Large cylinder segment pushing system and method based on visual positioning guidance

Abstract

The application provides a large cylinder segment propulsion system and method based on visual positioning guidance, which is used to solve the technical problems that the large cylinder segment propulsion system cannot be installed with traditional contact positioning systems or laser positioning systems for automatic guidance, and the installation of single-point visual systems leads to the failure to accurately obtain the docking pose information. The large cylinder segment propulsion system provided by the application comprises N targets which are circumferentially arranged on the outer sidewall of an inner cylinder and correspond to the front positions of N pin holes respectively; the N targets are symmetrically arranged about the center line in the vertical direction of the end face of the inner cylinder; N sets of visual systems are circumferentially arranged on the end face of a propulsion device close to an outer cylinder, and when the propulsion device is docked with the inner cylinder, the N sets of visual systems are arranged in one-to-one correspondence with the N targets; a control unit solves the docking pose information according to the image information collected by the visual systems; and the solved docking pose information is used to guide the attitude adjustment platform to adjust the attitude of the propulsion device.

Description

TECHNICAL FIELD

[0001] The present application relates to a large cylinder segment propulsion system, in particular to a large cylinder segment propulsion system and method based on visual positioning guidance. BACKGROUND

[0002] The large cylinder segment propulsion system is a special rigid body assembly docking system, which is usually used to realize the process of pushing or pulling the product out of the external cylinder segment during the transfer of space products. The propulsion system usually involves an outer cylinder, an inner cylinder and a propulsion device. A plurality of pin holes are arranged circumferentially on the outer side wall of the inner cylinder, and a plurality of telescopic pins are arranged axially on the inner side wall of the propulsion device. The propulsion device is sleeved on the inner cylinder of the large cylinder segment, and the propulsion process is as follows: the pose of the propulsion device is adjusted by the pose adjustment platform, so that the telescopic pins on the circumference of the propulsion device can be inserted into the pin holes on the inner cylinder, and then the inner cylinder is pushed into the outer cylinder or pulled out of the outer cylinder.

[0003] During the automatic docking process of the large cylinder segment propulsion system, an automatic guidance device needs to be set up to realize the positioning of the telescopic pins on the propulsion device and the pin holes on the inner cylinder by solving the pose relationship between the propulsion device and the inner cylinder segment. However, due to the small difference in diameter between the outer cylinder, the propulsion device and the inner cylinder in the large cylinder segment propulsion system, the space for installing the automatic guidance device on the propulsion device is very small, and it is impossible to install traditional contact positioning systems or laser positioning systems for automatic guidance. Single-point vision systems can be installed, but single-point vision systems will result in a small field of view of the vision system, which cannot accurately obtain the docking pose information. SUMMARY

[0004] The purpose of the present application is to solve the technical problems that the large cylinder segment propulsion system cannot install traditional contact positioning systems or laser positioning systems for automatic guidance, and that the installation of single-point vision systems leads to the inability to accurately obtain docking pose information, and to provide a large cylinder segment propulsion system and method based on visual positioning guidance.

[0005] In order to achieve the above purpose, the technical solution provided by the present application is as follows:

[0006] A large cylinder segment propulsion system based on visual positioning guidance, comprising an outer cylinder, an inner cylinder and a propulsion device; the outer side wall of the inner cylinder is circumferentially provided with M pin holes, and the inner side wall of the propulsion device is circumferentially provided with telescopic pins corresponding to the M pin holes 21, and M is an even number greater than or equal to 4; the special feature is that,

[0007] It further comprises N targets, N groups of vision systems and a control unit, N is less than or equal to M and greater than or equal to 4, and N is an even number;

[0008] N targets are arranged on the outer sidewall of the inner cylinder in a circumferential direction and correspond to the front positions of the N pin holes respectively; the N targets are symmetrically arranged about the center line in the vertical direction of the end face of the inner cylinder;

[0009] N sets of vision systems are arranged on the end face of the propulsion device close to the outer cylinder, and when the propulsion device is docked with the inner cylinder, the N sets of vision systems are arranged in one-to-one correspondence with the N targets; the lens of the vision system is arranged in the direction of facing the inner cylinder, and the camera optical axis is perpendicular to the center axis of the inner cylinder;

[0010] The output end of the vision system is connected with the control unit, and is used for solving the docking pose information of the propulsion device and the inner cylinder according to the image information collected by the vision system;

[0011] The propulsion device is arranged on the pose adjustment platform, and the pose adjustment platform is electrically connected with the control unit, and is used for guiding the pose adjustment platform to adjust the attitude of the propulsion device through the solved docking pose information.

