A vision system for space station robotic arms
By designing a multimodal vision system on the space station robot arm, including binocular cameras and Ethernet communication, the problems of single functions, large resource consumption and limited data transmission are solved, and multiple requirements for complex tasks are compatible and efficient. Position measurement.
Patent Information
- Application Number
- CN202210977497.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-15
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2042-08-15
AI Technical Summary
The existing space station robotic arm vision system has a single function and cannot adapt to complex tasks. The failure of a single camera leads to vision loss, large resource consumption, and limited data transmission bandwidth, which affects the efficiency of the control system.
Design a visual system including at least two wrist cameras, an elbow gimbal camera, a docking camera, a data communication device and a data processing device, and use a binocular camera to switch binocular measurement and monocular measurement modes. High-speed image transmission is realized through Ethernet, and pose calculation is performed on the camera side to reduce the computing burden of the controller.
It achieves compatibility with multiple requirements for complex tasks, ensuring that the object detection and measurement functions can still be identified and measured when a single object fails, reducing resource consumption, and improving the efficiency of control system and the reliability of measurement results.
Smart Images

Figure CN115493513B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a vision system applied to a space station robotic arm, which can meet the monitoring and measurement requirements of the space station robotic arm in on-orbit missions. Background Art
[0002] Large space manipulators are indispensable tools in the construction of space stations. As an essential component of space manipulators, the visual system is responsible for the two main tasks of monitoring and measurement. All working modes of the manipulator, except for open-loop control, cannot be independently realized without the guidance and assistance of the visual system. During the on-orbit operation of the space manipulator, it is not only necessary to be able to move over a large range, but also to complete a series of precise operations with high precision and strong stability. Therefore, in order to meet the application requirements of large space manipulators, the visual system is required to be able to clearly image the surrounding space environment and space targets within a large-span observation range (from hundreds of millimeters to tens of meters), and at the same time, it can complete detection, identification and high-precision three-dimensional measurement of various space targets with large size changes.
[0003] At present, most of the large space manipulators that are currently in orbit or have been flight-verified abroad are also equipped with visual systems, but most of these systems have one or more of the following problems:
[0004] (1) The tasks that the robotic arm needs to perform are relatively simple, and the functional requirements for the visual system are relatively simple, which cannot adapt to the current complex space station tasks;
[0005] (2) Most cameras in the visual system are monocular. Once the monocular camera product fails, the current visual and monitoring functions will be lost;
[0006] (3) Two surveillance cameras are arranged on the arm near the elbow of the robotic arm to visually monitor the areas where the two ends are located, which consumes a lot of on-orbit resources;
[0007] (4) The information processing resources of the visual system are relatively concentrated, resulting in room for improvement in the efficiency of the control system when the data transmission bandwidth is limited. Summary of the invention
[0008] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and meet the monitoring and measurement requirements of the space station robotic arm in on-orbit missions.
[0009] The purpose of the present invention is achieved through the following technical solutions:
[0010] A visual system for a space station manipulator, comprising at least two wrist cameras, at least one elbow gimbal camera, at least one docking camera, a data communication device, and a data processing device;
[0011] At least one elbow PTZ camera is mounted on the elbow of the robotic arm for monitoring;
[0012] At least one wrist camera is installed on each wrist of the robotic arm to obtain the position information of the target and further guide the end of the robotic arm to approach the target;
[0013] At least one docking camera is installed on the docking part of the space station cabin to obtain the position information of the target and further guide the robotic arm to complete the cabin docking;
[0014] All wrist cameras and elbow gimbal cameras communicate data with the data communication device;
[0015] The data communication device and all docking cameras perform data communication with the data processing device, and communicate with the outside through the data processing device.
[0016] Preferably, the elbow gimbal camera drives the camera to move via the gimbal so as to always keep monitoring the target.
[0017] Preferably, the wrist camera is also used to guide the end of the robotic arm to capture, grasp and lock the target adapter.
[0018] Preferably, the elbow gimbal camera is a monocular camera; and the wrist camera and the docking camera are both binocular cameras.
[0019] Preferably, the wrist camera and docking camera both have monocular measurement and binocular measurement modes, and both can switch autonomously.
[0020] Preferably, the monocular measurement mode is used to measure the three-dimensional pose of the target, and the binocular measurement mode is used to synchronously capture images of a binocular camera and calculate the target pose by matching image features.
