WO2022166113A1 - 多视角摄像头标定设备、标定方法及存储介质 - Google Patents

多视角摄像头标定设备、标定方法及存储介质 Download PDF

Info

Publication number
WO2022166113A1
WO2022166113A1 PCT/CN2021/105711 CN2021105711W WO2022166113A1 WO 2022166113 A1 WO2022166113 A1 WO 2022166113A1 CN 2021105711 W CN2021105711 W CN 2021105711W WO 2022166113 A1 WO2022166113 A1 WO 2022166113A1
Authority
WO
WIPO (PCT)
Prior art keywords
calibration
cameras
calibration plate
spatial position
group
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2021/105711
Other languages
English (en)
French (fr)
Inventor
韩亚宁
陈可
黄康
蔚鹏飞
王立平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
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 Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2022166113A1 publication Critical patent/WO2022166113A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration

Definitions

  • the embodiments of the present application relate to the field of photographing devices, for example, to a multi-view camera calibration device, a calibration method, and a storage medium.
  • the behavioral differences before and after drug use in model animals are important indicators for judging drug efficacy.
  • the estimation of animal motion and posture in three-dimensional space has become a hot topic in recent years.
  • the three-dimensional motion and pose estimation of animals is generally realized by multi-view cameras. Because of the large number of model animal species and experimental requirements, different species of animals and different experimental designs usually require different multi-view camera setup schemes. But no matter which kind of multi-view camera erection scheme, the multi-view camera needs to be calibrated.
  • the multi-camera calibration method either obtains the required calibration image by manually moving the calibration plate, or obtains the required calibration image by rotating multiple cameras to shoot the moving calibration plate, but no matter which method, there are The problem of low efficiency of calibration image acquisition.
  • the present application provides a multi-view camera calibration device, a calibration method and a storage medium.
  • the present application provides a multi-view camera calibration device, including:
  • a moving module arranged below the calibration plate, and configured to switch the spatial position state of the calibration plate within the shooting range of a plurality of cameras;
  • a control module connected to the mobile module and the plurality of cameras, is configured to switch the spatial position state of the calibration plate through the mobile module, and controls the plurality of cameras to complete the calibration plate on the mobile module.
  • the spatial position state is switched, the calibration plate is photographed at the same time, and at least two cameras in the plurality of cameras simultaneously capture the calibration image of the complete calibration plate in the corresponding spatial position state, and the calibration is repeatedly switched by the moving module.
  • the spatial position state of the board is controlled, and the plurality of cameras are controlled to shoot the calibration board at the same time when the mobile module completes the switching of the spatial position state of the calibration board.
  • the any two cameras are grouped into a group, until the number of calibration images simultaneously captured by two cameras in each group of cameras in all cameras reaches the target number, according to all cameras.
  • the calibration images captured by each group of cameras in the cameras determine the calibration results of each group of cameras to complete the calibration of all cameras.
  • the present application also provides a method for calibrating a multi-view camera, which is applicable to the multi-view camera calibration device described in any of the embodiments, including:
  • the calibration results of each group of cameras are determined according to the calibration images captured by each group of cameras in all cameras to complete the calibration of all cameras.
  • the present application further provides a storage medium containing computer-executable instructions, when executed by a computer processor, the computer-executable instructions are used to perform the method for calibrating a multi-view camera according to any embodiment.
  • FIG. 1 is a schematic diagram of a multi-view camera calibration device provided in Embodiment 1 of the present application;
  • FIG. 2 is a flowchart of a method for calibrating a multi-view camera provided in Embodiment 2 of the present application;
  • FIG. 3 is a schematic diagram of a computer device provided in Embodiment 3 of the present application.
  • FIG. 1 is a schematic diagram of a multi-view camera calibration device provided in Embodiment 1 of the present application.
  • the technical solution of this embodiment is applicable to the situation of automatically and quickly acquiring a calibration image for calibrating a multi-view camera, and automatically completing the calibration of the multi-view camera according to the acquired calibration image.
  • the device includes a calibration board 1, a mobile module 2 and a control module 3.
  • the mobile module 2 is arranged below the calibration board 1 and is set to switch the spatial position state of the calibration board 1 within the shooting range of the plurality of cameras 4; the control module 3 is connected to The mobile module 2 and the plurality of cameras 4 are set to switch the spatial position state of the calibration plate 1 through the mobile module 2, and control the plurality of cameras 4 to simultaneously perform the calibration plate 1 when the mobile module 2 completes the switching of the spatial position state of the calibration plate 1.
  • Shooting obtain at least two cameras in the plurality of cameras 4 to simultaneously capture the calibration image of the complete calibration plate in the corresponding spatial position state, repeatedly switch the spatial position state of the calibration plate 1 through the moving module 2, and control all the calibration images.
  • the any two cameras are grouped into a group, until the number of calibration images captured by the two cameras in each group of cameras in all cameras at the same time reaches the target number, according to each group of cameras in all cameras.
  • the calibration images captured by the cameras determine the calibration results of each group of cameras, so as to complete the calibration of all cameras.
  • the calibration board may be a flat board carrying a black and white checkerboard, or may be a display device displaying a black and white checkerboard, see Fig. 1 . It can be understood that, if the calibration board is a display device displaying a black and white checkerboard, the size, contrast and position of the black and white checkerboard on the display device can be adjusted, and the adjustment precision is at the pixel level.
  • the calibration image refers to an image that includes a complete calibration plate, or an image that includes at least two feature points that meet preset conditions while including the complete calibration plate.
  • the feature points of the calibration plate in Fig. 1 can be selected as the intersection points of black and white grids.
  • each spatial position state of the calibration plate corresponds to a set of position state data
  • the position state data includes the horizontal position of the calibration plate and the inclination parameters of the calibration plate.
  • the moving module includes a vertical moving unit 21 and a horizontal moving unit 22 for moving the calibration plate 1 and the vertical moving unit 21 .
  • the horizontal moving unit 22 is a four-wheeled vehicle seat, such as a trolley base, and the trolley base realizes the horizontal movement of the calibration plate 1 through the movement of the wheels.
  • the trolley base includes front and rear reverse wheels and left and right drive wheels.
  • the trolley base adopts four-wheel drive wheels as the driving hardware of the trolley base, so that the trolley base has high movement stability.
  • the trolley base adopts a Mecanum wheel as the driving wheel, and the Mecanum wheel can realize the horizontal omnidirectional movement of the trolley base, so that the trolley base has a high movement flexibility.
  • the vertical moving unit 21 includes a first vertical moving mechanism 211 and at least two second vertical moving mechanisms 212, and the vertical moving unit 21 is configured to pass through the first vertical moving mechanism 211 and at least two second vertical moving mechanisms 212.
  • the moving mechanism 212 adjusts the inclination angle of the calibration plate 1
  • the first vertical moving mechanism 211 is arranged to provide a rotation fulcrum for the center of the calibration plate 1
  • the at least two second vertical moving mechanisms 212 are respectively arranged on at least two of the calibration plate 1 .
  • Below the adjacent edges it is set to adjust the height of the corresponding edge of the calibration plate respectively. It can be understood that the inclination angle of the calibration plate can be adjusted by adjusting the heights of the moving ends of the at least two second vertical moving mechanisms respectively.
  • the calibration plate is placed horizontally; if the height of the top of the first vertical moving mechanism is lower than the height of the moving ends of the at least two second vertical moving mechanisms, then the calibration plate is in a tilted state , and the inclination amplitude depends on the difference between the height of the top end of the first vertical moving mechanism and the height of the moving end of any second vertical moving mechanism.
  • the calibration plate can be rotated around the top of the first vertical moving mechanism at any inclination angle. , that is, the calibration plate can be switched from an inclination angle to any inclination angle through the vertical movement unit.
  • the vertical movement unit includes a first vertical movement mechanism and two second vertical movement mechanisms, wherein the first vertical movement mechanism is a damping shaft arranged at the center of the bottom surface of the calibration plate, and is arranged as The rotation of the calibration plate provides a fulcrum, and the two second vertical moving mechanisms are connecting rods arranged on two adjacent edges of the bottom surface of the calibration plate, and each connecting rod is connected with a corresponding servo motor.
  • the link mechanism controlled by the servo motor can make the two small ball heads on the link (see the spherical structure on the second vertical moving mechanism 212 in FIG. 1 ) move up and down.
  • All the structures of the vertical moving unit are fixed to the trolley base, that is, to the horizontal moving unit, so that the horizontal moving unit drives the calibration plate to move horizontally, and also drives the vertical moving unit to move horizontally, and the horizontal moving unit
  • the horizontal movement and the vertical movement of the calibration plate can be simultaneously performed by the vertical movement unit and the calibration plate, thereby improving the switching efficiency of the spatial position state of the calibration plate.
  • the control module adopts an open-loop control mode to control the switching of the inclination angle of the calibration plate, that is, directly sends a preset angle of hitting command to the servo motor, and the servo motor controls the calibration plate through the linkage mechanism according to the preset angle of the command. Move to the inclination angle corresponding to the preset angle command.
  • the control module stores inclination angle control data of the calibration plate, for example, each unit change of the inclination angle of the calibration plate corresponds to the rotation data of the servo motors of the plurality of second vertical moving mechanisms.
  • control module can determine the target inclination angle of the calibration plate and the change amount of the inclination angle between the target inclination angle and the current inclination angle according to the target space position state of the calibration plate, and then according to the change amount of the inclination angle and the corresponding unit change amount
  • the rotation data of the plurality of servo motors control the rotation of the corresponding servo motors, so that the calibration plate rotates from the current tilt angle to the target tilt angle.