[0012] Further, the vision system is a monocular vision system.

[0013] The application also provides a large cylinder segment propulsion method based on visual positioning guidance, which is characterized in that the method comprises the following steps:

[0014] 1. Building a large cylinder segment propulsion system based on visual positioning guidance as described above;

[0015] 2. Calibrating the internal parameters and external parameters of the vision system;

[0016] 3. Artificially controlling the pose adjustment platform to drive the propulsion device to run to the position of docking with the inner cylinder, at this time, the image information of the corresponding targets is collected by the N cameras on the N vision systems respectively;

[0017] 4. Moving the pose adjustment platform to drive the propulsion device to the initial shooting position, the image information of the corresponding targets is collected by the N cameras respectively, and is compared with the N target image information obtained in step 3, so as to obtain the docking pose information of the propulsion device and the inner cylinder solved by a single vision system in the coordinate system of the propulsion device; the docking pose information comprises translation information and roll angle information;

[0018] 5. Multi-camera fusion of the docking pose information

[0019] 5.1. Solving the mean values of the up-down translation component, the left-right translation component and the front-back translation component of the translation information in the N sets of vision systems respectively, and taking the mean values of the components as the current solved translation information;

[0020] 5.2. Solving the mean value of the roll angle information in the N sets of vision systems; according to the positional relationship of the N sets of vision systems, the translation component of the roll angle reflected in the up-down direction of each vision system is solved Solve the roll angle by the radius of the inner cylinder, and reflect the roll angle component in the up-down direction of each visual system And add the obtained average roll angle to the current solved roll angle information

[0021] 6】Docking pose adjustment

[0022] 6.1】The adjustment platform drives the propulsion device to make a roll angle adjustment according to the current solved roll angle information in step 5.2; it is judged whether the roll angle at this time is less than the roll angle threshold value, if yes, the roll angle adjustment of the propulsion device is completed, step 6.2 is executed, if not, it is returned to step 4 for iteration until the roll angle is less than the roll angle threshold value

[0023] 6.2】The translation information of the propulsion device in the coordinate system solved in step 5.1 is converted to the coordinate system of the adjustment platform, and the up-down translation component, left-right translation component and front-back translation component of the converted translation information are obtained; the adjustment platform drives the propulsion device to make a movement along the up-down translation component, left-right translation component and front-back translation component of the converted translation information respectively; it is judged whether the up-down translation component, left-right translation component and front-back translation component at this time are less than the corresponding up-down translation component threshold value, left-right translation component threshold value and front-back translation component threshold value, if yes, the translation adjustment of the propulsion device is completed, step 7 is executed, if not, it is returned to step 4 for iteration until the up-down translation component, left-right translation component and front-back translation component are all less than the corresponding up-down translation component threshold value, left-right translation component threshold value and front-back translation component threshold value, and the positioning of the telescopic pin and the pin hole is completed

[0024] 7】The telescopic pin is inserted into the corresponding pin hole, the docking of the propulsion device and the inner cylinder is realized, and the propulsion work of the large cylinder segment is completed

[0025] Further, the center line of the N groups of visual systems 5 in the vertical direction of the end face of the inner cylinder 2 is divided into left and right two parts, and each part has n groups of visual systems 5. The up-down translation components solved by the n visual systems 5 located on the left side of the center line in the vertical direction of the end face of the inner cylinder 2 are respectively denoted as x 1l , x 2l …x nl , and the up-down translation components solved by the n visual systems 5 located on the right side of the center line in the vertical direction of the end face of the inner cylinder 2 are respectively denoted as x 1r , x 2r …x nr , then the roll angle component reflected in the up-down direction is It is obtained by calculation through the following formula:

[0026]

[0027] ​Wherein, R represents the radius of the inner cylinder.

