74results about How to "Accurate pose" patented technology

The present invention is applicable to the technical field of unmanned aerial vehicles, and provides a visual target tracking method for an unmanned aerial vehicle. The method comprises the following steps: step A, conducting FAST angular point extraction for a first target region selected for tracking, generating a weighted FAST angular point histogram for the first target region, and calculating, according to the generated FAST angular point histogram, a FAST angular point feature similarity between the first target region and candidate target regions selected for tracking; step B, calculating a color feature similarity between the candidate target regions selected for tracking and the first target region, fusing the FAST angular point feature similarity and the color feature similarity, and using a candidate target region which has a highest similarity to the first target region according to the fusing result as a tracking result for the current target region. According to the present invention, the FAST angular point feature and the color feature are fused, thereby effectively using partial information and global information of the image, and thus reducing the impacts caused by a vehicle-mounted sensor to target state estimation.

The invention relates to a Multi-Camera / Lidar / IMU-based multi-sensor SLAM method, which comprises the following steps: performing tight coupling joint initialization on data of a plurality of images obtained by a multi-view camera and data obtained by an IMU inertial measurement unit to obtain an initial pose of a system; acquiring point cloud data of a laser radar frame by a laser radar sensor, preprocessing the point cloud data, and dividing the point cloud into a strong angular point, a weak angular point, a strong plane point and a weak plane point; optimizing the pose of the laser radar frame through the initial pose of the system; optimizing the pose of the laser radar frame further through closed-loop detection; and carrying out map splicing by using the optimized laser radar framepose. According to the invention, the data obtained by the three sensors are fused, the obtained laser radar frame pose is more accurate, and the robustness and stability of the positioning system areeffectively improved.

The invention is suitable for the machine vision field and especially relates to a robot, and a pose estimation method and a device based on a marker. The method comprises the following steps of acquiring an image shot by a camera and identifying the marker; acquiring the image coordinate of the characteristic point of the marker, according to a demand, establishing a world coordinate system, andthrough measurement, acquiring the coordinate of the characteristic point of the marker in the world coordinate system; and according to a corresponding relation among the inner parameter of the camera, the position of the robot, the yaw angle of the robot, the offset of the robot and the coordinate of the characteristic point in the world coordinate system, calculating the yaw angle of the robotand the offset of the robot. In the invention, a p3p algorithm is not adopted so that the redundant pitch angle and the roll angle of the robot which only needs two-dimensional plane navigation do notneed to be calculated, and a calculated pose is accurate.

The invention discloses an odometer method and device for sports equipment; the odometer method comprises: acquiring current pose of sports equipment, current second laser frame and a local point cloud map; calculating according to the current pose, the second laser frame and the local point cloud map to obtain a current optimal pose; and other steps. In the odometer method, initial relative exercise amount and pose are acquired through the number of turns of a coded disc, laser inter-frame matching is performed by means of iterative nearest neighbor algorithm and the like, and accurate relative exercise amount and pose are acquired accordingly; the pose acquired in the above step is used as an initial value to perform matching on the laser frame and the local point cloud map so as to obtain more accurate pose and exercise amount that are then output. The odometer device provided herein enables the poor precision and accumulative error defects of an existing odometer to be overcome, and is applicable to general or special computing system environments or configurations.

The invention discloses a wireless camera sensor network position and posture calibration method for a wireless camera sensor network, and the method comprises the steps: enabling an unmanned plane carrying a photoelectric identification to enter a region covered by the wireless camera sensor network; taking a plurality of pictures comprising the photoelectric identification at equal intervals in a recognition range of one wireless camera sensor; carrying out the decoding of the photoelectric identification, extracting the pictures which have no overlapped view and are taken by the wireless camera sensor at three different moments, solving and obtaining the position and posture information of the wireless camera sensor; carrying out the feature matching of the wireless camera sensors which have an overlapped view, calculating an intrinsic matrix, decomposing and obtaining a rotation matrix; solving and obtaining the position and posture information of the wireless camera sensors which have the overlapped view, and determining the rotation matrix and a translation amount. The method achieves the quick, simple and precise calibration of the wireless camera sensor, and the wireless camera sensor network calibrated through the method can be quickly applied to disaster relief and other situations where the cameras need to be calibrated.