  • the vertical movement unit further includes a sensor for sensing the inclination angle of the calibration plate.
  • the vertical moving unit is also configured to adjust the height of the moving ends of the at least two second vertical moving mechanisms according to the current inclination angle of the calibration plate sensed by the sensor and the target inclination angle of the calibration plate, so as to move the calibration plate.
  • the tilt angle is adjusted from the current tilt angle to the target tilt angle.
  • the control module adopts the closed-loop control mode to control the vertical movement of the calibration plate, firstly reads the current value of the sensor, then calculates the difference between the target value and the current value, and converts the difference into the corresponding value through the closed-loop control method.
  • the servo motor cornering data controls the servo motor cornering. Because it is a closed-loop control, the above process will continue to cycle so that the inclination angle of the calibration plate reaches an angle that is less than the allowable error range of the target inclination angle.
  • the closed-loop control algorithm can use a proportional-integral-derivative (Proportion Integral Differential, PID) controller commonly used in the industry, or a more complex control algorithm.
  • the vertical movement unit includes a mechanical arm with a rotation function, and the switching of the inclination angle of the calibration plate is realized by the mechanical arm, so as to improve the switching speed and switching accuracy of the spatial position state of the calibration plate, thereby improving the speed of camera calibration and precision.
  • the device also includes a hardware controller arranged between the moving module and the calibration board, the hardware controller is configured to read the value of the sensor, control the movement of the moving module, supply power to the display device, and control the display mode of the display device.
  • the memory of the hardware controller stores at least one preset standard calibration plate image.
  • the user can operate the control module to select any preset standard calibration plate image through the hardware controller, and control the translation and rotation of the preset standard calibration plate image in the display device, and the translation and rotation are pixel-level. move. It can be understood that the rough horizontal movement of the calibration plate is realized through the horizontal moving unit, and the precise movement of the calibration plate is realized through the display device. The combination of the two can realize the accurate horizontal movement of the calibration plate at any amplitude, and the accuracy of the horizontal movement is pixel. Level, able to meet the needs of any shooting scene.
  • the device also includes a communication module 5 arranged between the horizontal movement unit 22 and the calibration board 1, and the communication module 5 is arranged to realize the communication between the hardware controller and the control module 3.
  • the control module sends a message to the hardware controller.
  • the hardware controller feeds back the execution information of the control instruction for the horizontal movement to the control module.
  • the communication module may use any wireless communication means in the related art to communicate, such as WiFi.
  • the device also includes a control module 3 connected to the mobile module 2 and all the cameras 4, the control module 3 is set to switch the spatial position state of the calibration board 1 through the mobile module 2, and control all the cameras 4 to complete the calibration board 1 in the mobile module 2.
  • the control module 3 is set to switch the spatial position state of the calibration board 1 through the mobile module 2, and control all the cameras 4 to complete the calibration board 1 in the mobile module 2.
  • the spatial position state is switched, the calibration plate 1 is photographed at the same time to obtain the initial image, and then the initial images captured by multiple cameras are detected, so as to eliminate the initial image that only contains part of the calibration plate, and retain the initial image that contains the complete calibration plate. If the number of initial images containing complete calibration plates is less than 2, all initial images of the batch will be deleted directly.
  • the number of initial images containing complete calibration plates is greater than or equal to 2
  • all initial images containing complete calibration plates will be used as calibration image, and simultaneously record the number of calibration images of the complete calibration board captured by any two cameras in the same spatial position.
  • the arbitrary The two cameras are grouped into a group, until the number of calibration images captured by the two cameras in each group of all cameras at the same time reaches the target number, then stop the shooting of all cameras, and then according to each group of cameras in all cameras.
  • the captured calibration images determine the calibration results of each group of cameras to complete the calibration of all cameras, and then analyze the calibration errors corresponding to the calibration results of each group of cameras. If the calibration error of any group of cameras exceeds the corresponding threshold range, output the prompt information of the failure of the calibration of the camera, and control the calibration board to return to zero. and recalibrate all cameras.
  • the control module while taking the initial image including the complete calibration plate as the calibration image, the control module, on the one hand, records the number of calibration images captured by any two cameras in the same spatial position, and on the other hand determines the The internal parameters and external parameters corresponding to the calibration images captured by any two cameras, and when the number of calibration images captured by any two cameras at the same time reaches the target number, the two cameras are grouped into a group, and the two cameras are determined. The calibration result of the camera, and no new calibration images will be added for the two cameras.
  • the control module controls all cameras except the two cameras to continue the calibration when the mobile module completes the switching of the spatial position status of the calibration board. The board is shot until the calibration results of all cameras are determined to complete the calibration of all cameras.
  • the control module when determining the calibration results of any two cameras, the control module also determines the calibration error corresponding to the calibration result. If the calibration error exceeds the corresponding threshold range, the calibration of other groups of cameras may be completed. Then, adjust the spatial position state of the calibration board according to the spatial positions of the two cameras to re-acquire the calibration images of the two cameras, and when the number of the calibration images of the two cameras reaches the target number, re-acquire the calibration images of the two cameras The camera is calibrated until the calibration error is within the corresponding threshold range.
  • the control module can optionally use Zhang's calibration method to extract at least two feature points that meet preset conditions from each calibration image collected by each group of cameras, and calculate corresponding internal parameters according to the at least two feature points. and external parameters, and determine the calibration results of each group of cameras according to all the internal parameters and external parameters corresponding to each group of cameras, so as to obtain the calibration results of all cameras.
  • the calculation method of the internal parameter and the external parameter can be selected as an optimization algorithm, such as an EM algorithm (Expectation-Maximum, expectation maximization algorithm).
  • the feature points of the calibration image are extracted by Zhang's calibration method, which avoids the problem of low quality and weak features of the feature points extracted according to the content of the environment image.
  • the method for determining the calibration error includes: determining the number of iteration steps in the current calibration process of the two cameras, the calculation error, and whether the calibration errors corresponding to the calibration results of the two cameras are all within the corresponding threshold range.
  • determining the number of iteration steps, the calculation error, and the calibration errors corresponding to the calibration results of the two cameras are all within the corresponding threshold range.
  • a prompt message indicating that the current two cameras are successfully calibrated is output;
  • the calibration board is controlled by the mobile module to return to zero, and all cameras are Recalibrate.
  • the EM algorithm can quickly converge when the optimization problem has a single pole, but when the optimization problem has multiple poles, the convergence speed will be reduced, and the calculation error is large, so the number of iteration steps and the calculation error of the EM algorithm can be used as one of the judgement standard.
  • the threshold range for calibration error can be chosen to be less than one pixel wide. Returning the calibration plate to zero means that the calibration plate is transported to a preset initial horizontal position by the horizontal moving unit, and moved to a preset initial vertical position by the vertical moving unit.
  • the method for determining the calibration error includes: the control module extracts, through the extraction unit, feature points at preset positions from each calibration image collected by each group of cameras, and selects, through the selection unit, from the extracted feature points that conform to the preset position At least two feature points of the condition, and then based on Zhang's calibration method, the camera corresponding to one of the calibration images is used as the main camera, and the camera corresponding to the other calibration image is used as the sub-camera, and the corresponding internal parameters are calculated according to the at least two feature points. and external parameters, and complete the calibration of each set of cameras according to all internal parameters and external parameters corresponding to each set of cameras to obtain a calibration result.
  • the control module can be selected as a control computer, and the control computer can output the calibration result, and output the calibration error and corresponding internal parameters and external parameters while outputting the calibration result.
  • control module that is, the control computer
  • the working process of the device includes: initializing all the cameras 4 through the control module 3, for example, checking whether all the cameras can shoot synchronously.
  • the horizontal movement unit 22 is controlled by the control module 3 to roughly move the calibration plate 1 to the center of the shooting site, and the vertical movement unit 21 is controlled to set the inclination angle of the calibration plate 1 to a preset inclination angle, such as a zero-degree inclination angle;
  • Control all cameras 4 to shoot the calibration board 1 at the same time to obtain an initial image, and control the horizontal moving unit to move the calibration board to the center of the shooting scene according to the initial image, and adjust the preset standard calibration board image (checkerboard) in the display device at the same time to correct the position of the calibration board on the shooting site.
  • the spatial position state of the calibration board at this time is taken as the zero point state of the calibration board, and all cameras 4 are activated to shoot the calibration board 1 synchronously, and after the shooting is completed, the movement module 2 is used to control the movement of the calibration board 1 to switch the spatial position of the calibration board. state.
  • the switching of the spatial position state of the calibration board includes two steps, namely horizontal movement and vertical movement.
  • the horizontal movement is controlled by the control module to change the horizontal position of the calibration board through the horizontal movement unit, such as the forward, backward, left and right of the four-wheeled vehicle.