[0028] Further, step 4 is specifically:

[0029] 4.1】The pose adjustment platform drives the propulsion device to move to a preset initial shooting position capable of collecting target images;

[0030] 4.2】The cameras on the four sets of vision systems simultaneously collect image information of corresponding targets on the inner cylinder;

[0031] 4.3】Solving the docking pose information of the propulsion device and the inner cylinder

[0032] 4.3.1】Extracting the target landmark points in the target image information obtained in step 4.2 to form a first target landmark point set; solving the three-dimensional coordinates of the first target landmark point set in the vision system coordinate system, and converting the three-dimensional coordinates solved at this time in the vision system coordinate system to the propulsion device coordinate system by using the external parameters calibrated in step 2;

[0033] 4.3.2】Extracting the target landmark points in the target image information obtained in step 3 to form a second target landmark point set; solving the three-dimensional coordinates of the second target landmark point set in the vision system coordinate system, and converting the three-dimensional coordinates solved at this time in the vision system coordinate system to the propulsion device coordinate system by using the external parameters calibrated in step 2;

[0034] 4.3.3】Using the SVD algorithm to obtain the conversion relationship of the propulsion device coordinate system in two states of step 4.3.1 and step 4.3.2, obtaining the roll angle, the up-down translation component, the left-right translation component and the front-back translation component in the corresponding vision system, that is, obtaining the docking pose information of the propulsion device and the inner cylinder solved by a single vision system in the propulsion device coordinate system.

[0035] Further, in step 6.2, the up-down translation component x p , the left-right translation component y p and the front-back translation component z p of the converted translation information are respectively:

[0036]

[0037] Wherein, x represents the fused up-down translation component, y represents the fused left-right translation component, and z represents the fused front-back translation component, represents the fused roll angle information when the pose is solved for the first time.

[0038] The beneficial effects of the present application compared with the prior art are as follows:

[0039] 1. This invention provides a large-scale cylindrical section propulsion system based on visual positioning guidance. Multiple targets are set on the outer wall of the inner cylinder, corresponding to the pin holes, directly in front of them. Multiple vision systems are also set on the end face of the propulsion device near the outer cylinder, directly opposite the telescopic pin. Together with the control unit, this forms a visual positioning guidance system, guiding the attitude adjustment platform to adjust the orientation of the propulsion device and complete the docking between the propulsion device and the inner cylinder section. The system is simple, accurate, and convenient, effectively solving the problem in large-scale cylindrical section propulsion systems where the inner cylinder diameter is large and the diameter difference between the propulsion device and the inner and outer cylinders is small, resulting in limited installation space for automatic guidance equipment and the inability to obtain the relative positional relationship between the propulsion device and the inner cylinder.

[0040] 2. The present invention provides a large-diameter cylindrical section propulsion method based on vision positioning guidance. By using multiple vision systems set on one end face of the propulsion device to quickly measure the target on the inner cylinder, the propulsion device and the inner cylinder are quickly positioned, and the relative pose information of the propulsion device and the inner cylinder section is obtained. This information is fused and processed to obtain an accurate docking pose, namely the adjustment pose in four directions: up and down, left and right, forward and roll. The pose adjustment platform is guided to adjust the orientation of the propulsion device, thereby completing the precise docking of the propulsion device and the inner cylinder section. This method can be widely applied to the propulsion of large-diameter cylindrical sections with a small field of view.

[0041] 3. The present invention provides a large-diameter cylindrical section propulsion method based on vision positioning guidance. It utilizes multi-vision system information fusion to obtain pose information, thus solving the problem of inaccurate pose determination by a single-point vision system caused by the small field of view of large-diameter cylindrical sections.

[0042] 4. The present invention provides a large-scale cylindrical section propulsion method based on visual positioning guidance. Through iterative process, it solves the problem that the propulsion device cannot reach the position in one go, resulting in inconsistent coordinate systems in the subsequent process. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of an embodiment of a large cylindrical section propulsion system based on visual positioning guidance according to the present invention (control unit not shown).

[0044] Figure 2 This is a schematic diagram of the structure of the vision system distributed on the propulsion device in an embodiment of the present invention.

