105 results about "Object motion" patented technology

The invention discloses an object level motion blur three-dimensional scene synthesis method and an object level motion blur three-dimensional scene synthesis device based on three-dimensional Gaussian splash (3D Gaussian splash, 3DGS). According to the method, a multi-view image and corresponding camera parameters are used as input, and three-dimensional Gaussian point scene representation including spatial position, scale, rotation, opacity and appearance characteristics is constructed; generating a binary mask of a moving foreground object and a static background by using an image segmentation model; on the basis, camera exposure time is dispersed into a plurality of time steps, space pose parameters changing along with time are introduced only for three-dimensional Gaussian points corresponding to the foreground object, instantaneous three-dimensional scene representation under each time step is obtained, integration or weighted averaging is carried out on rendering results of the plurality of time steps, and the real-time three-dimensional scene representation of the foreground object is obtained. Generating a motion blur rendering result with real physical significance; meanwhile, a mask-guided foreground-background separation rendering strategy and a partition loss function are adopted, clear rendering supervision is applied to a static background area, fuzzy rendering supervision is applied to a moving foreground area, and joint optimization of three-dimensional Gaussian scene parameters is achieved; in the reasoning stage, the clear background and the fuzzy foreground are fused and output according to the mask, and a three-dimensional scene reconstruction result with the clear background and the reasonable motion fuzzy effect of the foreground is obtained. According to the method, the exposure integral process and the object movement track are explicitly modeled in the three-dimensional geometric space, and object-level controllable fuzzy three-dimensional scene generation is realized.

The invention provides a spatial intelligent visual physical process inference method based on an implicit physical large model, and belongs to the field of spatial intelligent and artificial intelligence modeling calculation. The problems of low prediction precision and lack of physical consistency for complex physical scenes in the prior art are solved. The method comprises the following specific steps: acquiring multi-modal data of an environment, and representing the multi-modal data in a unified coordinate system; preprocessing the multi-modal data, designing a geometric coding model, a dynamic prediction model and an energy conservation constraint, introducing a space-time attention mechanism, predicting the motion state of an object according to the multi-modal data, and obtaining prediction data of the motion state of the object; acquiring real observation data of object motion, comparing the difference between the prediction data and the real observation data, and performing correction and weight updating on the prediction data of the model through a self-adaptive residual term; according to the method, visual information and implicit physical law modeling are fused, and the self-supervised physical consistency constraint is utilized, so that automatic learning and prediction of a potential physical process are realized.

Certain aspects of the present disclosure provide techniques and apparatus for accurately predicting the movement and location of objects in a multidimensional space. An example method generally includes receiving an input defining a location of an object in a three-dimensional space at a first time. The input generally includes a plurality of tokens. Using a transformer neural network, a predicted movement for tokens in the plurality of tokens is generated. Generally, the predicted movement includes a translation and a rotation associated with individual tokens in the plurality of tokens. The predicted movement is translated into an object-wise predicted movement. Based on the object-wise predicted movement and the location of the object in the three-dimensional space, a predicted location of the object at a second time subsequent to the first time is output.

The invention discloses a real-time three-dimensional modeling and hanging object size identification method and system applied to a tower crane, and belongs to the technical field of image processing, and the method comprises the following steps: obtaining a corrected real-time three-dimensional model, thereby obtaining a hanging object simulation dynamic swing amplitude; the movement envelope volume of the hanging object is obtained; whether static obstacle interference exists or not is judged, if yes, early warning is carried out, and if not, hoisting task construction is carried out; when hoisting task construction is carried out, whether dynamic obstacle interference exists or not is predicted in real time, if yes, corresponding speed adjustment is carried out, the obstacle avoidance speed is obtained for initial configuration, and if not, speed adjustment is not executed; according to the method, the mechanical performance influence parameters of the tower crane are obtained, the mechanical performance influence value is obtained through analysis, the obstacle avoidance speed is corrected, the correction execution obstacle avoidance speed is obtained, and actual configuration is carried out, so that the problem of inaccurate obstacle avoidance speed adjustment caused by inaccurate suspended object size analysis in the prior art is solved; the method is suitable for a tower crane construction scene and a port container lifting remote control scene.