  • the module changes part or all of the vertical position of the calibration plate through the vertical moving unit, so as to realize the change of the inclination angle of the calibration plate. Since the horizontal movement and vertical movement of the calibration plate are independent of each other, the control module can simultaneously complete the horizontal movement and vertical movement of the calibration plate through the horizontal movement unit and the vertical movement unit respectively, so as to shorten the switching of the spatial position state of the calibration plate time, thereby improving the acquisition efficiency of the calibration image.
  • the control module acquires the initial images captured by the multi-view camera in real time, and detects whether each initial image contains a complete calibration plate and feature points that meet the preset conditions. If the number of remaining initial images is less than 2, delete all initial images of this batch, and if the number of remaining initial images is greater than 2 or equal to 2, keep the remaining initial images, and use the remaining initial images as calibration images. When the number of calibration images captured by any two cameras at the same time reaches the target number, use the remaining initial images as calibration images. The cameras are divided into one group until the number of calibration images captured by the two cameras in each group of cameras at the same time reaches the target number.
  • Internal parameters and external parameters and determine the calibration results of each group of cameras according to all the internal parameters and external parameters of each group of cameras, thereby obtaining the calibration results of each group of cameras in all cameras, and the calibration results corresponding to the calibration error , and the calibration error is output graphically.
  • the switching of the spatial position state of the calibration board can be quickly completed through the cooperative use of the control module and the mobile module.
  • the Complete the switching of the spatial position status of the calibration board more accurately and quickly, and start the shooting of all cameras when the switching is completed, and count whether the number of calibration images captured by each group of cameras in all cameras at the same time reaches the target number.
  • the number of targets stops the switching of the spatial position status of the calibration board and the shooting of all cameras, and performs calibration processing on the calibration images.
  • the calibration of multi-view cameras can be made simpler, faster and more efficient. Calibration requirements of multi-view cameras for different shooting scenes.
  • Embodiment 2 is a multi-view camera calibration method provided in Embodiment 2 of the present application, which is suitable for the multi-view camera calibration device described in the foregoing embodiments.
  • the method can be implemented in software or hardware, including steps S201 to S203.
  • the calibration board 1 may be a flat board carrying a black and white checkerboard, or may be a display device displaying a black and white checkerboard, see FIG. 1 . It can be understood that, if the calibration board is a display device displaying a black and white checkerboard, the size, contrast and position of the black and white checkerboard on the display device can be adjusted, and the adjustment precision is at the pixel level.
  • the memory of the hardware controller connected to the display device stores at least two preset standard black and white checkerboards, and the user can select a suitable preset standard from the memory of the hardware controller through the control module as required Black and white checkerboard.
  • each spatial position state of the calibration plate corresponds to a set of position state data
  • the position state data includes the horizontal position of the calibration plate and the inclination parameters of the calibration plate.
  • the moving module 2 includes a vertical moving unit 21 and a horizontal moving unit 22 for moving the calibration plate 1 and the vertical moving unit 21 .
  • the horizontal moving unit is a four-wheeled vehicle seat, such as a trolley base, which realizes the horizontal movement of the calibration plate through the movement of the wheels.
  • the trolley base includes front and rear reverse wheels and left and right drive wheels.
  • the trolley base adopts four-wheel drive wheels as the driving hardware of the trolley base, so that the trolley base has high movement stability.
  • the trolley base adopts a Mecanum wheel as the driving wheel, and the Mecanum wheel can realize the horizontal omnidirectional movement of the trolley base, so that the trolley base has a high movement flexibility.
  • the vertical movement unit 21 is configured to adjust the inclination angle of the calibration plate 1 through a first vertical movement mechanism 211 and at least two second vertical movement mechanisms 212 , and the first vertical movement mechanism 211 is set to be the center of the calibration plate 1 .
  • a rotation fulcrum is provided, and the at least two second vertical moving mechanisms 212 are respectively disposed below at least two adjacent edges of the calibration plate 1 to adjust the heights of the corresponding edges of the calibration plate respectively. It can be understood that the inclination angle of the calibration plate can be adjusted by adjusting the heights of the moving ends of the at least two second vertical moving mechanisms respectively.
  • the calibration plate is placed horizontally; if the height of the top of the first vertical moving mechanism is lower than the height of the moving ends of the at least two second vertical moving mechanisms, then the calibration plate is in a tilted state , and the inclination amplitude depends on the difference between the height of the top end of the first vertical moving mechanism and the height of the moving end of any second vertical moving mechanism.
  • the calibration plate can be rotated around the top of the first vertical moving mechanism at any inclination angle. , that is, the calibration plate can be switched from an inclination angle to any inclination angle through the vertical movement unit.
  • the vertical movement unit includes a first vertical movement mechanism and two second vertical movement mechanisms, wherein the first vertical movement mechanism is a damping shaft arranged at the center of the bottom surface of the calibration plate, and is arranged as The rotation of the calibration plate provides a fulcrum, and the two second vertical moving mechanisms are connecting rods arranged on two adjacent edges of the bottom surface of the calibration plate, and each connecting rod is connected with a corresponding servo motor.
  • the linkage mechanism controlled by the servo motor can make the two small balls on the linkage (see Figure 1) move up and down.
  • All the structures of the vertical moving unit are fixed to the trolley base, that is, to the horizontal moving unit, so that the horizontal moving unit drives the calibration plate to move horizontally, and also drives the vertical moving unit to move horizontally, and the horizontal moving unit
  • the horizontal movement and the vertical movement of the calibration plate can be simultaneously performed by the vertical movement unit and the calibration plate, thereby improving the switching efficiency of the spatial position state of the calibration plate.
  • the vertical moving unit further includes a sensor for sensing the inclination angle of the calibration plate.
  • the vertical moving unit is also configured to adjust the height of the moving ends of the at least two second vertical moving mechanisms according to the current inclination angle of the calibration plate sensed by the sensor and the target inclination angle of the calibration plate, so as to move the calibration plate.
  • the tilt angle is adjusted from the current tilt angle to the target tilt angle.
  • the control module adopts an open-loop control mode to control the switching of the inclination angle of the calibration plate, that is, directly sends a preset angle of hitting command to the servo motor, and the servo motor controls the calibration plate through the linkage mechanism according to the preset angle of the command. Move to the inclination angle corresponding to the preset angle command.
  • the control module stores inclination angle control data of the calibration plate, for example, each unit change of the inclination angle of the calibration plate corresponds to the rotation data of the servo motors of the plurality of second vertical moving mechanisms.
  • control module can determine the target inclination angle of the calibration plate and the change amount of the inclination angle between the target inclination angle and the current inclination angle according to the target space position state of the calibration plate, and then according to the change amount of the inclination angle and the corresponding unit change amount
  • the rotation data of the plurality of servo motors control the rotation of the corresponding servo motors, so that the calibration plate rotates from the current tilt angle to the target tilt angle.
  • the control module uses a closed-loop control mode to control the vertical movement of the calibration plate.
  • the current value of the sensor is read, and then the difference between the target value and the current value is calculated, and the difference is converted into a closed-loop control method.
  • the corresponding servo motor cornering data controls the servo motor cornering. Because it is a closed-loop control, the above process will continue to cycle so that the inclination angle of the calibration plate reaches an angle that is less than the allowable error range of the target inclination angle.
  • the closed-loop control algorithm can use the PID hardware controller commonly used in the industry, and can also use a more complex control algorithm.
  • the vertical movement unit includes a mechanical arm with a rotation function, and the switching of the inclination angle of the calibration plate is realized by the mechanical arm, so as to improve the switching speed and switching accuracy of the spatial position state of the calibration plate, thereby improving the speed of camera calibration and precision.
  • the calibration image refers to an image that includes a complete calibration plate, or an image that includes at least two feature points that meet preset conditions while including the complete calibration plate.
  • the feature points of the calibration plate in Figure 1 refer to the intersections of the black and white grids.
  • control module 3 is set to switch the spatial position state of the calibration board 1 through the moving module 2, and control all the cameras 4 to photograph the calibration board 1 at the same time when the moving module 2 completes the switching of the spatial position state of the calibration board 1 to obtain an initial image , and then detect the initial images captured by the multiple cameras 4, so as to remove the initial images that only contain part of the calibration board, and retain the initial images that contain the complete calibration board. If the number of initial images containing the complete calibration board is less than 2, delete it directly For all initial images of this batch, if the number of initial images containing complete calibration plates is greater than or equal to 2, all initial images containing complete calibration plates will be used as calibration images, and the state of the same spatial position captured by any two cameras is recorded at the same time. the number of calibration images of the complete calibration board under the When the number of calibration images captured by two cameras in the camera reaches the target number at the same time, the shooting of all cameras is stopped.
  • the control module while taking the initial image including the complete calibration plate as the calibration image, the control module, on the one hand, records the number of calibration images captured by any two cameras in the same spatial position, and on the other hand determines the The internal parameters and external parameters corresponding to the calibration images captured by any two cameras, and when the number of calibration images captured by any two cameras at the same time reaches the target number, no new calibration images are added for the two cameras. It can be understood that, if the calibration results of any two cameras have been determined, the spatial positions of the two cameras need not be considered when adjusting the spatial position state of the calibration plate.