[0045] The specific reference numerals in the attached drawings are as follows: 1-outer cylinder; 2-inner cylinder; 21-pin hole; 3-propulsion device; 31-telescopic pin; 4-target; 5-vision system; 6-attitude adjustment platform. Detailed Implementation

[0046] To make the advantages and features of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] As Figure 1 shown, a large cylinder segment propulsion system based on multi-camera information fusion positioning includes an outer cylinder 1, an inner cylinder 2, a propulsion device 3, four targets 4, four sets of vision systems 5, and a control unit.

[0048] The outer lateral wall of the inner cylinder 2 is circumferentially provided with M pin holes 21, and the inner lateral wall of the propulsion device 3 is circumferentially provided with retractable pins 31 corresponding to the M pin holes 21, and in this embodiment, the number of pin holes 21 and retractable pins 31 is eight. Through the cooperation of the retractable pins 31 and the pin holes 21, the propulsion device 3 pushes the inner cylinder 2 into the outer cylinder 1 or pulls the inner cylinder 2 out of the outer cylinder 1.

[0049] The four targets 4 are circumferentially arranged on the outer lateral wall of the inner cylinder 2 at positions corresponding to the front of the four pin holes 21, respectively; at the same time, the four targets 4 are symmetrically arranged about the center line in the vertical direction of the end face of the inner cylinder 2. In this embodiment, the target 4 is pasted on the outer lateral wall of the inner cylinder 2, and the center of the target 4 is aligned with the center of the pin hole 21.

[0050] As Figure 2 shown, the four sets of vision systems 5 are circumferentially arranged on the end face of the propulsion device 3 close to the outer cylinder 1, so that when the propulsion device 3 is docked with the inner cylinder 2, the four sets of vision systems 5 are arranged one by one opposite to the four targets 4. Specifically, the vision system 5 includes a camera and a lighting system, the lens of the camera is arranged in the direction of the inner cylinder 2, and the camera optical axis is perpendicular to the center axis of the inner cylinder 2, so as to ensure that the camera can collect image information of the target 4. Four sets of vision systems 5 cooperate with the corresponding four targets 4 to complete the positioning of the retractable pins 31 and the pin holes 21, so as to realize the docking of the propulsion device 3 and the inner cylinder 2. Due to the distance limitation between the propulsion device 3 and the outer cylinder 1, the camera 5 in the present application preferably uses a monocular camera. The output end of the vision system 5 is connected with the control unit, which is used for inputting the image information collected by the vision system 5 into the control unit, and combining the vision positioning guidance technology to solve the docking pose information of the propulsion device 3 and the inner cylinder 2.

[0051] The propulsion device 3 is arranged on the pose adjusting platform 6, and the pose adjusting platform 6 is electrically connected with the control unit, which is used for guiding the pose adjusting platform 6 to adjust the attitude of the propulsion device 3 through the solved docking pose information, so as to complete the positioning and docking of the retractable pins 31 and the pin holes 21, and then complete the subsequent propulsion work.

[0052] The present application also provides a large cylinder segment propulsion method based on multi-camera information fusion positioning, which specifically includes the following steps:

[0053] 1】Build a large cylinder segment propulsion system based on vision positioning guidance.

[0054] 2】Calibrate the internal and external parameters of the vision system 5

[0055] 2.1】Calibration of internal parameters of vision system 5

[0056] The internal parameters of vision system 5 are calibrated by using a high-precision calibration board to obtain an intrinsic matrix. The purpose of the calibration of the internal parameters of vision system 5 is to obtain the hardware parameters of vision system 5, including the focal length of the lens, the coordinates of the intersection of the camera optical axis and the imaging physical plane in the image coordinate system, and the distortion coefficient.

[0057] 2.2】Calibration of external parameters of vision system 5

[0058] The calibration of the external parameters of vision system 5 is to obtain the conversion relationship between the vision system coordinate system and the propulsion device coordinate system. Since vision system 5 is fixed to propulsion device 3, the relative position of the vision system coordinate system and the propulsion device coordinate system is fixed and unchanged. The external parameters of vision system 5 are calibrated by using target 4 pasted on inner cylinder 2 to obtain the conversion relationship.

[0059] 3】Calibration of docking pose

[0060] The purpose of the calibration of the docking pose is to record the relationship between target 4 and vision system 5 in the docking state, and to associate inner cylinder 2 with vision system 5. The specific calibration process is as follows: the pose adjustment platform 6 is manually controlled to drive the propulsion device 3 to run to the position for docking with inner cylinder 2, i.e., the position where the four telescopic pins 31 are matched with the corresponding pin holes 21 to complete docking. At this time, the image information of the corresponding target 4 is collected by the four groups of cameras on vision system 5.