The invention discloses a transformer substation inspection robot positioning and navigation system integrated with multiple sensors. The system is divided into nine functional modules including a positioning module, an obstacle avoidance module, a map management module, a voice control module, an electric quantity monitoring module, a task scheduling module, a navigation module, a bottom layer control module and a safety guarantee module. According to the invention, a laser radar odometer and a visual mileage calculation method are introduced to a substation inspection robot for the first time; the positioning stability and accuracy of the outdoor substation inspection robot can be effectively improved; and meanwhile, a depth camera is introduced into the transformer substation inspectionrobot, low obstacles or pit obstacles in front of the robot can be conveniently obtained, the compatibility and safety of the robot to various environments are improved, and meanwhile the robot has higher automatic navigation and obstacle avoidance capacity in the inspection task.

The embodiment of the invention discloses a pose determination method and device, a storage medium and electronic equipment. The method comprises the following steps: acquiring current position information and a point cloud frame; and determining a local map point cloud according to the position information, determining an attitude angle of the electronic device according to the point cloud frameand the local map point cloud, and performing point cloud matching according to the position information, the point cloud frame, the local map point cloud and the attitude angle to obtain pose information of the electronic device. By adopting the embodiment of the invention, the accuracy of point cloud matching is improved, and the accurate pose of the electronic equipment can be obtained.

The invention is concerned with the computerized controlling double dimension positioning and double order tracking digitization operation system. It mainly consists of first deck (-), second deck (-), full-visual lens groupware (-), and the center control computerized unit, which includes the centerized computer, double dimension positioning and double order tracking module. The invention conducts the digitization scaling to the operation space using positioning and marking points up and down the decks, this way, it ensures the acrruate location of the operation position according to the two marked X-ray images are collectable by model C arm; the double order tracking theory, as well, tracks and identifies the video image to ensure the space location with marked project.

The invention discloses a method and a system for measuring the pose of an industrial mechanical arm. The method comprises the following steps: a target mounting step: mounting at least three targetsat the tail end of the industrial mechanical arm, so that at least three targets and a set point of the axle wire of an arm stand of the mechanical arm form a preset position relationship; a parameteracquiring step: acquiring the position information of the at least three targets on a mechanical arm carrier coordinate system and the structure size parameters and the mounting size parameters of the targets; and a calculation step of the pose of the tail end of the mechanism arm: calculating the pose information comprising position coordinate and posture of the tail end of the mechanical arm onthe mechanical arm carrier coordinate system according to the preset position relationship, the acquired position information and the structure size parameters and the mounting size parameters. Through the adoption of the method, the position and the posture of the mechanical arm can be accurately measured in real time, and operating personnel are assisted to control the movement of the mechanical arm to complete the job task well.

The invention relates to a method and a system for extracting and matching feature points in the simultaneous positioning and mapping technology. The method extracts n FAST corners from the first frame of the key frame image, and extracts BRIEF features from the FAST corners which are larger than the first k of the response value. For each frame from the second frame to the last frame, k FAST corners are extracted and BRIEF features are extracted, and the k FAST corners are matched with the feature points extracted from the 3D point cloud or the previous frame as feature points, and the camerapose is calculated. In this method, the matching process of some feature points is replaced by optical flow to improve the running speed, and the corner points of unextracted BRIEF feature are tracked by optical flow method, which saves the time of extracting descriptor from feature points. The invention ensures the extraction speed of the characteristic points, and at the same time, is conduciveto reconstructing more dense point clouds, and makes the distribution of the characteristic points more uniform.