[0021] Preferably, the wrist camera and the docking camera are both used to independently calculate the real-time position and posture of the target.
[0022] Preferably, the wrist camera, docking camera, and elbow gimbal camera are all provided with a light source for active lighting.
[0023] Preferably, the data communication device adopts an Ethernet switch.
[0024] Preferably, the visual system also includes visual markers for the wrist camera and / or docking camera to acquire position information of the target.
[0025] Compared with the prior art, the present invention has the following beneficial effects:
[0026] (1) In view of the numerous tasks of the space station's robotic arm and the complex requirements for visual monitoring and measurement, the present invention proposes a visual system for the space station's robotic arm, which is compatible with multiple mission requirements such as cabin crawling, cabin segment transfer and docking, extravehicular cargo handling, space station extravehicular status inspection, and hovering aircraft capture, and has both position measurement and speed measurement capabilities.
[0027] (2) In response to the reliability issue of monocular cameras, the present invention proposes that both the wrist camera and the docking camera have the function of autonomously switching between binocular vision measurement and monocular vision measurement. When both the left and right eyes are in monocular vision measurement mode, they have the function of autonomously optimizing the measurement mode. In both measurement modes, they have the function of autonomously judging the effectiveness of the measurement. When a point on the visual mark is missing or invisible, they still have the functions of target detection and recognition and three-dimensional posture measurement.
[0028] (3) The present invention addresses the problem that both end operations of the robotic arm require visual monitoring. By configuring a single dual-axis pan-tilt head at the elbow joint of the robotic arm, a large-scale monitoring is achieved with a relatively small number of cameras, thereby effectively enhancing the monitoring capability of the visual system.
[0029] (4) In order to solve the problem of large amount of data transmitted by the robot arm's visual camera image, the present invention adopts Ethernet as the communication channel for image data transmission, thereby realizing synchronous high-speed transmission of multiple visual images; at the same time, the camera is equipped with an image processing circuit and a posture solution circuit, and the posture calculation function is realized on the camera side, which does not occupy the computing resources of the robot arm controller and improves the efficiency of the control system; the measurement results output by the camera carry measurement status identification information and verification information, thereby improving the reliability of the control system using the measurement results. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 A block diagram of the visual system.
[0031] Figure 2 Visually mark the target adapter with pattern and size.
[0032] Figure 3 Mark the pattern and size for docking camera vision.
[0033] Figure 4 Information flow diagram of the visual system. DETAILED DESCRIPTION
[0034] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.
[0035] Embodiment 1:
[0036] The visual system described in the present invention can monitor the working area of the robotic arm at medium and long distances; at close distances, it can output the high-precision three-dimensional posture of the target, guide the end effector of the robotic arm to further adjust its own posture to gradually reduce the relative posture between the end effector coordinate system and the target adapter coordinate system, enter the capture area of the target adapter and complete the capture, grasping, locking and other operations of the target adapter; the docking camera can output the high-precision three-dimensional posture of the visual mark on the cabin, and guide the robotic arm to complete the transfer control of the cabin.
[0037] like Figure 1 and 4 As shown, the vision system includes wrist camera A, wrist camera B, elbow gimbal camera, docking camera, Ethernet switch, wrist camera visual marker, and docking camera visual marker.
[0038] Wrist camera A is installed on the end A, and wrist camera B is installed on the end B. They observe the operation target area of the end B, which can ensure that the space station's robotic arm can achieve the crawling task on the cabin. Both wrist camera A and wrist camera B use binocular cameras.
[0039] The elbow gimbal camera is installed at the elbow of the robotic arm. The elbow camera fixed on it is driven to move by the dual-axis gimbal to achieve large-scale monitoring of the monocular elbow camera.
[0040] The docking camera is installed on the cabin to guide the robotic arm to complete the secondary docking of the cabin. The docking camera uses a binocular camera.
[0041] The Ethernet switch is installed at the elbow of the robotic arm to interconnect the Ethernet network of the robotic arm and the Ethernet network in the space station cabin. The Ethernet switch receives the image data output by the camera, performs real-time detection and control of the data flow, and forwards the image data storage to the Ethernet switch in the cabin.