The embodiment of the invention provides a man-machine interaction method and device, equipment, a medium and a program product. The method comprises the steps that target media data are displayed in a target interface, and the target media data are currently played media data; in response to a starting operation of a target service, object motion information is determined according to device attitude data sent by an attitude data acquisition device, the target service comprises interaction configuration information, and the interaction configuration information represents candidate object motion information and a candidate interaction mode; and determining a target interaction mode according to the object motion information and the interaction configuration information, and executing an interaction operation on the target media data according to the target interaction mode. According to the embodiment of the invention, man-machine interaction is controlled based on different motion information of the object, the user is prevented from being in the same posture for a long time, and the user is guided to do proper motion and healthily and properly use the application program.

The application provides a kind of progressive motion-power alignment dexterous operation fast migration method and system. It is related to the technical field of robot operation control, the method of the application comprises: extracting hand-object motion and interaction information from RGB hand-object interaction video;Based on kinematic matching, the human hand operation trajectory is mapped to the primary control signal of the target dexterous hand;Under the reinforcement learning framework, the primary control signal is dynamically optimized in combination with a multi-level reward mechanism to generate a stable grasping control strategy;According to the object target trajectory, the dexterous hand wrist control signal is synthesized to maintain the consistency of the operation task semantics;Through multi-dexterous hand configuration adaptation mechanism, the control signal is migrated to different configurations of dexterous hand.The application can automatically convert human hand operation trajectory into control instructions of multiple types of dexterous hand, enabling it to perform complex tasks such as grasping, moving and operating, thereby reducing the cost of acquiring training data for dexterous hand and improving the operation reusability and generalization ability across devices.

A method and system for evaluation of object motion repetitiveness as a basis to distinguish between human motion and non-human motion, in order to facilitate control of device operation. An example method includes (i) a computing system receiving sensor data representing object motion detected by at least one sensor, the object motion defining motion of an object, (ii) the computing system making a determination, based at least on an evaluation of the received sensor data, of whether the detected object motion is repetitive, and (iii) based at least on the determination being that the detected object motion is not repetitive, the computing system responding to the detected object motion by causing a device to take an action corresponding with a prediction that the object is a human being.

The application relates to the technical field of computer vision, and discloses a target object motion state analysis method, device and equipment, the method comprising the following steps: reconstructing based on multiple images shot by a camera during target object motion to obtain multiple position points; determining an initial motion interval of the target object based on a preset starting point and a preset landing point; correcting the initial motion interval to obtain a target motion interval; performing trajectory fitting based on the target motion interval to obtain a flight trajectory; and performing prediction based on the flight trajectory to obtain a predicted landing point. The application preliminarily determines the initial motion interval by extracting the position information of the target object, thereby obtaining an accurate target motion interval; the flight trajectory is obtained by performing trajectory fitting based on the target motion interval, which is helpful for intuitively understanding the motion state of the target object; the landing point of the target object is predicted based on the trajectory, and the accuracy of trajectory fitting and landing point prediction is improved, so that accurate analysis of the motion state of the target object is realized.

This disclosure provides systems, methods, and devices that utilize machine learning models to determine corresponding spatial positions and motion trajectories for images. In one aspect, a method is provided that includes receiving an image of a scene captured by a camera; determining, with a first machine learning model, a plurality of position values relative to the camera for at least a subset of pixels within the image; and training a second machine learning model based on the determined position values. The method further includes determining basis trajectories based on movement of the positions relative to previous image frames, and determining, for each of a subset of pixels, a movement trajectory relative to the previous frames as a weighted combination of the basis trajectories. These techniques can be employed as part of a structure-from-motion pipeline to estimate multi-view three-dimensional geometry, camera poses, and per-pixel object motion. Additional aspects are provided.

A control apparatus that obtains a correction amount for correcting a motion blur attributed to a motion of an object is disclosed. The apparatus computes a first motion blur amount based on a position of an object region in a captured image and a second motion blur amount based on an object vector indicating a motion of an object between captured images. The apparatus obtains a motion blur correction amount by adding the first motion blur amount and the second motion blur amount with use of weights corresponding to relative reliabilities of the first motion blur amount and the second motion blur amount.

This invention provides a three-dimensional measurement method for the surface of a moving object on a production line based on phase-shifting profilometry. Employing a three-dimensional vision measurement system composed of a projector and a camera, and based on the measurement principle of phase-shifting profilometry, the method analyzes the phase error caused by the motion of the measured object and proposes a three-dimensional measurement method for the surface of a moving object that does not require phase error correction. By setting the installation position and angle of the projector and camera, the phase encoding direction in the projected stripe pattern is perpendicular to the direction of motion of the measured object, thereby eliminating the phase error caused by motion. This allows the three-dimensional surface measurement results of the measured object under motion conditions to essentially achieve the measurement accuracy of the measured object under static conditions, realizing high-precision online real-time measurement of the surface three-dimensional morphology of moving objects on actual production lines.