  • S203 Determine the calibration result of each group of cameras according to the calibration images captured by each group of cameras in all cameras to complete the calibration of all cameras.
  • the calibration results of each group of cameras are determined according to the calibration images captured by each group of cameras in all cameras, and all groups of cameras in all cameras are separately calibrated to complete the calibration of all cameras, and then the calibration results of all groups of cameras are analyzed.
  • Corresponding calibration error if the calibration errors corresponding to the calibration results of all groups of cameras are within the corresponding threshold range, then output the prompt information of successful calibration of all cameras, if the calibration results of any group of cameras correspond to the calibration error beyond the corresponding threshold range , then output the prompt information of the failure of the calibration of the group of cameras, control the calibration board to zero, and re-calibrate all the cameras.
  • the control module when determining the calibration results of any group of cameras, the control module also determines the calibration errors corresponding to the calibration results. If the calibration errors exceed the corresponding threshold range, the calibration of the cameras in other groups may be completed. After that, adjust the spatial position state of the calibration plate according to the spatial position of the group of cameras, so as to re-acquire the calibration images of the group of cameras, and when the number of the calibration images of the group of cameras reaches the target number, re-calibrate the two cameras, until the calibration error is within the corresponding threshold range.
  • the control module can optionally use Zhang's calibration method to extract at least two feature points that meet preset conditions from each calibration image collected by each group of cameras, and calculate corresponding internal parameters according to the at least two feature points. and external parameters, and determine the calibration results of each group of cameras according to all internal parameters and external parameters corresponding to each group of cameras, so as to obtain the calibration results of all cameras.
  • the calculation method of the internal parameters and the external parameters can be selected as an optimization algorithm, such as an EM algorithm.
  • the feature points of the calibration image are extracted by Zhang's calibration method, which avoids the problem of low quality and weak features of the feature points extracted according to the content of the environment image.
  • the method for determining the calibration error includes: determining the number of iteration steps and calculation errors of the EM algorithm in the current calibration process of the two cameras, and whether the calibration errors corresponding to the calibration results of the two cameras are all within the corresponding threshold range. When the number of iteration steps, the calculation error, and the calibration errors corresponding to the calibration results of the two cameras are all within the corresponding threshold range, a prompt message indicating that the current two cameras are successfully calibrated is output; When at least one of the iterative steps in the calibration process, the calculation error, and the calibration error corresponding to the calibration results of the two cameras is not within the corresponding threshold range, the calibration board is controlled by the mobile module to return to zero, and all cameras are Recalibrate.
  • the EM algorithm can quickly converge when the optimization problem has a single pole, but when the optimization problem has multiple poles, the convergence speed will be reduced, and the calculation error is large, so the number of iteration steps and the calculation error of the EM algorithm can be used as one of the judgement standard.
  • the threshold range for calibration error can be chosen to be less than one pixel wide. Returning the calibration plate to zero means that the calibration plate is transported to a preset initial horizontal position by the horizontal moving unit, and moved to a preset initial vertical position by the vertical moving unit.
  • the method for determining the calibration error includes: the control module extracts, through the extraction unit, feature points at preset positions from each calibration image collected by each group of cameras, and selects, through the selection unit, from the extracted feature points that conform to the preset position At least two feature points of the condition, and then based on Zhang's calibration method, the camera corresponding to one of the calibration images is used as the main camera, and the camera corresponding to the other calibration image is used as the sub-camera, and the corresponding internal parameters are calculated according to the at least two feature points. and external parameters, and complete the calibration of each set of cameras according to all internal parameters and external parameters corresponding to each set of cameras to obtain a calibration result.
  • the control module can be selected as a control computer, and the control computer can output the calibration result, and output the calibration error and corresponding internal parameters and external parameters while outputting the calibration result.
  • control module that is, the control computer
  • this embodiment does not limit the calibration processing timing of the calibration images, that is, when the calibration images corresponding to all groups of cameras reach the target number, batch processing is performed on the calibration images of each group of cameras to obtain the calibration images of each group of cameras. Calibration results; also in the acquisition of calibration images of multiple cameras, once it is detected that the number of calibration images of any two cameras reaches the target number, the calibration images of the two cameras are calibrated to obtain the two cameras. Calibration results.
  • the former has lower requirements on the parallel processing capability of the computer, but when the calibration of any two cameras fails, the calibration process of all cameras needs to be restarted. Calibrate the images, and perform calibration processing on the obtained calibration images until the two cameras are successfully calibrated. When the calibration of any two cameras fails, there is no need to restart the calibration process for all cameras, and the calibration process can be simplified when any two cameras fail to be calibrated. , although the parallel computing capability of the control module is higher than that of the former.
  • the switching of the spatial position state of the calibration board can be quickly completed through the cooperative use of the control module and the mobile module.
  • the Complete the switching of the spatial position status of the calibration board more accurately and quickly, and start the shooting of all cameras when the switching is completed, and count whether the number of calibration images captured by any group of cameras at the same time reaches the target number, and once the target is reached
  • the number of the calibration board stops the switching of the spatial position state of the calibration board and the shooting of all cameras, and the calibration image is calibrated.
  • the calibration of the multi-view camera can be simpler, faster and more efficient, and can meet the needs of different Calibration requirements for multi-view cameras that capture scenes.
  • FIG. 3 is a schematic structural diagram of a computer device provided in Embodiment 3 of the present application.
  • the device is a control module in the foregoing embodiment, including a processor 301, a memory 302, an input device 303, and an output device 304; the device
  • the number of processors 301 in the device can be at least one, and one processor 301 is taken as an example in FIG. 3 ; the processor 301, the memory 302, the input device 303, and the output device 304 in the device can be connected through a bus or other means, as shown in FIG. 3 Take connection via bus as an example.
  • the memory 302 may be configured to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the calibration method for a multi-view camera in the embodiments of the present application.
  • the processor 301 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 302 , that is, to implement the above-mentioned calibration method for a multi-view camera.
  • the memory 302 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Additionally, memory 302 may include high speed random access memory, and may also include nonvolatile memory, such as at least one magnetic disk storage device, flash memory device, or other nonvolatile solid state storage device. In some examples, memory 302 may include memory located remotely from processor 301, which may be connected to the device through a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
  • the input device 303 may be configured to receive input numerical or character information, and to generate key signal input related to user settings and function control of the device.
  • the output device 304 may include a display device such as a display screen, for example, a display screen of a user terminal.
  • Embodiments of the present application also provide a storage medium containing computer-executable instructions, when the computer-executable instructions are executed by a computer processor for executing a method for calibrating a multi-view camera, the method includes:
  • the calibration image of the complete calibration board repeatedly switching the spatial position state of the calibration board through the mobile module, and controlling the plurality of cameras to simultaneously monitor the spatial position state of the calibration board when the mobile module completes the switching of the spatial position state of the calibration board.
  • the step of shooting by the calibration board when the number of calibration images captured by any two cameras at the same time reaches the target number, the any two cameras are grouped into a group, until two cameras in each group of cameras in all cameras. The number of calibration images captured by each camera at the same time reaches the target number;
  • the calibration results of each group of cameras are determined according to the calibration images captured by each group of cameras in all cameras to complete the calibration of all cameras.
  • a storage medium containing computer-executable instructions provided by the embodiments of the present application, the computer-executable instructions of which are not limited to the above-mentioned method operations, and can also perform the calibration of the multi-view camera provided by any embodiment of the present application. related operations in the method.
  • the present application can be implemented by means of software and necessary general-purpose hardware, and certainly can also be implemented by hardware.
  • the technical solutions of the present application can be embodied in the form of software products in essence or the parts that make contributions to related technologies, and the computer software products can be stored in a computer-readable storage medium, such as a computer floppy disk, Read-Only Memory (ROM for short), Random Access Memory (RAM for short), Flash Memory (FLASH), hard disk or CD, etc., including several instructions to make a computer device (which can be a personal A computer, a server, or a network device, etc.) executes the multi-view camera calibration method described in the various embodiments of the present application.
  • a computer device which can be a personal A computer, a server, or a network device, etc.
  • the included units and modules are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be realized;
  • the names of the functional units are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Studio Devices (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