[0042] The wrist camera, elbow camera, and docking camera can autonomously control the on / off and brightness of the light source based on the statistical results of the pixel grayscale values in different areas of the image, and autonomously calculate the exposure time according to the frame rate constraint. The wrist camera and docking camera can also use the information of the image area where the visual marker is located to perform local exposure control based on the detection results of the visual marker. The wrist camera, elbow camera, and docking camera are all equipped with a sunshade, and the internal lens adopts a stray light elimination optical path design and is sprayed with stray light elimination paint.
[0043] The wrist camera and docking camera have both monocular measurement and binocular measurement modes. The monocular vision measurement mode uses a single monocular camera to complete the three-dimensional pose measurement of the target. In the binocular vision measurement mode, the binocular cameras are used to synchronously capture images, and the target pose is calculated by matching image features. The binocular measurement mode can not only effectively improve the measurement accuracy and increase the pose measurement function of non-cooperative targets, but also has the function of monocular measurement backup. Even if one of the two cameras fails, it can still provide pose data to complete the control of the robotic arm.
[0044] The wrist camera and the docking camera are both equipped with an image processing circuit and a posture calculation circuit, which can independently realize the real-time posture calculation function at the camera end, and send the posture calculation results to the robot controller through the communication bus. The image processing circuit and the posture calculation circuit integrate the number of visual markers detected in a single-frame image, the processing time of a single-frame image, the posture calculation results of continuous multi-frame images, and other information to autonomously determine the validity of the posture measurement results. When it is determined that one of the left-eye camera or the right-eye camera cannot work normally, it will switch from the binocular vision measurement mode to the monocular vision measurement mode. When it is determined that the left-eye camera and the right-eye camera are both normal, it will integrate the current measurement mode, the communication status between the left-eye camera and the right-eye camera, the time-consuming time to resume normal operation, and other information to autonomously switch from the monocular vision measurement mode to the binocular vision measurement mode. The boot program of the image processing circuit and the posture calculation circuit is stored in a single-particle immune chip. The main program is stored in three copies in the memory. When starting, the boot program reads the main program and performs a two-out-of-three vote. When the data is found to be inconsistent due to the single-particle effect, it is written back. During the program operation, the key data is processed with triple-mode redundancy, thereby improving the radiation resistance. The image processing circuit and the posture calculation circuit are also equipped with a watchdog circuit, which can reset and recover autonomously when a single-particle event occurs.
[0045] The visual markers include two categories: wrist camera visual markers and docking camera visual markers, which are respectively installed on the target adapter and the space station cabin, and are used for wrist camera and docking camera to detect, identify and calculate the position and posture of targets at close range.
[0046] The wrist camera A, wrist camera B, elbow gimbal camera, or docking camera are all equipped with a light source, which actively illuminates the target.
[0047] The wrist camera A, wrist camera B, elbow gimbal camera, docking camera, Ethernet switch, and visual marker all have modular interfaces and can be replaced on-orbit.
[0048] The visual mark pattern is mainly a solid circle and annular pattern. The visual mark only uses black and white as the basic colors of the pattern (no other colors except black and white are used).
[0049] Target adapter visual markers such as Figure 2 As shown, a total of 8 solid circle / ring patterns were designed, 4 ring patterns were all distributed on the surface of the target adapter visual mark flat plate, and 3 of the remaining 4 solid circle patterns were distributed on the surface of the target adapter visual mark flat plate, and 1 was distributed on the top plane of the target adapter visual mark cylindrical protrusion.
[0050] Docking camera visual markers such as Figure 3 As shown, a total of 8 solid circle / ring patterns are designed, 6 ring patterns are all distributed on the flat surface of the target adapter, and of the remaining 2 solid circle patterns, one is distributed on the target flat surface and the other is distributed on the top plane of the cylindrical protrusion.
[0051] The calculation formula of the target 3D pose under monocular vision measurement is:
[0052]
[0053] where s i is a non-zero scale factor, (u i ,v i ) is the center of the visual marker point P i The corresponding image pixel coordinates, (X i ,Y i ,Z i ) is the center of the visual marker point P i The three-dimensional coordinates in the target coordinate system, A is the camera intrinsic parameter matrix, R is the 3×3 orthogonal unit rotation matrix, t is the three-dimensional translation vector, R and t represent the three-dimensional pose of the target. The three-dimensional pose of the target can be obtained by solving the equation group.