A method, a device (800), a medium for controlling the movement of a virtual body (107), including: a target determining step, in the first frame, based on the moving direction and position of the virtual body (107), selecting a virtual object (103) in a predetermined area as a virtual target (202); a dynamic adjusting step, calculating a horizontal moving direction (302a) of the virtual body (107), a direction of a connecting line between a vertical projection (107′a) of the virtual body (107) on the horizontal plane and a vertical projection of the virtual target on the horizontal plane, and an included angle (θ1) between the horizontal moving direction (302a) of the virtual body (107) and the direction of the connecting line in each frame, and controlling the horizontal moving direction (302a) of the virtual body (107) based on the included angle (θ1) (204).

This application relates to a method, apparatus, and computer equipment for generating virtual object motion. The method includes the steps of: acquiring motion description text (202); analyzing the motion description text at a predetermined number of semantic levels by semantic analysis to acquire motion description information for each of the multiple semantic levels and acquiring a sampling noise signal for generating virtual object motion (204); encoding the motion description information for each of the multiple semantic levels to acquire a motion description representation for each of the multiple semantic levels (206); and performing noise reduction processing on the sampling noise signal based on the motion description representation of the first semantic level to acquire a motion feature vector output from the first semantic level (208); and multiple semantic levels Step (210) is to obtain a cascaded noise-reduced operation feature vector that has gone through multiple semantic levels by performing noise reduction processing on the sampling noise signal based on the operation feature vector output from the previous semantic level and the operation description representations of at least two semantic levels from the first semantic level to the current semantic level, wherein the operation granularity level represented by the operation feature vector output from the noise reduction processing of multiple semantic levels decreases in order from the highest semantic level to the lowest semantic level, and Step (212) is to obtain virtual object operation by decoding the cascaded noise-reduced operation feature vector.

For motion-compensated reconstruction in an imaging process, projection images (7) generated by the imaging process are obtained, representing an object from a multitude of different projection directions. Depending on the projection images (7), three-dimensional first image data (8) are reconstructed. Consistency data (9), which quantifies data consistency for each pair of projection images (7), are generated. Depending on the consistency data (9) and the three-dimensional image data (8), input data (10) are generated. Motion data (12), which characterizes object movement and / or motion artifacts in the first image data (8), are generated by applying a trained machine learning model (MLM) (11) to the input data (10). Depending on the projection images (7) and the motion data (12), three-dimensional, at least partially motion-compensated second image data are reconstructed.

The invention provides an object motion trail prediction method and device, and belongs to the technical field of computer vision and image processing, and the method comprises the steps: constructing an observation frame set of a target object; performing three-dimensional reconstruction on a target object by using an initial frame in the set to obtain an initial three-dimensional Gaussian particle set; predicting a particle set corresponding to each observation frame through a deformation auxiliary network according to the related information of the initial three-dimensional Gaussian particle and the timestamp corresponding to the target observation frame; performing speed prediction by using the three-dimensional position of the particle corresponding to each observation frame to generate a speed field of the particle in the current observation frame, and updating the three-dimensional position of the particle according to the speed field to obtain a particle set corresponding to a future frame; the particle set corresponding to the future frame is rendered to obtain a rendered image, the difference between the rendered image and the real image is calculated to optimize the particle set corresponding to the future frame, and the particle set is the predicted object motion trail. Based on the method, the invention further provides corresponding equipment. The method can effectively predict the sudden motion.

An object interference checking technique using point sets which uses CAD models of objects and obstacles and converts the CAD models to 3D points. The 3D point locations are updated based on object motion. The 3D points are then converted to 3D grid space indices defining space occupied by any point on any object or obstacle. The 3D grid space indices are then converted to 1D indices and the 1D indices are stored as a set per object and per position. Swept volumes for an object are created by computing a union of the 1D index sets across multiple motion steps. Interference checking between objects is performed by computing an intersection of the 1D index sets for a given motion step or position. The 1D indices are converted back to 3D coordinates to define the 3D shapes of the swept volumes and interferences.