本申请实施例公开了一种多视角摄像头标定设备、标定方法及存储介质，该设备包括：标定板；移动模块，设置于标定板的下方，设置为在多个摄像头的拍摄范围内切换标定板的空间位置状态；控制模块，连接移动模块和多个摄像头，设置为通过移动模块切换标定板的空间位置状态，并控制多个摄像头在移动模块完成标定板的空间位置状态切换时同时对标定板进行拍摄，以使至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，根据所有摄像头中每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果，以完成所有摄像头的标定。

Description

多视角摄像头标定设备、标定方法及存储介质
本申请要求在2021年2月8日提交中国专利局、申请号为202110181621.4的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及拍摄设备领域，例如涉及一种多视角摄像头标定设备、标定方法及存储介质。
背景技术
在神经精神类药物的研发方面，模式动物用药前后的行为差异是判断药效的重要指标。为了给药物开发提供更为精细的动物行为学参数，近年来对动物在三维空间中的运动以及姿态的估计成为热点。动物三维运动姿态估计一般采用多视角摄像头实现。因为模式动物种类数量和实验要求繁多，所以不同物种的动物以及不同的实验设计通常需要采用不同的多视角摄像头架设方案。但无论哪一种多视角摄像头架设方案，均需要对多视角摄像头进行标定。
多摄像头标定方法要么通过人工移动标定板的方式来获取所需的标定图像，要么通过旋转多个摄像头来拍摄移动标定板的方式来获取所需的标定图像，但无论哪一种方法，均存在标定图像获取效率较低的问题。
发明内容
本申请提供了一种多视角摄像头标定设备、标定方法及存储介质。
第一方面，本申请提供了一种多视角摄像头标定设备，包括：
标定板；
移动模块，设置于所述标定板的下方，设置为在多个摄像头的拍摄范围内切换所述标定板的空间位置状态；
控制模块，连接所述移动模块和所述多个摄像头，设置为通过所述移动模 块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄，得到多个摄像头中的至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果以完成所有摄像头的标定。
第二方面，本申请还提供了一种多视角摄像头的标定方法，适应于任意实施例所述的多视角摄像头标定设备包括：
通过移动模块切换所述标定板的空间位置状态；
控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄，以使至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量；
根据所有摄像头中每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果以完成所有摄像头的标定。
第三方面，本申请还提供了一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行如任意实施例所述的多视角摄像头的标定方法。
附图说明
图1是本申请实施例一提供的多视角摄像头标定设备的示意图；
图2是本申请实施例二提供的多视角摄像头的标定方法的流程图；
图3是本申请实施例三提供的计算机设备的示意图。
具体实施方式
以下将参照本申请实施例中的附图，通过实施方式清楚、完整地描述本申请的技术方案。
实施例一
图1是本申请实施例一提供的多视角摄像头标定设备的示意图。本实施例的技术方案适用于自动快速地获取用于标定多视角摄像头的标定图像并根据获取的标定图像自动完成多视角摄像头标定的情况。该设备包括标定板1、移动模块2和控制模块3，移动模块2设置于标定板1的下方，设置为在多个摄像头4的拍摄范围内切换标定板1的空间位置状态；控制模块3连接移动模块2和多个摄像头4，设置为通过移动模块2切换标定板1的空间位置状态，并控制多个摄像头4在移动模块2完成标定板1的空间位置状态切换时同时对标定板1进行拍摄，得到多个摄像头4中的至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块2切换所述标定板1的空间位置状态，并控制所述多个摄像头4在所述移动模块2完成所述标定板1的空间位置状态切换时同时对所述标定板1进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果，以完成所有摄像头的标定。
其中，标定板可以是携带有黑白棋盘格的平面板，也可以是显示有黑白棋 盘格的显示装置，参见图1。可以理解的是，如果标定板为显示有黑白棋盘格的显示装置，那么黑白棋盘格在显示装置上的大小、对比度和位置均是可调的，且调节精度为像素级。
其中，标定图像是指包含完整标定板的图像，或者在包含完整标定板的同时，还包含符合预设条件的至少两个特征点的图像。图1中的标定板的特征点可选为黑白格的交点。
其中，标定板的每个空间位置状态均对应一组位置状态数据，该位置状态数据包括标定板的水平位置和标定板的倾斜参数。
其中，移动模块包括竖直移动单元21以及用于移动标定板1和该竖直移动单元21的水平移动单元22。水平移动单元22为四轮车座，比如小车底座，该小车底座通过轮子的移动实现标定板1的水平移动。在一个实施例中，小车底座包括前后倒向轮和左右驱动轮。在又一实施例中，小车底座采用四轮驱动轮作为小车底座的驱动硬件，从而使小车底座具有较高的移动平稳性。在又一实施例中，小车底座采用麦克纳姆轮作为驱动轮，麦克纳姆轮可以实现小车底座的水平全向移动，使得小车底座具有较高的移动灵活性。
其中，竖直移动单元21包括第一竖直移动机构211和至少两个第二竖直移动机构212，竖直移动单元21设置为通过第一竖直移动机构211和至少两个第二竖直移动机构212调整标定板1的倾斜角度，第一竖直移动机构211设置为为标定板1中心提供旋转支点，该至少两个第二竖直移动机构212分别设置于标定板1的至少两个相邻边缘的下方，设置为分别调整标定板相应边缘的高度。可以理解的是，通过分别调整该至少两个第二竖直移动机构的移动端的高度即可调整标定板的倾斜角度，比如，如果第一竖直移动机构的顶端和该至少两个第二竖直移动机构的移动端的高度相同，那么标定板被水平放置；如果第一竖直移动机构的顶端的高度低于该至少两个第二竖直移动机构的移动端的高度，那么标定板处于倾斜状态，且倾斜幅度取决于第一竖直移动机构的顶端的高度与任一第二竖直移动机构的移动端的高度的差值。示例性的，通过调整该至少 两个第二竖直移动机构中的至少一个第二竖直移动机构的移动端的高度，可以使标定板以任一倾斜角度绕第一竖直移动机构的顶端旋转，即通过竖直移动单元可将标定板从一个倾斜角度切换至任意倾斜角度。
在一个实施例中，竖直移动单元包括一个第一竖直移动机构和两个第二竖直移动机构，其中，第一竖直移动机构为设置于标定板底面中心的阻尼转轴，设置为为标定板旋转提供支点，两个第二竖直移动机构为设置于标定板底面的两相邻边缘的连杆，且每个连杆均与对应伺服电机连接。伺服电机控制的连杆机构可以使得连杆上的两个小球头(参见图1中的第二竖直移动机构212上的球状结构)上下移动。竖直移动单元的所有结构均被固定至小车底座上，即被固定至水平移动单元，从而使水平移动单元在带动标定板水平移动的同时，也带动竖直移动单元水平移动，且水平移动单元和竖直移动单元可以分别同时执行标定板的水平移动和竖直移动，从而提高标定板空间位置状态的切换效率。
在一个实施例中，控制模块采用开环控制模式控制标定板的倾斜角的切换，即直接发送预设打角指令给伺服电机，伺服电机根据该预设打角指令通过连杆机构控制标定板移动至预设打角指令对应的倾斜角。示例性地，控制模块中存储有标定板倾斜角控制数据，比如，标定板的倾斜角的每个单位改变量均对应有多个第二竖直移动机构的伺服电机的旋转数据。这样控制模块既可根据标定板的目标空间位置状态确定标定板的目标倾斜角，以及目标倾斜角与当前倾斜角之间的倾斜角度改变量，然后根据该倾斜角度改变量和相应单位改变量对应的多个伺服电机的旋转数据控制对应伺服电机旋转，以使标定板从当前倾斜角旋转至目标倾斜角。
在一个实施例中，竖直移动单元还包括用于感应标定板倾斜角度的传感器。相应的，竖直移动单元还设置为根据传感器感应到的标定板的当前倾斜角角度与标定板的目标倾斜角角度调整该至少两个第二竖直移动机构的移动端的高度，以将标定板的倾斜角从当前倾斜角调整至目标倾斜角。示例性地，控制模块采用闭环控制模式控制标定板的竖直移动，首先读取传感器的当前数值，然后计 算目标数值与当前数值的差值，并通过闭环控制方法将该差值转换为对应的伺服电机打角数据控制伺服电机打角。因为是闭环控制，上述过程会不断循环从而使得标定板的倾斜角达到与目标倾斜角小于容许误差范围内的打角。闭环控制算法可以采用工业常用的比例-积分-微分(Proportion Integral Differential，PID)控制器，也可采用更为复杂的控制算法。
在一个实施例中，竖直移动单元包括具有旋转功能的机械臂，通过该机械臂实现标定板倾斜角的切换，以提高标定板空间位置状态的切换速度和切换精度，从而提高摄像头标定的速度和精度。
可以理解的是，通过移动模块来移动标定板，而不是移动摄像头，使得摄像头的架设更加灵活，不需要额外增加摄像头的硬件成本。
该设备还包括设置于移动模块与标定板之间的硬件控制器，该硬件控制器设置为读取传感器的数值、控制移动模块移动、为显示装置供电以及控制显示装置的显示方式。该硬件控制器的存储器存储有至少一个预设的标准标定板图像。而且用户可以操作控制模块通过该硬件控制器选择任一预设的标准标定板图像，以及控制该预设的标准标定板图像在显示装置中的平移和旋转，且该平移和旋转为像素级别的移动。可以理解的是，通过水平移动单元实现标定板的粗略水平移动，通过显示装置实现标定板的精确移动，二者结合可以实现标定板任一幅度的精确水平移动，且该水平移动的精度为像素级别，能够满足任何拍摄场景的需求。
该设备还包括设置于水平移动单元22与标定板1之间的通信模块5，该通信模块5设置为实现硬件控制器与控制模块3之间的通信，比如，控制模块向硬件控制器发送用于控制水平移动单元水平移动标定板的控制指令，硬件控制器向控制模块反馈水平移动的控制指令的执行信息。该通信模块采用任一相关技术中的无线通信手段进行通信即可，比如WiFi。
该设备还包括连接移动模块2和所有摄像头4的控制模块3，该控制模块3设置为通过移动模块2切换标定板1的空间位置状态，并控制所有摄像头4 在移动模块2完成标定板1的空间位置状态切换时同时对标定板1进行拍摄以得到初始图像，然后对多个摄像头拍摄的初始图像进行检测，从而剔除仅包含部分标定板的初始图像，保留包含完整标定板的初始图像，如果包含完整标定板的初始图像的数量小于2，则直接删除该批次的所有初始图像，如果包含完整标定板的初始图像的数量大于或等于2，则将所有包含完整标定板的初始图像作为标定图像，同时记录任意两个摄像头均拍摄到相同空间位置状态下的完整标定板的标定图像的数量，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，则停止所有摄像头的拍摄，然后根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果以完成所有摄像头的标定，再分析每组摄像头的标定结果对应的标定误差，若所有摄像头中每组摄像头的标定结果对应的标定误差均在相应阈值范围内，则输出所有摄像头标定成功的提示信息，若任意一组摄像头的标定结果对应的标定误差超出相应的阈值范围，则输出该组摄像头标定失败的提示信息，同时控制标定板归零，并对所有摄像头重新进行标定。