[0054] In the binocular vision measurement method, the coordinates of the marker in the binocular camera coordinate system are first solved according to the matching results of the camera model and the image features, and then the coordinates of the marker in the target coordinate system are matched with the target visual marker, that is, a one-to-one correspondence is established with the coordinates of the marker in the target coordinate system, and then the target posture is calculated.
[0055] The left camera coordinate system is used as the reference coordinate system. The intrinsic parameter matrices of the left and right cameras are A and l and A r , the relative position relationship between the two cameras is R c and t c (R c is a 3×3 orthogonal unit rotation matrix, t c is the three-dimensional translation vector), then the target three-dimensional pose calculation formula under binocular vision measurement is:
[0056]
[0057]
[0058] where s li 、s ri are all non-zero scale factors, (X ci ,Y ci ,Z ci ) is the three-dimensional coordinate of the center of the visual marker Pi in the left camera coordinate system, (u li ,v li ) and (u ri ,v ri ) are the image pixel coordinates of the center of the visual marker point Pi in the left camera and the right camera respectively.
[0059] Left camera coordinate system O c X c Y c Z c With the target coordinate system O o X o Y o Z o The conversion relationship between them can be expressed as
[0060]
[0061] Where (X ci ,Y ci ,Z ci ) is the three-dimensional coordinate of the center of the visual marker Pi in the left camera coordinate system, (X oi ,Y oi ,Z oi ) is the 3D coordinate of the center of the visual marker Pi in the target coordinate system, R is the 3×3 orthogonal unit rotation matrix, t is the 3D translation vector, R and t represent the 3D pose of the target. The 3D pose of the target can be obtained by solving the set of equations simultaneously.
[0062] The wrist camera, elbow camera, and Ethernet switch communicate with the manipulator controller through the 1553B bus, and the docking camera communicates with the space station information system in real time through the 1553B bus, realizing two-way data transmission of remote control commands and telemetry data (including the target 3D pose calculation results). The target 3D pose calculation results output by the camera carry measurement status identification information such as measurement mode, pose measurement validity, number of marker points based on, and verification information calculated by sum verification. The camera and Ethernet switch receive remote control commands through the 1553B bus respectively, complete camera imaging control, light source illumination control, and Ethernet switch data forwarding control; upload the target 3D pose data calculated in real time and related telemetry data of the camera, Ethernet switch and other equipment through the 1553B bus to provide necessary information for the on-orbit operation of the manipulator; the RT address of the wrist camera and docking camera can be set.
[0063] The visual system communicates with the space station information system in real time through the 100M Ethernet bus and the Ethernet switch, realizing real-time high-speed transmission of encoded image data. The instruments in the space station complete the decoding and display of the image, providing astronauts and ground operators with visual information in the working area of the robotic arm; the Ethernet port supports full-duplex data communication; image data is transmitted using the RTP protocol. The image communication protocol is transmitted using the RTP protocol.
[0064] In order to solve the complex illumination adaptability problems caused by the alternation of low-orbit illumination areas and ground shadow areas, the relative motion position between the camera and the target, and the diverse optical characteristics of the target background area, the present invention proposes that the wrist camera, elbow camera, and docking camera all have integrated light sources, and all have the function of autonomous light source control, multi-mode automatic exposure, manual exposure controlled by remote control commands, and stray light suppression capabilities, which effectively improves the imaging quality of the camera and further enhances the monitoring and measurement capabilities of the visual system.
[0065] In response to the demand for 10 to 15 years of on-orbit use, the camera and data communication device adopt a modular interface design and ergonomic design, and have the ability to be replaced on-orbit by astronauts outside the vehicle; the camera and data communication device have the function of on-orbit parameter injection and modification, and the function of on-orbit software injection, modification and design; the camera and data communication device jointly support the acquisition of multi-format images such as color compressed images and original grayscale images, thereby supporting on-orbit calibration of the camera and posture measurement using camera images on the ground.
[0066] In order to adapt to the harsh solar ultraviolet radiation, total ionizing dose, single particle effect and other radiation environment adaptability issues, the materials and processes of the camera lens film layer and visual marking paint layer meet the requirements for use in the space environment; the image processing circuit and posture solution circuit configured in the camera have the ability to autonomously detect and autonomously recover from faults caused by single particle events, which can ensure the reliability of posture calculation.