The embodiment of the invention discloses a spatial reasoning data set construction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a plurality of target videos, the target videos being videos with object motion; for each target video, determining a three-dimensional space trajectory of each object in the target video; on the basis of the three-dimensional space track, questions and answers corresponding to the questions are generated for the multiple motion attributes, the questions and the answers corresponding to the questions serve as question and answer data pairs, and the motion attributes indicate changing attributes of the object in the three-dimensional space; and generating a spatial reasoning data set based on the target video and the plurality of question and answer data pairs corresponding to the target video. According to the embodiment of the invention, the spatial reasoning data set covering a plurality of dynamic scenes and a plurality of motion attributes can be constructed, so that the visual language model understands and reasones spatial relationship changes among objects in the dynamic scenes.

A LiDAR system includes alight emitter, a light detector, and a controller. The controller is programmed to: activate the light emitter to emit a series of shots into a field of view of the light detector; activate the light detector to detect shots reflected from an object in the field of view; record the detected shots from a first subset of the series of shots; group a second subset of the series of shots not in the first subset; and based on the detected shots from the second subset of the series of shots, identify an object moving in the field of view.

The application relates to the technical field of target tracking detection, and discloses an object motion trajectory extraction method and device, computer equipment and a medium. The method comprises the following steps: acquiring a plurality of trajectory points, the sampling time of the plurality of trajectory points is different, and the plurality of trajectory points correspond to a plurality of target objects; the sampling time of the plurality of trajectory points is utilized to determine a trajectory endpoint corresponding to a first target object from the plurality of trajectory points; the trajectory endpoint is taken as a current query point, and a target screening area is determined with the current query point as the center; based on the size relationship between the sampling time of the trajectory point in the target screening area and the sampling time of the current query point, a next trajectory point belonging to the first target object is determined; the next trajectory point is updated to the current query point, the trajectory point belonging to the first target object is determined, and the motion trajectory of the first target object is obtained, thereby solving the problem that the trajectories of the target objects cannot be accurately and effectively extracted.

The application discloses a closed-chain operation hydraulic double-arm trajectory planning and motion control method considering multiple physical constraint limitations. The method comprises the following steps: establishing a hydraulic double-arm trajectory planner, modeling the physical constraints of the hydraulic double-arm, and generating a new expected trajectory according to an input instruction; establishing a hydraulic double-arm motion controller, receiving the expected trajectory from the trajectory planner, adaptively updating object inertia parameters for estimation of the physical constraint boundary, and performing motion control on the double arm. The application can be used for closed-chain operation of the hydraulic double-arm object, ensures that the object motion does not exceed the multiple physical constraint limitations of the system, so that the object does not deviate from the expected path, reduces the double-arm position error and internal force, and thus avoids the occurrence of unexpected accidents.

1. The name of the design product: bumper and its main body. 2. The use of the design product: both the overall and partial uses are for absorbing or reducing the impact force, vibration or energy generated during the movement of objects. 3. The design points of the design product: in shape. 4. The picture or photo that best indicates the design points: design 1 perspective view 1. 5. Design 1 is designated as the basic design. 6. Other circumstances that need to be explained: designs 2 and 3 are partial designs, and the dashed lines in designs 2 and 3 represent the parts not claimed.

An imaging system providing object motion preprocessing allows a switchable division of signals from different pixels into groups of center and surround pixels of arbitrary shapes and positions allowing tailoring of motion detection to particular objects whose motion should be identified, for example, using smaller center regions for smaller objects.

The application discloses a kind of mobile object intelligent tracking method and device for pan-tilt camera, first by relevant filtering to motion object position and size characteristics, correlate the same object appearing in successive frames, abstract two key points of object in each frame, construct motion trajectory in the panoramic space of camera view field;Second, build reinforcement learning model including multi-layer neural network and customized reward and punishment mechanism, according to the spatio-temporal information of object motion trajectory online decision adjustment pan-tilt rotation amplitude, fusion tracking the same type of object learning experience accelerates model convergence;Finally, use the decision generation and execution of pipeline processing rotation instruction, merge or eliminate the redundant instruction in cache queue, simultaneously adopt frame sampling mode to track low-frequency mobile object, according to object motion information adaptive adjustment sampling interval, to match pan-tilt rotation and object moving speed.The application realizes the efficient low-power consumption mobile object intelligent tracking method, greatly improves the effectiveness and intelligence of monitoring.

Certain aspects of the present disclosure provide techniques and apparatus for accurately predicting the movement and location of objects in a multidimensional space. An example method generally includes receiving an input defining a location of an object in a three-dimensional space at a first time. The input generally includes a plurality of tokens. Using a transformer neural network, a predicted movement for tokens in the plurality of tokens is generated. Generally, the predicted movement includes a translation and a rotation associated with individual tokens in the plurality of tokens. The predicted movement is translated into an object-wise predicted movement. Based on the object-wise predicted movement and the location of the object in the three-dimensional space, a predicted location of the object at a second time subsequent to the first time is output.

