41results about How to "Improve pose accuracy" patented technology

The invention discloses a visual SLAM method based on point-line fusion, and the method comprises the steps: firstly inputting an image, predicting the pose of a camera, extracting a feature point ofthe image, and estimating and extracting a feature line through the time sequence information among a plurality of visual angles; and matching the feature points and the feature lines, tracking the features in front and back frames, establishing inter-frame association, optimizing the pose of the current frame, and optimizing the two-dimensional feature lines to improve the integrity of the feature lines; judging whether the current key frame is a key frame or not, if yes, adding the key frame into the map, updating three-dimensional points and lines in the map, performing joint optimization on the current key frame and the adjacent key frame, and optimizing the pose and three-dimensional characteristics of the camera;and removing a part of external points and redundant key frames; and finally, performing loopback detection on the key frame, if the current key frame and the previous frame are similar scenes, closing loopback, and performing global optimization once to eliminate accumulated errors. Under an SLAM system framework based on points and lines, the line extraction speed and the feature line integrity are improved by utilizing the sequential relationship of multiple view angle images, so that the pose precision and the map reconstruction effect are improved.

The invention discloses a method used for measuring the pose of a non-cooperative target based on a complete light field camera. The method is applied in the non-cooperative target containing a star arrow docking ring. The star arrow docking ring on the non-cooperative target is a circle and is characterized in that step 1, the light field image of the non-cooperative target with circular features is obtained by adopting the compete light field camera; step 2, surface normal vectors, and the coordinates of the circular center of the circular features are calculated in the main lens coordinate system of the complete light field camera; step 3, false solutions are removed; step 4, the pose of the non-cooperative target is acquired. According to the invention, the false solutions based on the monocular vision pose estimations of the circle can be effectively removed, so that the pose of the non-cooperative target which meets accuracy requirements can be obtained.

A system and method for constructing fluoroscopic-based three dimensional volumetric data from two dimensional fluoroscopic images including a computing device configured to facilitate navigation of a medical device to a target area within a patient and a fluoroscopic imaging device configured to acquire a fluoroscopic video of the target area about a plurality of angles relative to the target area. The computing device is configured to determine a pose of the fluoroscopic imaging device for each frame of the fluoroscopic video and to construct fluoroscopic-based three dimensional volumetric data of the target area in which soft tissue objects are visible using a fast iterative three dimensional construction algorithm.

The invention discloses an anti-shake device of a mobile robotic vision system and an anti-shake compensation control method therefor, and relates to an anti-shake device of a vision system for a robot and an anti-shake compensation control method. Aiming at the shake of a mobile robot, machinery and control methods are used to fundamentally eliminate the shake of the robotic vision system in real time; therefore, the stable motion of the vision system can be achieved. The vision system anti-shake device comprises an anti-shake mechanism, a sensor system and an anti-shake control system which is used for realizing the anti-shake of the vision system. The method has the following steps: measuring the shake parameters, judging whether the mobile robotic vision system shakes or not, calculating the compensation amount, planning compensation track, giving feedback based on PD, calculating shake compensation control by the reverse dynamics The anti-shake device realizes the motion control of two vision sensors within a large range, and prevents and controls the shake generated by the two vision sensors. With the anti-shake compensation control method adopted, each pose parametric curve of the two vision sensors can accurately be close to the expected pose parametric curve.

The invention provides a method for optimizing off-line planned welding gun pose in robot pipeline welding, and belongs to the field of welding gun pose optimizing and adjusting in pipeline welding. According to the method, firstly, a structured light measurement system for off-line planned welding gun pose optimization is built, and system parameters are calibrated; then, pipeline welding path points are off-line planned; a welding gun is adjusted to a desired pose, a phase diagram corresponding to the pose is generated, visual features of visual servo are extracted from the phase diagram, and a visual servo control law is designed; and at each off-line planned path point, the pose of each path point is optimized based on the visual servo control law, and corresponding angles of robot joints are recorded. According to the method, the advantages that structured light measurement is high in precision and visual servo control is high in precision are combined, and the advantage that a robot is high in repeated positioning precision is fully utilized, so that the pose of the welding gun in the actual welding process can be effectively optimized.