The invention discloses a laser and visual information fused robust positioning and mapping method and system. The method comprises the following steps: performing fusion calculation on a non-visual pose Ttm; estimating a visual pose Ttv according to the visual image, and if the estimation fails, performing non-visual auxiliary repositioning by using the non-visual pose Ttm to calculate the visual pose Ttv; and fusing the visual pose Ttv and the non-visual pose Ttm to obtain a fused robot pose Ttf, carrying out closed-loop detection, and carrying out global optimization on the robot pose by adopting a closed-loop optimization algorithm based on graph optimization after a closed loop is detected, so as to obtain a two-dimensional grid map and a three-dimensional point cloud map. According to the method, information of various sensors can be fused under the severe conditions that the robot moves too fast, the illumination environment is extreme, and scene texture features are lost, the problem that visual SLAM tracking is lost is solved, the positioning precision of the robot is improved, and meanwhile, an accurate two-dimensional grid map and an accurate three-dimensional point cloud map are constructed.

The invention discloses a laser radar point cloud motion distortion correction method, and belongs to the technical field of unmanned vehicle automatic driving. The method aims at solving the problem that in laser radar point cloud motion distortion correction, the displacement accumulative error of an external sensor is large. The method comprises the following steps: adopting an RTK as a pulse generator, and synchronously triggering a laser radar and an IMU; establishing a global coordinate system for the first frame point cloud of the laser radar; establishing a local coordinate system for non-first frame point clouds; converting the corresponding IMU coordinate system into the local coordinate system; obtaining the IMU pose synchronously; obtaining an IMU pose corresponding to each data point in the point cloud of the current frame by adopting an interpolation method, and correcting the motion distortion of the point cloud of the current frame of the laser radar according to the IMU pose; converting the local coordinate system to the global coordinate system, and calibrating the IMU pose by using the RTK pose; and performing next frame point cloud correction until correction of the last frame laser radar point cloud motion distortion is completed. According to the invention, the robustness and stability of laser radar point cloud correction are enhanced.

The invention discloses an indoor map construction method of a mobile robot. An odometer, WiFi equipment and a laser radar in an environment are adopted as sensing units to achieve the collection of WiFi fingerprint data, odometer data and laser data, the odometer provides a rough pose change of the robot, the WiFi fingerprint data is fused to provide a more precise pose, on the basis, laser datais combined to construct an indoor map, and the method is a method for accurately constructing an indoor environment map by fusing the data. WiFi data is fused in the indoor map construction process,a similarity-distance model is trained, fingerprint closed-loop detection is carried out, accumulated errors of the odometer are effectively reduced, map construction of the laser radar in a complex and changeable environment can be assisted, and the accuracy of indoor map construction is improved. The method can be widely applied to map construction in various complex environments, and has relatively high universality.

The invention discloses a double-shaft positioner mechanical structure. The double-shaft positioner mechanical structure comprises a base, a roll-over table and a worktable, and also comprises a horizontal servo motor and a horizontal RV reducer driven thereby and a vertical servo motor and a vertical RV reducer driven thereby, wherein one end of the roll-over table is mounted on an output shaft of the horizontal RV reducer, and the other end is supported on a bearing of the base; the vertical RV reducer is mounted on the roll-over table, and an output shaft of the vertical RV reducer is provided with a rotating seat; the worktable is fixed on the rotating seat and is driven to rotate when the vertical servo motor rotates; and the roll-over table is provided with a conductive block, the conductive block is connected with a ground wire, and the upper surface of the conductive block is always in contact with the lower surface of the worktable. According to the invention, the servo motors and the high-accuracy RV reducers are used to ensure the moving accuracy, so that workpieces fixed on the worktable can be accurately positioned, thereby ensuring welding quality.

The invention discloses a classical orbit three-dimensional spatial relationship construction method based on an OSG three-dimensional engine. Spatial relationships are directly established based on six elements of an orbital system, and geometrical angle changing processes of the right ascension of an ascending node, the orbit inclination, the perigee argument and the true anomaly can be dynamically displayed. An orbital element parameter representation process is based on the space geometry relationships, and satellite visual angle tracking and position calculation are not subjected to secondary coordinate transformation. The orbital elements can be visually displayed, and an orbital system (VVLH) visual angle positioning problem under a traditional model is solved by the introduction of two layers of new coordinate systems. Four attitude input interfaces of matrixes, Euler angles, Euler axial angles and the quaternion of a relative orbital system are provided by a spacecraft, and the spacecraft body axis and the orbit axis are displayed and customized through the spacecraft, the analysis of the correctness of the in-orbit attitude of the spacecraft is facilitated, and problems that the orbital elements cannot be visually displayed, the orbital elements need the secondary transformation, the attitude adjustment and visual angle adjustment of the orbital system require additional data, and the position and orientation of the spacecraft cannot be accurately displayed are solved.