[0061] 4】Solution of docking pose information of a single vision system 5

[0062] 4.1】Make the pose adjustment platform 6 drive the propulsion device 3 to move to a preset initial shooting position where the target image can be collected.

[0063] 4.2】The cameras on the four groups of vision system 5 simultaneously collect the image information of the corresponding target 4 on inner cylinder 2.

[0064] 4.3】Solution of docking pose information of a single vision system 5

[0065] 4.3.1】Extract the target marker points in the target 4 image information obtained in step 4.2 to form a first target marker point set; solve the three-dimensional coordinates of the first target marker point set in the vision system coordinate system, and convert the three-dimensional coordinates solved in the vision system coordinate system to the propulsion device coordinate system by using the external parameters calibrated in step 2.

[0066] 4.3.2 Extract the target mark points in the target 4 image information obtained in step 3 to form a second target mark point set; solve the three-dimensional coordinates of the second target mark point set in the vision system coordinate system, and convert the three-dimensional coordinates solved in the vision system coordinate system to the propulsion device coordinate system by using the external parameters calibrated in step 2.

[0067] 4.3.3 Obtain the conversion relationship of the propulsion device coordinate system in two states of step 4.3.1 and step 4.3.2 by using the SVD algorithm, to obtain the rotation matrix R and the translation vector T, i.e. the roll angle information and the translation information.

[0068] By using the relationship between the Euler angle and the rotation matrix R, the rotation matrix R corresponding to the i-th vision system 5 is decomposed into three Euler angles: the pitch angle θ i , the yaw angle ψ i , and the roll angle Since the inner cylinder 2 and the propulsion device 3 always maintain the axis parallel during the docking process, i.e. the pitch angle θ i and the yaw angle ψ i are approximately equal to 0, the corresponding roll angle can be directly solved by the rotation matrix R.

[0069] Each item of the translation vector T corresponds to a translation amount, wherein the translation vector T of the i-th vision system 5 can be decomposed into the up-down translation component x i , the left-right translation component y i , and the front-back translation component z i .

[0070] The roll angle , the up-down translation component x i , the left-right translation component y i , and the front-back translation component z i are obtained, i.e. the docking pose information of the propulsion device 3 and the inner cylinder 2 solved by the i-th vision system 5 (single vision system) in the propulsion device coordinate system is obtained.

[0071] 5 Multi-vision system fusion of docking pose information

[0072] Since the field of view range of the single vision system 5 is relatively very small compared with the entire inner cylinder 2, the translation information and the roll angle information solved by the single vision system in step 4 are not real values. Considering the installation directions of the four groups of vision systems 5, a large part of the roll angle between the inner cylinder 2 and the propulsion device 3 is reflected on the up-down translation component, and the influence on the left-right translation component and the front-back translation component can be ignored, so the up-down translation component and the roll angle information solved by the single vision system 5 are not accurate, and the docking pose information obtained by the four groups of vision systems 5 needs to be fused to solve the correct docking pose.

[0073] 5.1 Displacement information fusion

[0074] The mean values of the up-down displacement component, the left-right displacement component and the front-back displacement component are solved respectively under the four sets of vision systems 5, and the mean values of the components are taken as the current solved displacement information.

[0075] Since the four up-down displacement components solved in step 4 all contain the roll angle reflected on the up-down displacement components of each vision system 5, according to the distribution of the four sets of vision systems 5, the up-down displacement components solved by the four sets of vision systems 5 can be respectively expressed as:

[0076]

[0077] In the above formula, x is the actual up-down displacement component, is the displacement component in the up-down direction reflected by the roll angle.

[0078] The finally solved fused up-down displacement component is x=(x1+x2+x3+x4) / 4.

[0079] The finally solved fused left-right displacement component is y=(y1+y2+y3+y4) / 4, wherein y1, y2, y3, y4 are respectively the left-right displacement components solved by the four sets of vision systems 5.

[0080] The finally solved fused front-back displacement component is z=(z1+z2+z3+z4) / 4, wherein z1, z2, z3, z4 are respectively the front-back displacement components solved by the four sets of vision systems 5.