在一个实施例中，控制模块在将包含完整标定板的初始图像作为标定图像的同时，一方面记录任意两个摄像头均拍摄到相同空间位置状态下的标定图像的数量，另一方面确定所述任意两个摄像头拍摄的标定图像对应的内参数和外参数，并在任意两个摄像头同时拍摄到的标定图像的数量均达到目标数量时，将该两个摄像头分为一组，确定该两个摄像头的标定结果，同时不再为该两个摄像头添加新的标定图像，控制模块控制所有摄像头中除该两个摄像头之外的摄像头在移动模块完成标定板的空间位置状态切换时同时继续对标定板进行拍摄，直至确定完所有摄像头的标定结果以完成所有摄像头的标定。可以理解的是，如果任意两个摄像头的标定结果已确定，则在调整标定板的空间位置状态时，无需考虑该两摄像头的空间位置。在又一实施例中，控制模块在确定任意 两个摄像头的标定结果的同时，还确定该标定结果对应的标定误差，如果该标定误差超出相应的阈值范围，则可选在其他组摄像头标定完成后，根据该两个摄像头的空间位置调整标定板的空间位置状态，以重新获取该两个摄像头的标定图像，以及在该两个摄像头的标定图像的数量达到目标数量时，重新对该两个摄像头进行标定，直至标定误差在相应的阈值范围内。
本实施例中，控制模块可选采用张氏标定法从每组摄像头采集的每幅标定图像上提取符合预设条件的至少两个特征点，以及根据该至少两个特征点计算对应的内参数和外参数，并根据所述每组摄像头对应的所有内参数和外参数确定所述每组摄像头的标定结果，从而得到所有摄像头的标定结果。其中，内参数与外参数的计算方法可选为优化算法，比如EM算法(Expectation-Maximum，期望最大化算法)。采用张氏标定法提取标定图像的特征点，避免了根据环境图像内容提取特征点存在的特征点质量不高，特征弱的问题。
其中，标定误差的确定方法包括：确定当前两个摄像头标定过程中的迭代步数、计算误差，以及该两个摄像头的标定结果对应的标定误差是否均在相应的阈值范围内，在当前两个摄像头标定过程中的迭代步数、计算误差，以及该两个摄像头的标定结果对应的标定误差均在相应阈值范围内的情况下，输出当前两个摄像头标定成功的提示信息；在当前两个摄像头标定过程中的迭代步数、计算误差，以及该两个摄像头的标定结果对应的标定误差中的至少一个不在相应的阈值范围内的情况下，通过移动模块控制标定板归零，并对所有摄像头重新进行标定。其中，EM算法在优化问题存在单极点时，能够迅速收敛，而在优化问题存在多极点时收敛速度会降低，并且计算误差较大，因此可将EM算法的迭代步数和计算误差作为其中一个判定标准。标定误差的阈值范围可选为小于一个像素宽度。标定板归零，是指标定板被水平移动单元运送至预设起始水平位置，以及被竖直移动单元移动至预设初始竖直位置。
在一个实施例中，标定误差的确定方法包括：控制模块通过提取单元从每组摄像头采集的每幅标定图像上提取预设位置的特征点，通过选取单元从提取 的特征点中选择符合预设条件的至少两个特征点，然后基于张氏标定法将其中一标定图像对应摄像头作为主摄像头，将另一标定图像对应的摄像头作为副摄像头，并根据该至少两个特征点计算对应的内参数和外参数，并根据每组摄像头对应的所有内参数和外参数完成所述每组摄像头的标定以得到标定结果。标定结果确定后，对所有符合预设条件的特征点进行反投影计算，即通过标定的内参数和外参数将对应两摄像头拍摄到的特征点分别投影到相应摄像头所对应的标定板上以得到两个反投影点，反投影点与真实拍摄到的标定板的特征点之间会存在一定的误差，通过欧式距离量化所有特征点与反投影点之间的距离以作为量化的标定误差。其中，控制模块可选为控制计算机，控制计算机可输出标定结果，且在输出标定结果的同时还输出标定误差以及对应的内参数和外参数。
在一个实施例中，控制模块即控制计算机能够以条形图的形式显示标定误差，以及在三维空间中显示计算得到的每组摄像头的内参数和外参数，即所有摄像头在真实三维空间中的位置。
该设备的工作过程包括：通过控制模块3对所有摄像头4进行初始化，比如，检查所有摄像头能否进行同步拍摄。初始化完成后，通过控制模块3控制水平移动单元22将标定板1粗略移动至拍摄场地的中心，控制竖直移动单元21将标定板1的倾斜角设置为预设倾斜角，比如零度倾斜角；控制所有摄像头4同时拍摄标定板1以得到初始图像，并根据初始图像控制水平移动单元将标定板移动至拍摄场景的中心位置，同时调整预设的标准标定板图像(棋盘格)在显示装置中的显示位置，以修正标定板在拍摄场地的位置。将标定板此时的空间位置状态作为标定板的零点状态，并启动所有摄像头4对标定板1进行同步拍摄，并在拍摄完毕后通过移动模块2控制标定板1移动以切换标定板的空间位置状态。其中，标定板空间位置状态切换包括两个步骤，分别是水平移动和竖直移动，其中，水平移动由控制模块通过水平移动单元改变标定板的水平位置，比如四轮车的前进、后退、左转、右转，以及通过控制预设标定图像在 显示装置中的显示位置和状态来修正标定板的水平位置，比如，标定板在显示装置上的平移和旋转等；其中，竖直移动由控制模块通过竖直移动单元改变标定板的部分或全部的竖直位置，从而实现标定板倾斜角的变化。由于标定板的水平移动和竖直移动是相互独立的，因此控制模块可以分别通过水平移动单元和竖直移动单元同时完成标定板的水平移动和竖直移动，以缩短标定板空间位置状态的切换时间，从而提高标定图像的获取效率。控制模块实时获取多视角摄像头拍摄的初始图像，并检测每幅初始图像是否包含完整的标定板以及符合预设条件的特征点，在初始图像不包含完整的标定板以及符合预设条件的特征点的情况下，剔除相应初始图像；确定剩余的初始图像的数量是否小于2，在剩余的初始图像的数量小于2的情况下，删除此批次的所有初始图像，在剩余的初始图像的数量大于或等于2的情况下，保留剩余的初始图像，并将该剩余的初始图像作为标定图像，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，然后根据提取的标定图像的特征点批量确定所有摄像头中每组摄像头摄像头的内参数和外参数；以及根据所述每组摄像头的所有内参数和外参数确定所述每组摄像头的标定结果，从而得到所有摄像头中每组摄像头的的标定结果，以及标定结果对应的标定误差，同时将该标定误差以图示化形式输出。
本申请实施例提供的多视角摄像头标定设备的技术方案，通过控制模块与移动模块的配合使用可以快速完成标定板空间位置状态的切换，相较于手动变换标定板的空间位置状态来说，可以更加准确快速地完成标定板的空间位置状态的切换，并在切换完成的同时启动所有摄像头的拍摄，以及统计所有摄像头中的每组摄像头同时拍摄到的标定图像的数量是否达到目标数量，一旦达到目标数量则停止标定板空间位置状态的切换和所有摄像头的拍摄工作，并对标定图像进行标定处理，相较于相关技术来说，可使多视角摄像头的标定更加简单、快捷、高效，能够满足不同拍摄场景的多视角摄像头的标定需求。
实施例二
图2是本申请实施例二提供的多视角摄像头的标定方法，适应于前述实施例所述的多视角摄像头标定设备，该方法可采用软件或硬件的方式实现，包括步骤S201至S203。
S201、通过移动模块切换标定板的空间位置状态。
其中，标定板1可以是携带有黑白棋盘格的平面板，也可以是显示有黑白棋盘格的显示装置，参见图1。可以理解的是，如果标定板为显示有黑白棋盘格的显示装置，那么黑白棋盘格在显示装置上的大小、对比度和位置均是可调的，且调节精度为像素级。在一个实施例中，连接该显示装置的硬件控制器的存储器存储有至少两个预设的标准黑白棋盘格，用户可以根据需要通过控制模块从硬件控制器的存储器中选择合适的预设的标准黑白棋盘格。
其中，标定板的每个空间位置状态均对应一组位置状态数据，该位置状态数据包括标定板的水平位置和标定板的倾斜参数。
如图1所示，移动模块2包括竖直移动单元21以及用于移动标定板1和该竖直移动单元21的水平移动单元22。水平移动单元为四轮车座，比如小车底座，该小车底座通过轮子的移动实现标定板的水平移动。在一个实施例中，小车底座包括前后倒向轮和左右驱动轮。在又一实施例中，小车底座采用四轮驱动轮作为小车底座的驱动硬件，从而使小车底座具有较高的移动平稳性。在又一实施例中，小车底座采用麦克纳姆轮作为驱动轮，麦克纳姆轮可以实现小车底座的水平全向移动，使得小车底座具有较高的移动灵活性。
其中，竖直移动单元21设置为通过第一竖直移动机构211和至少两个第二竖直移动机构212调整标定板1的倾斜角度，第一竖直移动机构211设置为为标定板1中心提供旋转支点，该至少两个第二竖直移动机构212分别设置于标定板1的至少两个相邻边缘的下方，设置为分别调整标定板相应边缘的高度。可以理解的是，通过分别调整该至少两个第二竖直移动机构的移动端的高度即 可调整标定板的倾斜角度，比如，如果第一竖直移动机构的顶端和该至少两个第二竖直移动机构的移动端的高度相同，那么标定板被水平放置；如果第一竖直移动机构的顶端的高度低于该至少两个第二竖直移动机构的移动端的高度，那么标定板处于倾斜状态，且倾斜幅度取决于第一竖直移动机构的顶端的高度与任一第二竖直移动机构的移动端的高度的差值。示例性的，通过调整该至少两个第二竖直移动机构中的至少一个第二竖直移动机构的移动端的高度，可以使标定板以任一倾斜角度绕第一竖直移动机构的顶端旋转，即通过竖直移动单元可使标定板从一个倾斜角度切换至任意倾斜角度。
在一个实施例中，竖直移动单元包括一个第一竖直移动机构和两个第二竖直移动机构，其中，第一竖直移动机构为设置于标定板底面中心的阻尼转轴，设置为为标定板旋转提供支点，两个第二竖直移动机构为设置于标定板底面的两相邻边缘的连杆，且每个连杆均与对应伺服电机连接。伺服电机控制的连杆机构可以使得连杆上的两个小球头(参见图1)上下移动。竖直移动单元的所有结构均被固定至小车底座上，即被固定至水平移动单元，从而使水平移动单元在带动标定板水平移动的同时，也带动竖直移动单元水平移动，且水平移动单元和竖直移动单元可以分别同时执行标定板的水平移动和竖直移动，从而提高标定板空间位置状态的切换效率。
在一实施例中，竖直移动单元还包括用于感应标定板倾斜角度的传感器。相应的，竖直移动单元还设置为根据传感器感应到的标定板的当前倾斜角角度与标定板的目标倾斜角角度调整该至少两个第二竖直移动机构的移动端的高度，以将标定板的倾斜角从当前倾斜角调整至目标倾斜角。
在一个实施例中，控制模块采用开环控制模式控制标定板的倾斜角的切换，即直接发送预设打角指令给伺服电机，伺服电机根据该预设打角指令通过连杆机构控制标定板移动至预设打角指令对应的倾斜角。示例性地，控制模块中存储有标定板倾斜角控制数据，比如，标定板的倾斜角的每个单位改变量均对应有多个第二竖直移动机构的伺服电机的旋转数据。这样控制模块既可根据标定 板的目标空间位置状态确定标定板的目标倾斜角，以及目标倾斜角与当前倾斜角之间的倾斜角度改变量，然后根据该倾斜角度改变量和相应单位改变量对应的多个伺服电机的旋转数据控制对应伺服电机旋转，以使标定板从当前倾斜角旋转至目标倾斜角。