The invention is applicable to the field of computer vision, and provides a pose estimation method and device based on a rectangular identifier and a robot. The method comprises the steps of carryingout distortion correction on image coordinates of the four corners of a rectangular sign through the internal parameters and the distortion parameters of a camera, outputting the four angular points after the distortion correction; randomly selecting three angular points from the four angular points after distortion correction to serve as solving points of the P3P algorithm, and the other corner point serving as a check point to generate four groups of points, for the case of the TS4, obtaining two sets of solutions for each set of points, carrying out verification through the remaining points, so as to obtain eight groups of solutions; and selecting the point with minimal reprojection error as the final estimated pose. Compared with the current P3P algorithm, in which three points are randomly selected to serve as calculating points and the remaining point as a verification point, the method has more theoretical completeness. Redundancy can not be caused; and the calculated pose is accurate.

The invention relates to a mobile robot autonomous charging method based on ArUco code guidance. The method comprises the steps: setting a charging contact and a monocular camera for a robot, and arranging a charging pile with an ArUco code in an operation environment of the robot; establishing an environment map, and marking the position and orientation of the charging pile; setting a navigationtarget point and a target posture, and planning a movement track of the mobile robot; recognizing an ArUco code on the charging pile, and obtaining an angular point coordinate of the ArUco code on theimage; combining the robot pose collected by the monocular camera with the linear speed and the angular speed provided by the robot to acquire an accurate mobile robot pose; and a PID method is adopted to control the robot to go to the charging pile for butt joint, and autonomous charging is carried out. Compared with the prior art, ArUco codes are used for guiding stably, and on the premise thatlight exists, the 100% charging success rate can be achieved; and guiding is smooth, and the robot can be in butt joint with the charging pile quickly and smoothly.

The invention relates to an object pose estimation method based on a deformation convolution network, and the method comprises the steps: S1, obtaining a color image and a depth image of a target object, inputting the color image of the target object into a trained instance segmentation network, and obtaining an instance segmentation result; S2, respectively cutting out a color image block and a depth image block containing the target object from the color image and the depth image according to the instance segmentation result, and converting the depth image block into a point cloud represented by a three-channel image; S3, setting a pixel value of an area, which does not contain the target object, in the color image block and the point cloud image to be 0, and then inputting the pixel value into the trained deformation convolution network to obtain a target object pose estimation result; and the receptive field of the deformation convolution network extraction features is concentratedin the region where the target object is distributed on the color image block and the point cloud image. Compared with the prior art, the method has the advantages of high precision, high efficiencyand the like.

The invention is suitable for the technical field of computers, and provides a pose estimation method, which comprises the steps of collecting a target image presenting a target object; inputting thetarget image into a pre-trained control point prediction model for processing to obtain coordinate information of a plurality of control points; determining attitude parameters of the target object based on the coordinate information of the plurality of control points, with the attitude parameters being used for conversion between a first coordinate system and a second coordinate system, the firstcoordinate system being a coordinate system taking the center of the target object as an original point, and the second coordinate system being a coordinate system of a camera for collecting a targetimage; and determining the pose of the target object according to the pose parameter. The converted attitude parameters between the coordinate system of the target object and the coordinate system ofthe camera are determined through the coordinates of the plurality of control points, so that the relative position and the rotation angle of the target object under the coordinate system of the camera can be obtained, and the more accurate pose of the target object can be obtained.