[0081] 5.2 Roll angle information fusion

[0082] According to the positional relationship of the four sets of vision systems 5, the displacement component of the roll angle reflected in the up-down direction of each vision system 5 is solved Then, the roll angle component of the roll angle reflected in the up-down direction of each vision system 5 is solved by using the radius of the inner cylinder 2 And is added to the four roll angle mean values obtained in step 4, as the solved fused roll angle

[0083] Specifically, the fused roll angle is obtained by calculation through the following formula:

[0084]

[0085] In the above formula, wherein, are respectively the roll angles solved by the four sets of vision systems 5; The roll angle reflected in the up-down direction roll angle component of the roll angle.

[0086] The n groups of vision systems 5 in the present application are divided into left and right parts along the center line in the vertical direction of the end face of the inner cylinder 2, and each of the left and right parts has n groups of vision systems 5. The up-down translation components solved by the n vision systems 5 located on the left side of the center line in the vertical direction of the end face of the inner cylinder 2 are respectively denoted as x 1l , x 2l … x nl , and the up-down translation components solved by the n vision systems 5 located on the right side of the center line in the vertical direction of the end face of the inner cylinder 2 are respectively denoted as x 1r , x 2r … x nr The translation component of the roll angle reflected in the up-down direction of each vision system 5 can be solved by the following formula

[0087]

[0088] The roll angle component of the roll angle reflected in the up-down direction of each vision system 5 is further solved by the following formula:

[0089]

[0090] In the above formula, R represents the radius of the inner cylinder 2.

[0091] In the present embodiment, x1 and x4 are the up-down translation components solved by the 2 groups of vision systems 5 located on the left side of the center line in the vertical direction of the end face of the inner cylinder 2, and x2 and x3 are the up-down translation components solved by the 2 groups of vision systems 5 located on the right side of the center line in the vertical direction of the end face of the inner cylinder 2. Therefore, in the present embodiment, the roll angle component of the roll angle reflected in the up-down direction is calculated by the following formula:

[0092]

[0093] 6】Pose iteration

[0094] Since the weight and volume of the propulsion device 3 are large, when the pose adjustment platform 6 drives the propulsion device 3 to adjust the displacement after receiving the pose adjustment information of the vision system 5, the pose cannot be adjusted to the position at one time, that is, the roll angle and the translation cannot be cleared at one time. Therefore, multiple pose iterations are required, that is, the docking pose information of the vision system 5 is re-solved according to the last pose adjustment, and the pose adjustment platform 6 adjusts the displacement of the propulsion device 3 according to the latest docking pose information until the docking pose information of the vision system 5 solved is less than a preset threshold.

[0095] 6.1】Roll angle pose adjustment

[0096] After the four degrees of freedom of the docking pose information is obtained, the pose adjustment platform 6 first adjusts the roll angle. The pose adjustment platform 6 drives the propulsion device to adjust the roll angle once according to the roll angle information solved in step 5.2; it is judged whether the roll angle is less than the roll angle threshold value at this time, if yes, the roll angle adjustment of the propulsion device is completed, step 6.2 is executed, if not, it is returned to step 4 for iteration until the roll angle is less than the roll angle threshold value. The selection of the roll angle threshold value depends on the gap of the telescopic pin 31 and the pin hole 21 in the roll angle direction.

[0097] 6.2】Adjusting the pose in the translation direction

[0098] After the roll angle adjustment is completed, the pose adjustment in the translation direction is performed. At this time, the relative position between the vision system 5 and the propulsion device 3 is known and unchanged, but the relative position between the pose adjustment platform 6 and the vision system 5 has changed, therefore, the translation information in the propulsion device coordinate system solved in step 5.1 needs to be converted to the pose adjustment platform coordinate system. The upper translation component x p , the left translation component y p and the front translation component z p of the converted translation information are respectively:

[0099]

[0100] The pose adjustment platform 6 drives the propulsion device 3 to move along the converted upper translation component x p , left translation component y p and front translation component z p respectively; it is judged whether the upper translation component x p , left translation component y p and front translation component z p are less than the corresponding upper translation component threshold value, left translation component threshold value and front translation component threshold value at this time, if yes, the translation adjustment of the propulsion device is completed, step 7 is executed, if not, it is returned to step 4 for iteration until the upper translation component x p , left translation component y p and front translation component z p are all less than the corresponding upper translation component threshold value, left translation component threshold value and front translation component threshold value, the positioning of the telescopic pin 31 and the pin hole 21 is completed. The selection of the front translation component threshold value depends on the gap of the telescopic pin 31 and the pin hole 21 in the front direction (axial direction), the threshold values of the upper translation component and the left translation component are the same, which are selected according to the empirical value in the equipment debugging process.

