装置、视觉点云构建方法和装置
CN201811459199.9 2018-11-30

Publications (1)

Publication Number Publication Date
WO2020107931A1 true WO2020107931A1 (zh) 2020-06-04

Family

ID=65851743

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/099207 WO2020107931A1 (zh) 2018-11-30 2019-08-05 位姿信息确定方法和装置、视觉点云构建方法和装置

Country Status (2)

Country Link
CN (1) CN109544630B (zh)
WO (1) WO2020107931A1 (zh)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109544630B (zh) * 2018-11-30 2021-02-02 南京人工智能高等研究院有限公司 位姿信息确定方法和装置、视觉点云构建方法和装置
CN111829489B (zh) * 2019-04-16 2022-05-13 杭州海康机器人技术有限公司 一种视觉定位的方法及装置
CN112097742B (zh) * 2019-06-17 2022-08-30 北京地平线机器人技术研发有限公司 一种位姿确定方法及装置
CN112444242B (zh) * 2019-08-31 2023-11-10 北京地平线机器人技术研发有限公司 一种位姿优化方法及装置
CN110738699A (zh) * 2019-10-12 2020-01-31 浙江省北大信息技术高等研究院 一种无监督绝对尺度计算方法及系统
CN110820447A (zh) * 2019-11-22 2020-02-21 武汉纵横天地空间信息技术有限公司 一种基于双目视觉的轨道几何状态测量系统及其测量方法
CN110889871B (zh) * 2019-12-03 2021-03-23 广东利元亨智能装备股份有限公司 机器人行驶方法、装置及机器人
CN113748693B (zh) * 2020-03-27 2023-09-15 深圳市速腾聚创科技有限公司 路基传感器的位姿校正方法、装置和路基传感器
CN113034594A (zh) * 2021-03-16 2021-06-25 浙江商汤科技开发有限公司 位姿优化方法、装置、电子设备及存储介质
CN113793381A (zh) * 2021-07-27 2021-12-14 武汉中海庭数据技术有限公司 单目视觉信息和轮速信息融合的定位方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103139446A (zh) * 2011-10-14 2013-06-05 斯凯普公司 接收的视频稳定化
CN106873619A (zh) * 2017-01-23 2017-06-20 上海交通大学 一种无人机飞行路径的处理方法
CN107481292A (zh) * 2017-09-05 2017-12-15 百度在线网络技术（北京）有限公司 车载摄像头的姿态误差估计方法和装置
US20180199039A1 (en) * 2017-01-11 2018-07-12 Microsoft Technology Licensing, Llc Reprojecting Holographic Video to Enhance Streaming Bandwidth/Quality
CN108648240A (zh) * 2018-05-11 2018-10-12 东南大学 基于点云特征地图配准的无重叠视场相机姿态标定方法
CN108827306A (zh) * 2018-05-31 2018-11-16 北京林业大学 一种基于多传感器融合的无人机slam导航方法及系统
CN109544630A (zh) * 2018-11-30 2019-03-29 南京人工智能高等研究院有限公司 位姿信息确定方法和装置、视觉点云构建方法和装置

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9984301B2 (en) * 2015-04-20 2018-05-29 Qualcomm Incorporated Non-matching feature-based visual motion estimation for pose determination
US10282859B2 (en) * 2016-12-12 2019-05-07 The Boeing Company Intra-sensor relative positioning
CN108225345A (zh) * 2016-12-22 2018-06-29 乐视汽车（北京）有限公司 可移动设备的位姿确定方法、环境建模方法及装置
CN106997614B (zh) * 2017-03-17 2021-07-20 浙江光珀智能科技有限公司 一种基于深度相机的大规模场景3d建模方法及其装置
CN107945265B (zh) * 2017-11-29 2019-09-20 华中科技大学 基于在线学习深度预测网络的实时稠密单目slam方法与系统
CN108171728B (zh) * 2017-12-25 2020-06-19 清华大学 基于混合相机系统的无标记运动物体姿态恢复方法及装置
CN108364319B (zh) * 2018-02-12 2022-02-01 腾讯科技（深圳）有限公司 尺度确定方法、装置、存储介质及设备
CN108629793B (zh) * 2018-03-22 2020-11-10 中国科学院自动化研究所 使用在线时间标定的视觉惯性测程法与设备