在一个实施例中，控制模块采用闭环控制模式控制标定板的竖直移动，首先读取传感器的当前数值，然后计算目标数值与当前数值的差值，并通过闭环控制方法将该差值转换为对应的伺服电机打角数据控制伺服电机打角。因为是闭环控制，上述过程会不断循环从而使得标定板的倾斜角达到与目标倾斜角小于容许误差范围内的打角。闭环控制算法可以采用工业常用的PID硬件控制器，也可采用更为复杂的控制算法。
在一个实施例中，竖直移动单元包括具有旋转功能的机械臂，通过该机械臂实现标定板倾斜角的切换，以提高标定板空间位置状态的切换速度和切换精度，从而提高摄像头标定的速度和精度。
可以理解的是，通过移动模块来移动标定板，而不是移动摄像头，使得摄像头的架设更加灵活，不需要额外增加摄像头的硬件成本。
S202、控制该多个摄像头在移动模块完成标定板的空间位置状态切换时同时对标定板进行拍摄，得到多个摄像头中的至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量。
其中，标定图像是指包含完整标定板的图像，或者在包含完整标定板的同时，还包含符合预设条件的至少两个特征点的图像。图1中的标定板的特征点是指黑白格的交点。
其中，控制模块3设置为通过移动模块2切换标定板1的空间位置状态，并控制所有摄像头4在移动模块2完成标定板1的空间位置状态切换时同时对标定板1进行拍摄以得到初始图像，然后对多个摄像头4拍摄的初始图像进行检测，从而剔除仅包含部分标定板的初始图像，保留包含完整标定板的初始图像，如果包含完整标定板的初始图像的数量小于2，则直接删除该批次的所有初始图像，如果包含完整标定板的初始图像的数量大于或等于2，则将所有包含完整标定板的初始图像作为标定图像，同时记录任意两个摄像头均拍摄到相同空间位置状态下的完整标定板的标定图像的数量，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量，则停止所有摄像头的拍摄。
在一个实施例中，控制模块在将包含完整标定板的初始图像作为标定图像的同时，一方面记录任意两个摄像头均拍摄到相同空间位置状态下的标定图像的数量，另一方面确定所述任意两个摄像头拍摄的标定图像对应的内参数和外参数，并在任意两个摄像头同时拍摄到的标定图像的数量均达到目标数量时，不再为该两个摄像头添加新的标定图像。可以理解的是，如果任意两个摄像头的标定结果已确定，则在调整标定板的空间位置状态时，无需考虑该两摄像头的空间位置。
S203、根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果以完成所有摄像头的标定。
根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果，对所有摄像头中的所有组摄像头均分别进行标定，以完成所有摄像头的标定，再分析所有组摄像头的标定结果对应的标定误差，若所有组摄像头的标定结果对应的标定误差均在相应阈值范围内，则输出所有摄像头标定成功的提示信息，若任意一组摄像头的标定结果对应的标定误差超出相应的阈值范围，则输出该组摄像头标定失败的提示信息，同时控制标定板归零，并对所有摄像 头重新进行标定。
在又一实施例中，控制模块在确定任意一组摄像头的标定结果的同时，还确定该标定结果对应的标定误差，如果该标定误差超出相应的阈值范围，则可选在其他组摄像头标定完成后，根据该组摄像头的空间位置调整标定板的空间位置状态，以重新获取该组摄像头的标定图像，以及在该组摄像头的标定图像的数量达到目标数量时，重新对该两摄像头进行标定，直至标定误差在相应的阈值范围内。
本实施例中，控制模块可选采用张氏标定法从每组摄像头采集的每幅标定图像上提取符合预设条件的至少两个特征点，以及根据该至少两个特征点计算对应的内参数和外参数，并根据每组摄像头对应的所有内参数和外参数确定所述每组摄像头的标定结果，从而得到所有摄像头的标定结果。其中，内参数与外参数的计算方法可选为优化算法，比如EM算法。采用张氏标定法提取标定图像的特征点，避免了根据环境图像内容提取特征点存在的特征点质量不高，特征弱的问题。
其中，标定误差的确定方法包括：确定当前两摄像头标定过程中EM算法的迭代步数、计算误差，以及该两摄像头的标定结果对应的标定误差是否均在相应的阈值范围内，在当前两个摄像头标定过程中的迭代步数、计算误差，以及该两个摄像头的标定结果对应的标定误差均在相应阈值范围内的情况下，输出当前两个摄像头标定成功的提示信息；在当前两个摄像头标定过程中的迭代步数、计算误差，以及该两个摄像头的标定结果对应的标定误差中的至少一个不在相应的阈值范围内的情况下，通过移动模块控制标定板归零，并对所有摄像头重新进行标定。其中，EM算法在优化问题存在单极点时，能够迅速收敛，而在优化问题存在多极点时收敛速度会降低，并且计算误差较大，因此可将EM算法的迭代步数和计算误差作为其中一个判定标准。标定误差的阈值范围可选为小于一个像素宽度。标定板归零，是指标定板被水平移动单元运送至预设起始水平位置，以及被竖直移动单元移动至预设初始竖直位置。
在一个实施例中，标定误差的确定方法包括：控制模块通过提取单元从每组摄像头采集的每幅标定图像上提取预设位置的特征点，通过选取单元从提取的特征点中选择符合预设条件的至少两个特征点，然后基于张氏标定法将其中一标定图像对应摄像头作为主摄像头，将另一标定图像对应的摄像头作为副摄像头，并根据该至少两个特征点计算对应的内参数和外参数，并根据所述每组摄像头对应的所有内参数和外参数完成所述每组摄像头的标定以得到标定结果。标定结果确定后，对所有符合预设条件的特征点进行反投影计算，即通过标定的内参数和外参数将对应两摄像头拍摄到的特征点分别投影到相应摄像头所对应的标定板上以得到两个反投影点，反投影点与真实拍摄到的标定板的特征点之间会存在一定的误差，通过欧式距离量化所有特征点与反投影点之间的距离以作为量化的标定误差。其中，控制模块可选为控制计算机，控制计算机可输出标定结果，且在输出标定结果的同时还输出标定误差以及对应的内参数和外参数。
在一个实施例中，控制模块即控制计算机能够以条形图的形式显示标定误差，以及在三维空间中显示计算得到的每组摄像头的内参数和外参数，即所有摄像头在真实三维空间中的位置。
需要说明的是，本实施例不对标定图像的标定处理时机进行限定，既可以在所有组摄像头对应的标定图像均达到目标数量时，对每组摄像头的标定图像进行批处理以得到每组摄像头的标定结果；也可以在多个摄像头的标定图像的获取中，一旦检测到任意两个摄像头的标定图像的数量达到目标数量，就对该两个摄像头的标定图像进行标定处理以得到该两摄像头的标定结果。前者对计算机的并行处理能力要求较低，但在任意两摄像头标定失败时，需要重新启动所有摄像头的标定流程，后者能够及时发现标定失败的任意两摄像头，并可以及时重新获取该两摄像头的标定图像，并对获取的标定图像进行标定处理，直至将该两摄像头标定成功，在任意两摄像头标定失败时，无需重新启动对所有摄像头的标定流程，可以在任意两摄像头标定失败时简化标定流程，虽然相较 于前者对控制模块的并行计算能力要求更高。
本申请实施例提供的多视角摄像头标定方法的技术方案，通过控制模块与移动模块的配合使用可以快速完成标定板空间位置状态的切换，相较于手动变换标定板的空间位置状态来说，可以更加准确快速地完成标定板的空间位置状态的切换，并在切换完成的同时启动所有摄像头的拍摄，以及统计所有摄像头任意一组摄像头同时拍摄到的标定图像的数量是否达到目标数量，一旦达到目标数量则停止标定板空间位置状态的切换和所有摄像头的拍摄工作，并对标定图像进行标定处理，相较于相关技术来说，可使多视角摄像头的标定更加简单、快捷、高效，能够满足不同拍摄场景的多视角摄像头的标定需求。
实施例三
图3为本申请实施例三提供的计算机设备的结构示意图，如图3所示，该设备为前述实施例中的控制模块，包括处理器301、存储器302、输入装置303以及输出装置304；设备中处理器301的数量可以是至少一个，图3中以一个处理器301为例；设备中的处理器301、存储器302、输入装置303以及输出装置304可以通过总线或其他方式连接，图3中以通过总线连接为例。
存储器302作为一种计算机可读存储介质，可设置为存储软件程序、计算机可执行程序以及模块，如本申请实施例中的多视角摄像头的标定方法对应的程序指令/模块。处理器301通过运行存储在存储器302中的软件程序、指令以及模块，从而执行设备的各种功能应用以及数据处理，即实现上述的多视角摄像头的标定方法。
存储器302可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的应用程序；存储数据区可存储根据终端的使用所创建的数据等。此外，存储器302可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实例中，存储器302可包括相对于处理器301远程设 置的存储器，这些远程存储器可以通过网络连接至设备。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
输入装置303可设置为接收输入的数字或字符信息，以及产生与设备的用户设置以及功能控制有关的键信号输入。
输出装置304可包括显示屏等显示设备，例如，用户终端的显示屏。
实施例四
本申请实施例还提供了一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行一种多视角摄像头的标定方法，该方法包括：
通过移动模块切换所述标定板的空间位置状态；
控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄，得到多个摄像头中的至少两个摄像头同时拍摄到相应空间位置状态下的完整标定板的标定图像，重复通过所述移动模块切换所述标定板的空间位置状态，并控制所述多个摄像头在所述移动模块完成所述标定板的空间位置状态切换时同时对所述标定板进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量；
根据所有摄像头中的每组摄像头拍摄的标定图像确定所述每组摄像头的标定结果以完成所有摄像头的标定。
当然，本申请实施例所提供的一种包含计算机可执行指令的存储介质，其计算机可执行指令不限于如上所述的方法操作，还可以执行本申请任意实施例所提供的多视角摄像头的标定方法中的相关操作。
通过以上关于实施方式的描述，所属领域的技术人员可以清楚地了解到，本申请可借助软件及必需的通用硬件来实现，当然也可以通过硬件实现。基于 这样的理解，本申请的技术方案本质上或者说对相关技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如计算机的软盘、只读存储器(Read-Only Memory，简称ROM)、随机存取存储器(Random Access Memory，简称RAM)、闪存(FLASH)、硬盘或光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述的多视角摄像头的标定方法。
值得注意的是，上述多视角摄像头的标定装置的实施例中，所包括的各个单元和模块只是按照功能逻辑进行划分的，但并不局限于上述的划分，只要能够实现相应的功能即可；另外，各功能单元的名称也只是为了便于相互区分，并不用于限制本申请的保护范围。