[0101] 7】The telescopic pin 31 is inserted into the corresponding pin hole 21, the propulsion device 3 is docked with the inner cylinder 2, and the propulsion work of the large cylinder segment is completed.

[0102] The above merely aims to explain the technical solutions of the present application, and is not intended to limit the same. Those skilled in the art can make modifications to the specific technical solutions described in the above embodiments, or make equivalent replacements to some of the technical features, and these modifications or replacements do not cause the corresponding technical solutions to deviate from the scope of the technical solutions protected by the present application.

Claims

1. A method for propulsion of large cylindrical sections based on visual positioning guidance, characterized in that, Includes the following steps:

1. A large cylindrical section propulsion system based on visual positioning guidance is constructed. The large cylindrical section propulsion system based on visual positioning guidance includes an outer cylinder (1), an inner cylinder (2), a propulsion device (3), N targets (4), N sets of vision systems (5), and a control unit. The outer side wall of the inner cylinder (2) is provided with M pin holes (21) in the circumferential direction. The inner side wall of the propulsion device (3) is provided with telescopic pins (31) corresponding to the M pin holes (21) in the circumferential direction. M is an even number greater than or equal to 4, and N is an even number less than or equal to M. N targets (4) are circumferentially arranged on the outer wall of the inner cylinder (2), corresponding to the positions directly in front of the N pin holes (21); the N targets (4) are symmetrically arranged about the center line of the vertical direction of the end face of the inner cylinder (2); N sets of vision systems (5) are circumferentially arranged on one end face of the propulsion device (3) near the outer cylinder (1), and when the propulsion device (3) is docked with the inner cylinder (2), the N sets of vision systems (5) are arranged facing the N targets (4) one by one; the lens of the vision system (5) is set facing the inner cylinder (2), and the optical axis of the camera is perpendicular to the central axis of the inner cylinder (2); The output of the vision system (5) is connected to the control unit and is used to solve the docking posture information of the propulsion device (3) and the inner cylinder (2) based on the image information collected by the vision system (5). The propulsion device (3) is mounted on the attitude adjustment platform (6), which is electrically connected to the control unit and is used to guide the attitude adjustment platform (6) to adjust the attitude of the propulsion device (3) through the solved docking posture information. 2】Calibrate the internal and external parameters of the vision system (5); 3】Manually control the attitude adjustment platform (6) to drive the propulsion device (3) to the position where it docks with the inner cylinder (2). At this time, the cameras on the N vision systems (5) respectively collect the image information of the corresponding targets (4); 4】The attitude adjustment platform (6) drives the propulsion device (3) to the initial shooting position. N cameras respectively collect the image information of the corresponding target (4) and compare it with the image information of the N targets obtained during the calibration in step 3 to obtain the docking pose information of the propulsion device (3) and the inner cylinder (2) solved by a single vision system (5) in the coordinate system of the propulsion device. The docking pose information includes translation information and roll angle information.

5. Perform multi-camera fusion on docking pose information. 5.1】Calculate the mean values ​​of the vertical translation components, horizontal translation components, and forward and backward translation components of the translation information under N visual systems (5), and use the mean value of each component as the translation information to be solved. 5.2】Calculate the mean roll angle of the roll angle information under N visual systems (5); based on the positional relationship of the N visual systems (5), calculate the translation component of the roll angle in the vertical direction of each visual system (5). Then, using the radius of the inner cylinder (2), the roll angle components reflected in the vertical direction of each visual system (5) are calculated. And Add the average roll angle obtained to obtain the roll angle information for the current solution; In this context, N sets of visual systems (5) are defined as being divided into left and right parts along the center line of the vertical direction of the inner cylinder (2) end face. The vertical translation components of the n visual systems (5) located to the left of the center line of the vertical direction of the inner cylinder (2) end face are denoted as x. 1l x 2l …x nl The vertical translation components of the n visual systems (5) located to the right of the center line in the vertical direction of the inner cylinder (2) end face are denoted as x. 1r x 2r …x nr The roll angle is reflected in the translational components in the vertical direction of each visual system (5). The roll angle is reflected in the roll angle components in the vertical direction of each visual system (5). Solve using the following formulas respectively: Where R represents the radius of the inner cylinder (2); 6) Docking position adjustment 6.1】The attitude adjustment platform (6) drives the propulsion device (3) to perform a roll angle adjustment based on the roll angle information obtained in step 5.