The invention discloses a method for object identification and three-dimensional pose measurement of a complex surface, wherein the information is acquired by using a device for object identification and three-dimensional pose measurement of the complex surface; the method specifically comprises the following steps: setting up an object characteristic database by an offline process, wherein the object characteristic database comprises an object color characteristic database and an object depth characteristic database; and performing the object identification and acquiring a three-dimensional pose parameters on line. Due to the fusion of the depth and color characteristic data acquired by a body sensor and the combination of the object identification and the pose measurement, not only the repeated computation of intermediate parameters is avoided, but also the intelligence level is improved. As the depth object characteristic database and the color object characteristic database are set up by the offline mode, respectively, the efficiency of the object identification and the pose measurement is greatly improved. The invention aims to acquire higher pose precision under the condition that the resolution and precision of the image acquired by the body sensor are relatively low, and the real-time performance is improved.

The invention provides a drill bit position and posture optimizing method for robot drilling and belongs to the field of drill bit position and posture optimization and adjustment in robot drilling operation. The method comprises the following steps: firstly, building a robot drilling system based on a fringe projection method and visual servo control, performing system parameter calibration and performing off-line planning of robot drilling operation to obtain a series of route points; secondly, at each route point, calculating a phase diagram of the drill bit under the expected position andposture of the route point, and designing visual feature and deducing an interaction matrix and a visual servo control law; finally, optimizing the drill bit position and posture of each route point subjected to off-line planning by the visual servo control law until the visual servo convergence of each route point is completed to improve the perpendicularity of the axis of the drill bit and the section of a to-be-drilled point, so that the drilling quality and the surface flatness are improved. By the drill bit position and posture optimizing method for robot drilling, the advantages of highmeasuring precision of a fringe projection outline measuring method and high visual servo control precision are combined, and the position and posture of the drill bit in the actual robot drilling operation can be optimized effectively.

The invention discloses a square root volume fuzzy adaptive Kalman filtering SLAM (Simultaneous Localization And Mapping) method. The method comprises the steps of: modeling a mobile robot, to establishing a dynamic model and an observation model; by means of a fuzzy adaptive noise dynamic adjustment algorithm, setting control weights for motion noise and observation noise in the dynamic model andthe observation model and fuzzily adjusting the noise weights by dynamically adjusting innovation mean and variance; predicting pose information of the robot at the moment k through the pose information of the robot at the moment k-1; and updating the post information in a calculator after carrying out iteration preset times. The method has the beneficial effects that an iterative method is combined with strong tracking and aiming at the problems of the motion noise and the observation noise of the robot, different noises are adaptively dynamically adjusted by adopting an improved fuzzy adaptive method; and the complexity of the algorithm is reduced, the distortion problem of a sampling point in the nonlinear case is solved, the trajectory deviation phenomenon caused by the increase of feature points is corrected and the pose accuracy is increased.

The invention relates to a hydraulic control system for a parallel six-degree-of-freedom turntable. The hydraulic control system mainly comprises a motion planning computer, a direct control computer, a communication interface, a position and speed detecting part, a D / A (Digital / Analog) conversion and power amplifier and an electro-hydraulic servo valve. The main functions of all parts of a turntable control system are as follows: the motion planning computer is used for directly receiving an outside control command and parameter, monitoring the motion of a robot in real time, displaying information and sending various control commands to subordinate; the computer is directly controlled to receive the commands and the corresponding given information for the position and speed sent by the higher-level computer; the state value of the monitored object is detected in real time; the communication interface is an important part for transferring information between two-level computers; through the interface, a given value for a motion speed at a certain moment and necessary information for coordination and control of the two levels of computers are supplied to a speed control computer by a position control computer; the electro-hydraulic servo valve can control the flux and the flow direction through an electric signal, so that the speed, direction and size of the joint motion can be controlled.

The invention provides an aircraft landing pose estimation method based on vision-inertia tight coupling. The method comprises the steps of: an actual feature point detection process: acquiring video,enhancing an image, detecting a target and outputting a feature point, a synthesized feature point generation process: reading a pose parameter of an inertia measurement unit, reading geographic information of an airport landmark point and calculating a synthesized feature point, and a relative pose solution process: reading the actual feature point and the synthesized feature point, reading thegeographic information of the airport landmark point, and calculating a relative pose between an aircraft and a landing platform.

The present invention discloses a method and a device for measuring a visual pose of a simulated satellite on an air floating platform. The method includes the following steps: arranging a cooperativetarget, establishing a world coordinate system to obtain three-dimensional coordinates of the cooperative target; obtaining a current frame image in real time by using a camera module; processing theimage to identify the cooperative target, matching two-dimensional pixel coordinates with the three-dimensional coordinates to obtain a matching point pair; and obtaining the pose for the matching point pair by solving according to an PNP algorithm. The present invention has low cost and is simple to be implemented, and it can be ensured that the cooperation target is included a camera vision ina moving range of the analog satellite, so that the pose can be stably outputted in real-time.