[0067] Embodiment 2:
[0068] A vision system applied to a large space station robotic arm can meet the monitoring and measurement needs of the space station robotic arm during on-orbit missions.
[0069] When the space station's robotic arm performs tasks on orbit, it goes through different stages, from long-distance positioning of targets to medium-distance approaching targets to close-range precise operation of targets. The visual system is required in each stage. The visual system consists of wrist cameras, elbow gimbal cameras, docking cameras, Ethernet switches, visual markers, etc.
[0070] During the operation of the manipulator on orbit, the elbow camera is far away from the cabin and can overlook the area below the manipulator from a bird's-eye view. The field of view is not easily blocked, which can provide a good on-orbit monitoring effect. The manipulator uses the elbow camera to perform safety checks before the manipulator starts, monitor the working status of the manipulator during operation, and monitor the cabin section transfer and secondary docking process. During the movement of the manipulator, the manipulator controller controls the movement of the elbow camera pan / tilt to drive the elbow camera fixed on it to follow the movement to always keep monitoring the target.
[0071] When the effective observation area of the target adapter enters the field of view of the wrist camera, the robotic arm uses the wrist camera to obtain high-precision three-dimensional pose information of the target adapter's visual markers, guides the robotic arm's end effector to further approach the target adapter and complete high-precision and stable precision operations such as capturing, grasping, and locking the target adapter.
[0072] When the effective observation area of the docking mechanism enters the field of view of the docking camera, the robotic arm uses the docking camera to obtain high-precision three-dimensional posture information of the docking camera's visual markers, guiding the robotic arm to complete the secondary docking of the cabin.
[0073] The present invention proposes a vision system applied to a space station robotic arm, comprising a vision system consisting of a wrist camera, an elbow gimbal camera, a docking camera, an Ethernet switch, and a vision marker.
[0074] Both the wrist camera and the docking camera have the function of autonomous switching between binocular vision measurement and monocular vision measurement; and in both measurement modes, when a point on the visual marker is missing or invisible, they still have the functions of target detection and recognition and 3D posture measurement.
[0075] Both the wrist camera and the docking camera integrate image processing circuits and posture calculation circuits, which can independently realize real-time posture calculation functions on the camera end without occupying the computing resources of the robotic arm controller, and send the posture calculation results to the robotic arm controller through the communication bus.
[0076] The elbow gimbal camera is installed on the elbow joint. The double-axis gimbal drives the elbow camera fixed on it to move, so that a large-scale monitoring of the monocular elbow camera can be achieved.
[0077] An Ethernet switch is configured to simultaneously receive image data output by multiple cameras and forward it to the space station cabin through an Ethernet bus, reducing the number of cables running through the cabin.
[0078] The camera pose calculation software, Ethernet switch operating system software and application software in the visual system all have the function of on-orbit injection modification and upgrade. Imaging control parameters, image data compression encoding parameters and pose measurement parameters support on-orbit injection changes.
[0079] All devices in the vision system adopt a modular interface design, which enables on-orbit maintenance and replacement.
[0080] The 100M Ethernet bus is used to achieve real-time high-speed transmission of image data.
[0081] The camera of the present invention has two image data output modes, namely, uncompressed image and compressed image. The uncompressed image supports on-orbit calibration of the camera, and the compressed image reduces the code stream to support simultaneous transmission of multiple images while satisfying bandwidth constraints.
[0082] The present invention has a redundant backup function. The wrist camera adopts a dual-machine hot backup mode under a monocular vision measurement mode, the elbow camera pan / tilt motor adopts a dual-winding backup mode, and the Ethernet switch adopts a dual-machine cold backup mode.
[0083] The visual system uses circles and rings to form a variety of visual marking patterns, realizing the measurement tasks of different cameras. The black and white visual markings are achieved by spray painting, and the paint and process are sufficient for long-term use in orbit.
[0084] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.
[0085] Although the present invention has been disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art may make possible changes and modifications to the technical solution of the present invention by using the methods and technical contents disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.