The invention relates to an object motion detection method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a model corresponding to a scene where an object is located, and enabling the model to have a collision grid; acquiring a blocking direction corresponding to the collision grid of the model, wherein the blocking direction is used for indicating a motion direction when the object is blocked by the corresponding collision grid in a motion state; voxelization processing is carried out on the collision grid based on the blocking direction corresponding to the collision grid to obtain voxel data corresponding to the scene, each voxel in the voxel data carries a type identifier, and the type identifier is determined based on the blocking direction of the collision grid corresponding to the voxel; acquiring a motion route of the object in the scene; based on the voxel data and the motion route, the motion process of the object in the scene is detected, a detection result is obtained, and the detection result is used for indicating whether the object is blocked or not in the process of moving along the motion route. The application scene of motion detection of the object is expanded.

The invention discloses an image distortion real-time correction method and system for an industrial infrared camera, and relates to the technical field of image processing, and the method comprises the following steps: constructing a distortion expression of a lens, and carrying out the first correction of an image according to the distortion expression of the lens, and constructing a response distortion expression of the mobile monitoring object according to the expression of the lens distortion amount, and carrying out second correction on the image. According to the invention, through constructing the lens inherent distortion expression and the dynamic distortion constructed for the mobile monitoring object, the static lens distortion and the composite distortion introduced by the dynamic scene are respectively eliminated, the geometric accuracy of the image is obviously improved, the deformation of the mobile object caused by the lens distortion can be quantified in response to the distortion expression, and the accuracy of the image is improved. The misjudgment of distortion aggravated by object motion is avoided, the complex distortion is decoupled into two parts of lens inherent and object response by a step-by-step correction strategy, the calculation complexity of single correction is reduced, and the real-time requirement is considered.

The invention relates to the technical field of artificial intelligence, and provides a self-adaptive scene matching safety monitoring video artificial intelligence identification method and system, and the method comprises the steps: detecting object motion state data in a target site, determining scene attribute information, and adjusting a video monitoring operation state of the target site; based on the feature information of the video monitoring operation state and the first picture feature of the monitoring video, the monitoring video is divided and classified to obtain a plurality of monitoring sub-video sets, each monitoring sub-video set is input into a matched neural network model for recognition, and a video picture recognition result is obtained; different monitoring sub-videos are subjected to matched artificial intelligence identification, so that effective monitoring sub-videos in the monitoring sub-videos are screened and reserved; the abnormal event occurrence attribute information of the target site is determined based on the respective video picture recognition results of all the effective monitoring sub-videos, and the monitoring video is subjected to identification processing, so that scene matching processing of the target site is realized, and the security recognition analysis flexibility and accuracy of the monitoring video are improved.

In this pressurizing device, damage to a workpiece is further reduced by improving response performance to feedback from a sensor that measures the value of pressure applied to the workpiece. A pressurizing device (1) applies a pressure having a target value to an object to be pressurized, the pressurizing device (1) comprising: an actuator (10) that pressurizes the object to be pressurized by linear motion converted from rotational motion of a motor (11) and driving a pressurizing portion (14) toward the object to be pressurized; a force sensor (15); and a motor driver (17) that controls the rotational direction and rotational speed of the motor (11). The value of the pressure applied to the object to be pressurized by the pressurizing portion (14) and measured by the force sensor (15) is input to the motor driver (17) as feedback.

This invention relates to the field of imitation learning technology, and particularly to a method and apparatus for imitation learning of a robotic arm based on a single-teaching demonstration. The method includes: acquiring and annotating an assembly video of an arm installing a target object; parsing the annotated assembly video to obtain an arm trajectory sequence; constructing a kinematic model of the target robotic arm and a motion model of the target object based on the arm trajectory sequence; inputting the annotated assembly video into an OpenPose keypoint detection model to extract keypoints of the arm and the target object; training a one-shot imitation learning model using the arm keypoints, target object keypoints, kinematic model, and motion model to generate an adaptive decision model; and integrating the adaptive decision model into the target robotic arm control system to generate an actual assembly strategy to control the target robotic arm for automated assembly. This solves the problems of the time and resources required to construct a robotic arm assembly model, and the lack of flexibility.