The invention provides a tail-end pose detection system and method for a flexible parallel platform with photoelectronic packaging. The system is composed of a moving platform, a fixed platform, a moving platform connecting element, and five sets of pose detection units. The five sets of pose detection units are used for detecting and controlling pose values with six freedom degrees of the moving platform in real time. In addition, the method includes macroscopic pose detection and microcosmic pose detection. During the macroscopic pose detection, five laser displacement sensors detect the pose of the moving platform. The overall detection system has advantages of small size and low cost. The moving platform connecting element is arranged below the fixed platform without any contact, thereby avoiding any interference to the motion of the moving plane. Two laser displacement sensors are arranged outside the flexible parallel platform, thereby avoiding increasing of the pressure on the moving platform. Besides, during the microcosmic pose detection, adjusting is carried out by using an optical power meter and fine tuning is carried out based on the light transmitting value of the optical power meter.

The invention provides a pose optimization method, a pose optimization device, a computer readable storage medium and electronic equipment, and relates to the technical field of computer vision. The pose optimization method comprises the steps: acquiring multiple frames of scene images collected for a target scene; processing the first image sequence through a pose estimation model to obtain an image relative pose of the first image sequence; optimizing the image relative pose of the second image sequence by using the pose of the reference image in the scene image and the image relative pose of the first image sequence; and updating the pose of the scene image in the second image sequence according to the optimized image relative pose of the second image sequence. According to the method,the sensitivity of pose tracking to environmental changes is reduced, and the algorithm robustness and pose precision are improved.

A method of moving an object to a target position by an industrial robot is provided, wherein the robot comprises at least one arm and an effector unit mounted at the end of said at least one arm. The method comprises: causing the robot to pick the object at a first position by means of the effector unit and move the picked object to a second position; and in response to a deviation between the second position and the target position, causing the robot to correct the deviation by a motion of the effector unit.

The invention relates to a positioning method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the feature point matching of an image collected by image collection equipment with an offline map corresponding to an area where the image collection equipment is located, and obtaining a feature point matching pair which comprises a three-dimensional matching pair and a two-dimensional matching pair; and determining the pose of the image acquisition device according to the feature point matching pair. According to the embodiment of the invention, the positioning precision can be improved.

The invention discloses a two-rotation-freedom-degree one-translation-freedom-degree motion platform, and belongs to the field of motion simulation. The motion platform comprises an upper platform anda lower platform which are each provided with three hinge installation points distributed in an isosceles triangle shape, two isosceles triangles are similar, and the vertex angles of the two isosceles triangles are in the same direction; the perpendicular bisector of the height of the isosceles triangle on the upper platform is provided with a fixing point and a linear slide rail, and the linearslide rail is arranged in the perpendicular bisector of the height of the isosceles triangle; the upper ends and the lower ends of three pose adjusting cylinders are connected with the hinge installation points on the upper platform and the lower platform correspondingly; and the lower ends of two telescopic limiting pieces are fixed to the perpendicular bisector of the height of the isosceles triangle on the lower platform, and the upper ends of the two telescopic limiting pieces are connected with the fixing point and a slide block on the linear slide rail. The two-rotation-freedom-degree one-translation-freedom-degree motion platform can achieve three-freedom-degree motion simulation, that is, conducts rotation around the X axis and the Y axis and translation in the Z-axis direction, and accordingly is suitable for simulating rolling, pitching, heaving and other two-rotation one-translation occasions when a ship is moored on the sea.

The embodiment of the invention discloses a multi-sensor fusion vehicle positioning method and device based on map element data. The method comprises the following steps: estimating an external parameter matrix based on historical map element perception data recognized from historical sensing data acquired by a camera at the last moment and map element data in a vehicle-mounted electronic navigation map at the last moment, and performing vehicle positioning on a target vehicle at the current moment through the estimated external parameter matrix. Since the estimated external parameter matrix is calculated based on the data of the last moment, the accuracy of the estimated external parameter matrix is higher than that of a pre-calibrated external parameter matrix when the camera is installed loosely and / or the inertial measurement unit is installed loosely, and the accuracy of the estimated external parameter matrix is higher than that of the pre-calibrated external parameter matrix, sothat the pose precision obtained by carrying out vehicle positioning on the target vehicle at the current moment based on the estimated external parameter matrix is relatively high. By applying the scheme provided by the embodiment of the invention, the pose precision of the target vehicle obtained by positioning can be improved, and the driving safety of the target vehicle is improved.