Claims
1. A visual system applied to a space station robotic arm, characterized in that: including at least two wrist cameras, at least one elbow gimbal camera, at least one docking camera, a data communication device, a data processing device, and visual markers; At least one elbow PTZ camera is mounted on the elbow of the robotic arm for monitoring; At least one wrist camera is installed on each wrist of the robotic arm to obtain the position information of the target and further guide the end of the robotic arm to approach the target; At least one docking camera is installed on the docking part of the space station cabin to obtain the position information of the target and further guide the robotic arm to complete the cabin docking; All wrist cameras and elbow gimbal cameras communicate data with the data communication device; The data communication device and all docking cameras communicate with the data processing device and communicate with the outside through the data processing device; The visual markers are in the form of solid circles and circular rings. The visual markers include wrist camera visual markers and docking camera visual markers, which are installed on the target adapter and the space station compartment, respectively, and are used for the wrist camera and docking camera to perform target detection, recognition, and pose calculation. The target adapter visual mark is 8 solid circle / ring patterns, 4 of which are all distributed on the surface of the target adapter visual mark plate, 3 of the remaining 4 solid circle patterns are distributed on the surface of the target adapter visual mark plate, and 1 is distributed on the top plane of the target adapter visual mark cylindrical protrusion; the docking camera visual mark is 8 solid circle / ring patterns, 6 of which are all distributed on the surface of the target adapter visual mark plate, and 1 of the remaining 2 solid circle patterns is distributed on the surface of the target adapter visual mark plate, and 1 is distributed on the top plane of the cylindrical protrusion; The wrist camera and the docking camera are both binocular cameras; the wrist camera and the docking camera have both monocular measurement and binocular measurement modes, and can switch autonomously; The wrist camera, docking camera, and elbow gimbal camera are all equipped with light sources for active lighting, and are capable of multi-mode automatic exposure, manual exposure controlled by remote control commands, and stray light suppression; the wrist camera, elbow gimbal camera, and docking camera autonomously control the on / off and brightness of the light source based on the statistical results of the pixel grayscale values in different areas of the image, and autonomously calculate the exposure time according to the frame rate constraint. The wrist camera and docking camera also perform local exposure control based on the information of the image area where the visual mark is located according to the detection results of the visual mark; the wrist camera, elbow gimbal camera, and docking camera are all equipped with a sunshade, and the interior of the lens adopts a stray light elimination optical path design and is sprayed with stray light elimination paint; The wrist camera and the docking camera are both equipped with an image processing circuit and a posture calculation circuit, which can independently realize the real-time posture calculation function on the camera end, and send the posture calculation results to the robot controller through the communication bus; the image processing circuit and the posture calculation circuit integrate the number of visual markers detected in a single-frame image, the processing time of a single-frame image, and the posture calculation results of continuous multiple frames of images to autonomously judge the validity of the posture measurement results; when it is determined that one of the left-eye camera or the right-eye camera cannot work normally, it will autonomously switch from the binocular vision measurement mode to the monocular vision measurement mode; when it is determined that both the left-eye camera and the right-eye camera are normal, it will integrate the current measurement mode, The communication status between the left camera and the right camera and the time-consuming information for resuming normal operation can be switched autonomously from the monocular vision measurement mode to the binocular vision measurement mode; the boot program of the image processing circuit and the posture calculation circuit is stored in a single-particle immune chip, and the main program is stored in three copies in the memory. When starting, the boot program reads the main program and performs a two-out-of-three vote. When data inconsistency caused by a single-particle effect is found, it is written back. During the program running, triple-mode redundant processing is performed on key data, thereby improving the radiation resistance; the image processing circuit and the posture calculation circuit are also equipped with a watchdog circuit, which can be reset and restored autonomously when a single-particle event occurs.
2. The visual system according to claim 1, characterized in that The elbow pan-tilt camera drives the camera to move via the pan-tilt to always keep monitoring the target.
3. The visual system according to claim 1, characterized in that The wrist camera is also used to guide the end of the robotic arm to capture, grasp and lock the target adapter.
4. The visual system according to claim 1, characterized in that: The elbow gimbal camera adopts a monocular camera.
5. The visual system according to claim 1, characterized in that: The monocular measurement mode is used to measure the three-dimensional position and posture of the target, and the binocular measurement mode is used to synchronously collect images of binocular cameras and calculate the target position and posture through image feature matching.
6. The visual system according to any one of claims 1 to 5, characterized in that: The data communication device adopts an Ethernet switch.
Citation Information
Patent Citations
Butt joint ring capture point measuring method and system based on binocular stereoscopic vision
CN112556658A