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103139446A (zh) * 2011-10-14 2013-06-05 斯凯普公司 接收的视频稳定化
US20180199039A1 (en) * 2017-01-11 2018-07-12 Microsoft Technology Licensing, Llc Reprojecting Holographic Video to Enhance Streaming Bandwidth/Quality
CN106873619A (zh) * 2017-01-23 2017-06-20 上海交通大学 一种无人机飞行路径的处理方法
CN107481292A (zh) * 2017-09-05 2017-12-15 百度在线网络技术（北京）有限公司 车载摄像头的姿态误差估计方法和装置
CN108648240A (zh) * 2018-05-11 2018-10-12 东南大学 基于点云特征地图配准的无重叠视场相机姿态标定方法
CN108827306A (zh) * 2018-05-31 2018-11-16 北京林业大学 一种基于多传感器融合的无人机slam导航方法及系统
CN109544630A (zh) * 2018-11-30 2019-03-29 南京人工智能高等研究院有限公司 位姿信息确定方法和装置、视觉点云构建方法和装置

Also Published As

Publication number Publication date
CN109544630B (zh) 2021-02-02
CN109544630A (zh) 2019-03-29

Similar Documents

Publication Publication Date Title
WO2020107931A1 (zh) 位姿信息确定方法和装置、视觉点云构建方法和装置
WO2020140431A1 (zh) 相机位姿确定方法、装置、电子设备及存储介质
US20230360260A1 (en) Method and device to determine the camera position and angle
CN110243358B (zh) 多源融合的无人车室内外定位方法及系统
CN109887057B (zh) 生成高精度地图的方法和装置
WO2021232470A1 (zh) 基于多传感器融合的slam制图方法、系统
CN109544629B (zh) 摄像头位姿确定方法和装置以及电子设备
CN110880189B (zh) 联合标定方法及其联合标定装置和电子设备
CN108932737B (zh) 车载相机俯仰角标定方法和装置、电子设备以及车辆
US8698875B2 (en) Estimation of panoramic camera orientation relative to a vehicle coordinate frame
JP2019526101A (ja) シーンにおけるカメラの姿勢を特定するシステム及び方法
US20180075614A1 (en) Method of Depth Estimation Using a Camera and Inertial Sensor
CN113551665B (zh) 一种用于运动载体的高动态运动状态感知系统及感知方法
CN112050806B (zh) 一种移动车辆的定位方法及装置
CN113516692A (zh) 一种多传感器融合的slam方法和装置
CN113112413A (zh) 图像生成方法、图像生成装置和车载抬头显示系统
CN110793526A (zh) 基于可穿戴单目视觉和惯性传感器融合的行人导航方法及系统
CN110728716B (zh) 一种标定方法、装置及飞行器
CN114777768A (zh) 一种卫星拒止环境高精度定位方法、系统及电子设备
CN114662587A (zh) 一种基于激光雷达的三维目标感知方法、装置及系统
CN114022561A (zh) 一种基于gps约束和动态校正的城区单目测图方法和系统
CN108961337B (zh) 车载相机航向角标定方法和装置、电子设备以及车辆
CN116429098A (zh) 低速无人飞行器视觉导航定位方法及系统
CN112652018B (zh) 外参确定方法、外参确定装置及电子设备
CN110836656B (zh) 用于单目adas的防抖测距方法、装置和电子设备

Legal Events

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

Ref document number: 19888820

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19888820

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 19888820

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 16.12.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19888820

Country of ref document: EP

Kind code of ref document: A1