A system and method for constructing fluoroscopic-based three dimensional volumetric data from two dimensional fluoroscopic images including a computing device configured to facilitate navigation of a medical device to a target area within a patient and a fluoroscopic imaging device configured to acquire a fluoroscopic video of the target area about a plurality of angles relative to the target area. The computing device is configured to determine a pose of the fluoroscopic imaging device for each frame of the fluoroscopic video and to construct fluoroscopic-based three dimensional volumetric data of the target area in which soft tissue objects are visible using a fast iterative three dimensional construction algorithm.

The invention is applicable to the technical field of robots, and provides a method and device for robot action replay. The method comprises the following steps of responding to a replay instruction;according to a replay time starting point included by the replay instruction, obtaining a moment of triggering the replay instruction and a plurality of historical sampling moments of the replay timestarting point; sequentially selecting an unselected historical sampling moment from the plurality of historical sampling moments to be used as a target historical sampling moment; obtaining a currentpose parameter of the current moment and historical pose parameters of the target historical sampling moments; generating a control parameter according to the current pose parameter of the current moment and the historical pose parameters of the target historical sampling moments; generating a control instruction according to the control parameter and the historical pose parameters; and executingthe control instruction. By using the mode, the goal of correcting the practical pose of the robot in the replay process through the current pose parameter in the replay process of the robot is achieved, and the pose precision of the robot action replay is improved.

The embodiment of the invention relates to an error estimation method, an integrated navigation processing system and a storage medium. The method comprises the following steps: acquiring a first observed quantity collected by an MEMS IMU under a first coordinate system; performing mechanical arrangement on the first observed quantity to obtain position data, speed data and attitude information of the carrier under the second coordinate system; acquiring state parameters according to the position data, the speed data and the attitude information; and estimating a three-dimensional misalignment angle error between the first coordinate system and the third coordinate system and a translation error of the first coordinate system and the third coordinate system in the X direction according to the position data, the speed data, the attitude information and the state parameters. By means of the mode, the application range of the non-integrity constraint is improved, and errors caused by the non-integrity constraint are reduced, so the purposes of improving the positioning and attitude determination precision of the system are achieved.

The invention discloses an odometer motion estimation method based on binocular vision, and relates to an odometer motion estimation method based on binocular vision. The invention aims to solve the problem that the robot pose obtained by the existing method is low in accuracy. The method comprises the steps that 1, a binocular camera collects images of a current scene at the position of the moment t and the position of the moment t + 1; 2, preprocessing the acquired image; 3, extracting feature points from the preprocessed image; 4, matching the extracted feature points; 5, obtaining the three-dimensional information of the current scene feature point at the moment t and the three-dimensional information of the current scene feature point at the moment t + 1; 6, acquiring the three-dimensional information of the current scene feature point at the moment t and the three-dimensional information of the current scene feature point at the moment t + 1 which are successfully matched; and 7,performing motion estimation on the obtained three-dimensional information of the feature points according to a motion estimation method. The method is applied to the field of intelligent mobile robot autonomous navigation.

The invention discloses a two-rotation-freedom-degree one-translation-freedom-degree motion platform, and belongs to the field of motion simulation. The motion platform comprises an upper platform anda lower platform which are each provided with three hinge installation points distributed in an isosceles triangle shape, two isosceles triangles are similar, and the vertex angles of the two isosceles triangles are in the same direction; the perpendicular bisector of the height of the isosceles triangle on the upper platform is provided with a fixing point and a linear slide rail, and the linearslide rail is arranged in the perpendicular bisector of the height of the isosceles triangle; the upper ends and the lower ends of three pose adjusting cylinders are connected with the hinge installation points on the upper platform and the lower platform correspondingly; and the lower ends of two telescopic limiting pieces are fixed to the perpendicular bisector of the height of the isosceles triangle on the lower platform, and the upper ends of the two telescopic limiting pieces are connected with the fixing point and a slide block on the linear slide rail. The two-rotation-freedom-degree one-translation-freedom-degree motion platform can achieve three-freedom-degree motion simulation, that is, conducts rotation around the X axis and the Y axis and translation in the Z-axis direction, and accordingly is suitable for simulating rolling, pitching, heaving and other two-rotation one-translation occasions when a ship is moored on the sea.