The invention provides a robot anti-interference positioning method based on sliding window pose map optimization, and the method comprises the steps: obtaining laser data, and carrying out the repositioning of a robot: judging a search range according to whether the robot has an initial pose, and determining the current pose of the robot; tracking real-time pose during robot movement: judging whether enough key frames are accumulated in the tracking process or not, if yes, deleting the earliest key frame, and storing the latest key frame; judging the current positioning state of the robot , if the current positioning is lost, obtaining the latest accurate positioning again, updating the positioning , and continuing pose tracking with the latest pose; otherwise, performing inner point screening and accumulated error elimination so as to allow the robot to continuously output poses in the map to achieve real-time accurate positioning . The method has the beneficial effects that global repositioning can be realized without long-time movement, and the anti-interference capability is improved.

The invention relates to a camera pose determination method and device and a virtual object display method and device. The method comprises the following steps: acquiring a current frame image, wherein the current frame image comprises an object to be tracked; determining a camera initial pose corresponding to the current frame image; projecting the three-dimensional model of the to-be-tracked object to a camera imaging plane according to the initial pose of the camera to obtain a projection image, wherein the three-dimensional model is obtained according to a target image with the highest similarity with the current frame image in a preset three-dimensional object database, and the target image is obtained by performing surrounding shooting on a preset three-dimensional marked object; and correcting the initial pose of the camera according to the difference between the projection image and the current frame image to obtain a target camera pose. By adopting the method, the camera pose can be accurately positioned.

The invention discloses a robot pose determination method, and the method comprises the steps: obtaining a two-dimensional code image collected by a camera of the robot and sensing data collected by an inertial sensor of the robot, and determining the mileage information of the robot in a preset duration, wherein the sensing data comprises an acceleration and an angular speed; determining a firstpose of the camera according to two-dimensional code information and a pose estimation function recognized from the two-dimensional code image, and determining a second pose of the inertial sensor according to the sensing data; performing tight coupling optimization processing based on the first pose and the second pose to obtain a third pose; and carrying out fusion processing on the third pose and the mileage information to obtain the pose of the robot. The invention further discloses a robot pose determination device, the robot and a computer readable storage medium. According to the invention, the accuracy of positioning the pose of the robot in a complex scene is improved.

The invention discloses an autonomous positioning method of an end effector of an under-constraint cable-driven parallel robot. A PCF algorithm is combined with a factor graph algorithm, the positionof the end effector is accurately positioned under the condition of not relying on external signal input by using a multi-sensor data fusion technology, the IMU inertial sensor data is processed by using a PCF method, and the posture of the end effector of the under-constraint cable-driven parallel robot is accurately obtained, and meanwhile, the measurement precision of the position of the end effector is improved; and moreover, the selected sensor is a low-cost IMU, a dynamometer, a motor tachometer and the like, and the low-cost can be realized while the same positioning precision of the end effector is achieved. According to the autonomous positioning method, the problem of the positioning accuracy of the end effector of the under-constraint cable-driven parallel robot at the low costand complex environment can be solved, oscillation suppression control of the under-constraint cable-driven parallel robot is realized, and the application range of the under-constraint cable-driven parallel robot is expanded.

The invention discloses a method and a device for drawing an electronic map. The method comprises the steps: obtaining an image sequence collected by an image collector on collection equipment; determining the pose when the image collector collects a frame of image for each frame of image in the image sequence, and enabling the pose to serve as the pose corresponding to the frame of image; taking positioning data obtained by a positioning sensor arranged on the collection equipment and a standard coordinate system corresponding to the positioning sensor as constraints, and optimizing the pose corresponding to the frame of image to obtain an optimized pose of the frame of image; constructing a point cloud space of a geographic area corresponding to the image sequence based on the optimized pose of each frame of image; according to the point cloud space and the image range of traffic elements in each frame of image, determining actual positions of the traffic elements in each frame of image, and drawing the traffic elements in an electronic map according to the actual positions of the traffic elements in each frame of image. In this way, the cost can be saved, and the accuracy of positioning the traffic elements is ensured.