Claims (10)

  1. 一种多视角摄像头标定设备，包括：
    标定板(1)；
    移动模块(2)，设置于所述标定板(1)的下方，设置为在多个摄像头(4)的拍摄范围内切换所述标定板(1)的空间位置状态；
    控制模块(3)，连接所述移动模块(2)和所述多个摄像头(4)，设置为通过所述移动模块(2)切换所述标定板(1)的空间位置状态，并控制所述多个摄像头(4)在所述移动模块(2)完成所述标定板(1)的空间位置状态切换时同时对所述标定板(1)进行拍摄，得到多个摄像头(4)中的至少两个摄像头(4)同时拍摄到相应空间位置状态下的完整标定板(1)的标定图像，重复通过所述移动模块(2)切换所述标定板(1)的空间位置状态，并控制所述多个摄像头(4)在所述移动模块(2)完成所述标定板(1)的空间位置状态切换时同时对所述标定板(1)进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头中的每组摄像头中的两个摄像头(4)同时拍摄到的标定图像的数量均达到目标数量，根据所有摄像头中每组摄像头(4)拍摄的标定图像确定所述每组摄像头(4)的标定结果以完成所有摄像头(4)的标定。
  2. 根据权利要求1所述的设备，其中，所述标定板(1)的空间位置状态包括标定板(1)的倾斜角度，所述移动模块(2)包括竖直移动单元(21)，竖直移动单元(21)包括第一竖直移动机构(211)和至少两个第二竖直移动机构(212)；
    所述竖直移动单元(21)设置为通过第一竖直移动机构(211)和至少两个第二竖直移动机构(212)调整所述标定板(1)的倾斜角度，所述第一竖直移动机构(211)设置为为所述标定板(1)中心提供旋转支点，所述至少两个第二竖直移动机构(212)分别设置于所述标定板(1)的至少两个相邻边缘的下方，设置为分别调整所述标定板(1)相应边缘的高度。
  3. 根据权利要求2所述的设备，其中，所述竖直移动单元(21)还包括 用于感应所述标定板(1)倾斜角度的传感器；
    所述控制模块(3)还设置为根据所述标定板(1)的目标倾斜角角度和所述传感器感应到的标定板(1)的当前倾斜角角度调整至少一个第二竖直移动机构(212)的高度，以将所述标定板(1)的倾斜角度从所述当前倾斜角调整至所述目标倾斜角。
  4. 根据权利要求2所述的设备，其中，所述标定板(1)的空间位置状态包括标定板(1)的水平位置，所述移动模块(2)还包括水平移动单元(22)；
    所述水平移动单元(22)为四轮车，设置为带动所述竖直移动单元(21)和所述标定板(1)在水平面移动。
  5. 根据权利要求1所述的设备，其中，所述标定板(1)为显示有预设标定图像的显示装置，所述预设标定图像在所述显示装置上的显示位置是可调的。
  6. 根据权利要求1所述的设备，其中，所述根据所有摄像头中每组摄像头(4)拍摄的标定图像确定所述每组摄像头(4)的标定结果以完成所有摄像头(4)的标定，包括：
    采用张氏标定法从所述每组摄像头(4)采集的每幅标定图像上提取符合预设条件的至少两个特征点，以及根据所述至少两个特征点计算所述每组摄像头(4)对应的内参数和外参数，并根据所述每组摄像头(4)对应的所有内参数和外参数确定所述每组摄像头(4)的标定结果，以完成所有摄像头(4)的标定。
  7. 根据权利要求6所述的设备，其中，所述内参数与外参数的计算方法为期望最大化EM算法，所述控制模块(3)在确定所述每组摄像头(4)的标定结果的同时，还确定所述标定结果的标定误差，所述控制模块(3)还设置为：
    获取当前标定结果确定过程中所述EM算法的迭代步数和计算误差，并确定所述迭代步数、计算误差和标定误差是否均在相应的阈值范围内；
    在所述迭代步数、计算误差和标定误差均在相应的阈值范围内的情况下， 输出当前所述每组摄像头(4)标定成功的提示信息。
  8. 一种多视角摄像头的标定方法，应用于权利要求1-7任一所述的多视角摄像头标定设备，包括：
    通过移动模块(2)切换所述标定板(1)的空间位置状态；
    控制所述多个摄像头(4)在所述移动模块(2)完成所述标定板(1)的空间位置状态切换时同时对所述标定板(1)进行拍摄，得到多个摄像头中的至少两个摄像头(4)同时拍摄到相应空间位置状态下的完整标定板(1)的标定图像，重复通过所述移动模块(2)切换所述标定板(1)的空间位置状态，并控制所述多个摄像头(4)在所述移动模块(2)完成所述标定板(1)的空间位置状态切换时同时对所述标定板(1)进行拍摄的步骤，在任意两个摄像头同时拍摄到的标定图像的数量达到目标数量的情况下，将所述任意两个摄像头分为一组，直至所有摄像头(4)中每组摄像头中的两个摄像头同时拍摄到的标定图像的数量均达到目标数量；
    根据所有摄像头(4)中每组摄像头(4)拍摄的标定图像确定所述每组摄像头(4)的标定结果，以完成所有摄像头(4)的标定。
  9. 根据权利要求8所述的方法，其中，所述标定板(1)的空间位置状态包括标定板(1)的倾斜角度，所述切换所述标定板(1)的空间位置状态，包括：
    通过移动模块(2)的竖直移动单元(21)的第一竖直移动机构(211)和至少两个第二竖直移动机构(212)调节所述标定板(1)的倾斜角度，其中，所述第一竖直移动机构(211)设置为为所述标定板(1)中心提供旋转支点，所述至少两个第二竖直移动机构(212)分别设置于所述标定板(1)的至少两个相邻边缘的下方，设置为分别调整所述标定板(1)相应边缘的高度。
  10. 一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行如权利要求8或9中任一所述的多视角摄像头的标定方法。
PCT/CN2021/105711 2021-02-08 2021-07-12 多视角摄像头标定设备、标定方法及存储介质 Ceased WO2022166113A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110181621.4 2021-02-08
CN202110181621.4A CN112862900B (zh) 2021-02-08 2021-02-08 多视角摄像头标定设备、标定方法及存储介质

Publications (1)

Publication Number Publication Date
WO2022166113A1 true WO2022166113A1 (zh) 2022-08-11

Family

ID=75989557

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/105711 Ceased WO2022166113A1 (zh) 2021-02-08 2021-07-12 多视角摄像头标定设备、标定方法及存储介质

Country Status (2)

Country Link
CN (1) CN112862900B (zh)
WO (1) WO2022166113A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115227398A (zh) * 2022-09-19 2022-10-25 杭州三坛医疗科技有限公司 一种配准板自动摆位方法和装置
CN115760998A (zh) * 2022-10-31 2023-03-07 浙江宇视科技有限公司 摄像机的零点位置标定方法、装置、电子设备和存储介质
CN116767868A (zh) * 2023-08-04 2023-09-19 安徽得壹能源科技有限公司 一种极片取料协同控制方法及取料装置
CN116907799A (zh) * 2023-06-16 2023-10-20 谷东科技有限公司 一种ar光机视场角的测量方法

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112862900B (zh) * 2021-02-08 2023-08-08 中国科学院深圳先进技术研究院 多视角摄像头标定设备、标定方法及存储介质
CN115811663B (zh) * 2021-09-14 2025-08-08 智博汽车科技(上海)有限公司 前视摄像头标定装置
CN114419166A (zh) * 2022-01-19 2022-04-29 白犀牛智达(北京)科技有限公司 一种相机内参标定方法、装置、电子设备及存储介质
CN115861443B (zh) * 2022-12-22 2026-02-03 北京百度网讯科技有限公司 一种多相机内参标定方法、装置、电子设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106097300A (zh) * 2016-05-27 2016-11-09 西安交通大学 一种基于高精度运动平台的多相机标定方法
CN110675458A (zh) * 2019-10-08 2020-01-10 Oppo广东移动通信有限公司 用于摄像头标定的方法、装置及存储介质
US20200262080A1 (en) * 2019-02-14 2020-08-20 InTune Products LLC Comprehensive model-based method for gantry robot calibration via a dual camera vision system
CN112862900A (zh) * 2021-02-08 2021-05-28 中国科学院深圳先进技术研究院 多视角摄像头标定设备、标定方法及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106097300A (zh) * 2016-05-27 2016-11-09 西安交通大学 一种基于高精度运动平台的多相机标定方法
US20200262080A1 (en) * 2019-02-14 2020-08-20 InTune Products LLC Comprehensive model-based method for gantry robot calibration via a dual camera vision system
CN110675458A (zh) * 2019-10-08 2020-01-10 Oppo广东移动通信有限公司 用于摄像头标定的方法、装置及存储介质
CN112862900A (zh) * 2021-02-08 2021-05-28 中国科学院深圳先进技术研究院 多视角摄像头标定设备、标定方法及存储介质

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115227398A (zh) * 2022-09-19 2022-10-25 杭州三坛医疗科技有限公司 一种配准板自动摆位方法和装置
CN115227398B (zh) * 2022-09-19 2023-03-03 杭州三坛医疗科技有限公司 一种配准板自动摆位方法和装置
CN115760998A (zh) * 2022-10-31 2023-03-07 浙江宇视科技有限公司 摄像机的零点位置标定方法、装置、电子设备和存储介质
CN116907799A (zh) * 2023-06-16 2023-10-20 谷东科技有限公司 一种ar光机视场角的测量方法
CN116767868A (zh) * 2023-08-04 2023-09-19 安徽得壹能源科技有限公司 一种极片取料协同控制方法及取料装置

Also Published As

Publication number Publication date
CN112862900A (zh) 2021-05-28
CN112862900B (zh) 2023-08-08

Similar Documents

Publication Publication Date Title
WO2022166113A1 (zh) 多视角摄像头标定设备、标定方法及存储介质
US12140680B1 (en) System and method of capturing and generating panoramic three-dimensional images
EP3764630B1 (en) Method and apparatus for adjusting scanning status
CN113658221B (zh) 一种基于单目相机的agv行人跟随方法
TWI552598B (zh) 一種攝影機之自動校正系統及其自動校正方法
KR101166508B1 (ko) 도화이미지 및 지피에스정보 합성을 통해 수정정보를 반영하는 수치지도 업데이트 시스템
WO2022062153A1 (zh) 一种高尔夫球落地式检测方法、系统及存储介质
CN112702587A (zh) 一种智能跟踪投影方法及系统
WO2025194747A1 (zh) 物体位姿的确定方法、系统、设备、介质及程序产品
CN112083621B (zh) 一种可自主定制的多目视觉三维近景摄影装置
CN116309870B (zh) 机动车全景影像系统的标定装置及方法
CN119450226A (zh) 一种机器人拍摄定位方法、装置、设备及介质
WO2025214315A1 (zh) 一种投影控制方法、设备及存储介质
CN110555887B (zh) 应用于相机标定系统的标定方法和相机标定系统
CN114755782B (zh) 点光谱镜头安装方法、装置、系统、存储介质以及产品
CN116203532A (zh) 雷达水平度一致性的调节方法、装置、设备及存储介质
CN115984099A (zh) 全景成像设备和方法
CN112166396A (zh) 云台及其控制方法、外置控制装置
CN114625157A (zh) 一种空中倾斜摄影结合采集数据的自动建模方法及系统
TWI906033B (zh) 影像收集方法及影像擷取裝置
CN120782870A (zh) 基于自适应多视角优化算法的三维标定方法、系统及介质
CN112399068B (zh) 一种图像处理系统和图像处理方法
CN121660879A (zh) 多目相机的图像拼接方法、装置及电子设备
CN121600066A (zh) 一种目标定位方法及装置
CN119887943A (zh) 一种无人机相机标定方法、装置、介质、产品及系统

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: 21924117

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: 21924117

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 16.01.2024)

122 Ep: pct application non-entry in european phase

Ref document number: 21924117

Country of ref document: EP

Kind code of ref document: A1