The invention provides a photoelectron packaging flexible parallel platform end pose detection system and method; the system comprises 6 pose detection devices used for detecting pose values of 6 degrees of freedom of a mobile platform; 6 sensors are respectively fixed on a support outside the mobile platform; only a detection baffle plate is arranged on the mobile platform so as to mitigate the pressure on the mobile platform, and preventing interferences on mobile platform motions; the sensor is small in size, easy to mount, and low in cost; the mobile platform pose detection of the flexible parallel platform comprises macroscopic pose detection and microscopic pose detection; the macroscopic displacement allows the pose precision to be controlled within 10 micrometers; 6 pose detection devices can detect real poses of the 6 degrees of freedom platform; microscopic pose detection can be adjusted by an optical power meter; the optical power meter can be used to observe light transmission amount data, thus carrying out fine adjustment.

The invention discloses a laser SLAM positioning system and method based on deep learning, and the system and the method comprises the modules and steps: 1, a point cloud preprocessing module which is used for extracting feature of ground points, planes and edges; 2, a laser speedometer module which is used for optimizing the pose in two steps, forming height constraint by ground points and optimizing the distance between the features; 3, a laser mapping module which is used for performing local optimization after down-sampling of adjacent frames, and optimizing mapping pose data in combination with odometer pose data; 4, a loopback detection module which is used for solving inter-frame similarity by using descriptors in a SegMatch framework and combining a twin neural network; and step 5, carrying out repositioning after the loopback is detected, and obtaining a corrected current pose after global optimization is carried out. Based on the method, the positioning precision of a mobile robot in indoor and outdoor environments is effectively improved.

The invention discloses an omni-directional mobile robot docking mechanism control system and method based on visual guidance. The control system comprises an upper computer, a controller, a driver, a photoelectric detection device, an encoder and a camera, wherein the controller is connected with the upper computer; the encoder is connected with the driver and then connected with the controller; the photoelectric detection device and the industrial camera are connected with the controller; and the encoder is fixed at the joint of a mechanical arm. The method comprises the steps that the industrial camera on an active moving module is turned on, mark information on a passive moving module is searched, then the information is calculated to obtain relative pose information of the active moving module and the passive moving module, the active moving module gradually gets close to the passive moving module according to a set moving sequence, when the two modules are close to a docking state, the two modules stop moving and enter the docking process, the tail end of the mechanical arm moves to the docking position on an active connector, and a locking mechanism locks a passive connector. The system and the method are accurate and reliable in docking, and the design difficulty of the docking mechanism is reduced.

The invention belongs to the related technical field of machining, and discloses a robot machining path optimization method based on deformation and dexterity. The robot machining path optimization method comprises the steps that the dexterity of a robot at a path point is obtained, so as to establish a function relation between the dexterity and a rotation angle and a joint angle of a robot joint; the cutting force obtained by a cutter at each path point is obtained, and the deformation amount of the robot is obtained according to the cutting force; weight values are distributed for the dexterity and the deformation to obtain a target function; the target function is solved within the constraint range of the joint angle, the dexterity and the deformation, and then the optimal rotation angle of the robot at the path point is obtained; and according to a robot inverse kinematics calculation method, the posture of the robot at the optimal rotation angle of the path point is converted, and the optimal rotation angle of all joints of the robot at the optimal rotation angle can be obtained. According to the robot machining path optimization method based on the deformation and the dexterity, the machining path of the robot under the constraint of the deformation and the dexterity can be obtained, and the machining path is more reasonable and accurate.

The invention discloses a real-time self-generating pose calculation method based on a variance component estimation theory, and belongs to the technical field of simultaneous localization and map construction. According to the method, firstly, initialization of a system is completed by utilizing vision measurement information, then three-dimensional space coordinates of feature points in an image are calculated according to imaging characteristics of a measurement device, then, clustering processing is performed, weight distribution of measurement information is calculated by utilizing variance component estimation, and finally, calculation results are substituted into a nonlinear optimization equation of the whole system. In the using process, the method is not only suitable for a visual navigation algorithm, but also has an optimization effect on a visual / inertial fusion navigation algorithm. By using the method to optimize the random error in the system, the accuracy and reliability of the system are improved.

A method of moving an object to a target position by an industrial robot is provided, wherein the robot includes at least one arm and an actuator unit mounted at an end of the at least one arm. The method includes causing the robot to pick up an object at a first location via an actuator unit and moving the picked object to a second location; and causing the robot to pass through the actuator unit in response to a deviation between the second location and a target location. Motion correction bias.