2. It is determined whether the roll angle is less than the roll angle threshold. If yes, the roll angle adjustment of the propulsion device (3) is completed and step 6.2 is executed. If no, it returns to step 4 for iteration until the roll angle is less than the roll angle threshold. 6.2】Transform the translation information in the current propulsion device coordinate system obtained in step 5.1 to the attitude adjustment platform coordinate system to obtain the vertical translation components, horizontal translation components and forward and backward translation components of the transformed translation information; The attitude adjustment platform (6) drives the propulsion device (3) to move once along the vertical translation components, horizontal translation components and forward and backward translation components of the transformed translation information; Determine whether the vertical translation components, horizontal translation components and forward and backward translation components are less than their corresponding vertical translation component threshold, horizontal translation component threshold and forward and backward translation component threshold. If yes, the translation adjustment of the propulsion device (3) is completed and step 7 is executed. If no, return to step 4 for iteration until the vertical translation components, horizontal translation components and forward and backward translation components are all less than their corresponding vertical translation component threshold, horizontal translation component threshold and forward and backward translation component threshold, and the positioning of the telescopic pin (31) and the pin hole (21) is completed. 7】 Insert the telescopic pin (31) into the corresponding pin hole (21) to realize the docking of the propulsion device (3) and the inner cylinder (2) and complete the propulsion work of the large cylinder section.

2. The method for propulsion of large cylindrical sections based on visual positioning guidance according to claim 1, characterized in that, Step 4 is as follows: 4.1】The attitude adjustment platform (6) drives the propulsion device (3) to move to the preset initial shooting position where the target image can be acquired; 4.2】The cameras on the four sets of vision systems (5) simultaneously acquire image information of the corresponding target (4) on the inner cylinder (2); 4.3 Solving for the docking pose information of the propulsion device (3) and the inner cylinder (2) 4.3.1 Extract the target marker points from the target (4) image information obtained in step 4.2 to form the first target marker point set; solve the three-dimensional coordinates of the first target marker point set in the visual system coordinate system, and use the external parameters calibrated in step 2 to transform the three-dimensional coordinates in the visual system coordinate system to the propulsion device coordinate system. 4.3.2 Extract the target marker points from the target (4) image information obtained in step 3 to form a second target marker point set; solve the three-dimensional coordinates of the second target marker point set in the visual system coordinate system, and use the external parameters calibrated in step 2 to transform the three-dimensional coordinates in the visual system coordinate system to the propulsion device coordinate system. 4.3.3】Use the SVD algorithm to obtain the transformation relationship of the propulsion device coordinate system in the two states of step 4.3.1 and step 4.3.2, and obtain the roll angle, vertical translation component, horizontal translation component and forward and backward translation component under the corresponding vision system (5), that is, obtain the docking pose information of the propulsion device (3) and the inner cylinder (2) solved by a single vision system (5) under the propulsion device coordinate system.

3. The method for propulsion of large cylindrical sections based on visual positioning guidance according to claim 2, characterized in that: In step 6.2, the vertical translation component x of the obtained transformed translation information p , left and right translation components y p and the forward and backward translation components z p They are respectively: Where x represents the fused vertical translation component, y represents the fused horizontal translation component, and z represents the fused forward / backward translation component. This indicates the roll angle information after merging.

4. The method for propulsion of large cylindrical sections based on visual positioning guidance according to claim 1, characterized in that: In step 1, the vision system (5) is a monocular vision system.

5. A method for propulsion of large cylindrical sections based on visual positioning guidance according to claim 1 or 4, characterized in that: In step 1, N = 4.

Images