The invention discloses a target detection method and a target detection device. The method comprises the following steps: acquiring image data, determining position information and size information of a target object in the image data for each target object contained in the image data, inputting the position information and the size information of the target object in the image data into a presetfirst pose estimation model, estimating pose information of the target object to serve as a first pose, determining a correction amount for the first pose according to the first pose and the image data, and correcting the first pose through the correction amount to obtain an actual pose corresponding to the target object. It can be guaranteed that unmanned equipment carries out safe driving basedon the determined accurate actual pose of the target object.

The invention relates to a pose determination method and device, electronic equipment and a storage medium, which are used for efficiently and accurately determining the pose of a target object in a video, and the method comprises the following steps: obtaining an initial pose containing a specified target object, a first video frame and a second video frame, and adopting two different calculation modes to calculating each feature point of the target object to obtain first two-dimensional coordinate information and second two-dimensional coordinate information of each feature point of the target object in the first video frame, and obtaining a target pose positioning result of the target object in the first video frame based on the first two-dimensional coordinate information and the second two-dimensional coordinate information.

The invention relates to an automatic library book taking and placing robot based on an ROS. The automatic library book taking and placing robot comprises a movable chassis, a lifting platform, a grabbing device, a multi-grid storage box, a touch screen and a power supply unit. The lifting platform, the multi-grid storage box, the touch screen and the power supply unit are all arranged on the movable chassis, and the grabbing device is arranged on the lifting platform. The power supply unit is in circuit connection with the movable chassis, the lifting platform, the grabbing device, the multi-grid storage box and the touch screen. The robot can create a map in a library, conduct automatic navigation and grab and place targets on all bookshelf layers, manpower and material resources needed by workers for placing books are saved, needed books are automatically found for readers and given to the readers, the time consumed for searching for books is shortened, the reading experience is enhanced, the contact between people in the public place is effectively reduced, and the probability of epidemic disease infection is reduced.

The invention discloses a pose tracking method, a pose tracking device and electronic equipment. The pose tracking method comprises the following steps: acquiring continuous multi-frame images of a scanned object and an initial pose of an image capture unit; taking the initial pose as an initial value, and based on a previous frame image and a current frame image in the continuous multi-frame images, using a first algorithm to obtain a first calculation pose of the current frame; taking the first calculation pose as an initial value, and obtaining a second calculation pose of a current frame by using a second algorithm based on the current frame image and a current frame reconstruction model; and updating the initial pose of an image capture unit according to the second calculation pose, and repeating the above steps to realize pose tracking of the image capture unit. The pose of the image capture unit can be accurately estimated, an approximately accurate pose can be provided when a scanned object deviates from a visual field range, and the pose tracking robustness of the image capture unit is improved.

The invention discloses an ORBSLAM2-based unmanned aerial vehicle autonomous navigation method, and belongs to the field of visual synchronous localization and map construction. The method is mainly composed of visual front-end and visual rear-end global optimization and mapping and three-dimensional path planning. The visual front-end mainly comprises ORB feature point extraction and matching and calculation of a current unmanned aerial vehicle pose by a visual odometer, and the part realizes a positioning function. Visual rear-end global optimization and mapping mainly comprise BA global graph optimization and dense point cloud image mapping, and the part realizes global optimization and mapping functions. Three-dimensional path planning realizes indoor and outdoor obstacle avoidance operation of the unmanned aerial vehicle, and the part realizes a path planning function. ORB pixel points are extracted by using improved FAST angular points, and ORB feature points are subjected to block optimization. Pose estimation is made by using an improved ICP-PnP algorithm.

The invention discloses a pose determination method and device for head-mounted equipment, equipment and a storage medium. The method comprises the following steps: acquiring a current frame image anda previous frame image of the current frame image; performing feature point extraction operation on the current frame image and the previous frame image to obtain a first feature point set of the current frame image and a second feature point set of the previous frame image; wherein the set feature point extraction operation comprises the steps of extracting an overall feature point set of a to-be-processed image; extracting a feature point set of a moving object in the to-be-processed image; removing a feature point set of the set moving object from the set overall feature point set, and outputting the feature point set; and determining whether the pose of the head-mounted device changes or not according to the first feature point set and the second feature point set.