863 results about "Tracking model" patented technology

One embodiment of the present invention provides a method and apparatus to track demand partners in a pay-per call performance based advertising system. In one embodiment, the method and apparatus include assigning at least one telephonic reference to an advertiser, the reference is to be included in an advertisement on a media channel of a demand partner. The advertiser is assessed a charge based at least in part on telephonic connections to the advertiser via the telephonic reference. The demand partner tracked with a predefined tracking model based at least in part on activity related to the telephonic reference.

The invention relates to a crashing detecting and warning method based on vehicle network technology. The method comprises the following steps: collecting vehicle information and driving information; establishing a track model with the assistance of historical driving data and an electronic map of a vehicle and according to current driving information of the vehicle for predicting the track of the vehicle in the future n seconds; and judging if there exists danger of crashing according to a security boundary of the vehicle at a moment in the future, and performing warning if the danger of crashing exists. Track information of surrounding vehicles is known through a wireless receiving-transmitting module for receiving and transmitting a vehicle network broadcast protocol dada packet. According to the invention, a driver can be warned in time to avoid the danger, and the driving efficiency with the precondition of ensured security is improved.

One embodiment of the present invention provides a method and apparatus to track demand partners in a pay-per call performance based advertising system. In one embodiment, the method and apparatus include assigning at least one telephonic reference to an advertiser, the reference is to be included in an advertisement on a media channel of a demand partner. The advertiser is assessed a charge based at least in part on telephonic connections to the advertiser via the telephonic reference. The demand partner tracked with a predefined tracking model based at least in part on activity related to the telephonic reference.

The present invention relates to systems and methods for detecting deviations from an orthodontic treatment plan. One method includes receiving a tracking model, performing a matching step between individual teeth in a plan model and the tracking model, comparing the tracking model with the plan model, and detecting one or more positional differences.

The invention provides an MS-KCF (MobileNet-SSD-kernelized correlation filters) based rapid and stable detection method of a human face in an image sequence. To solve the problem about detection of the human face with larger angle change and very severe shielding in the image sequence, a new automatic DTD (detection-tracking-detection) mode integrating rapid and accurate target detection model MSand rapid tracking model KCF is provided, that is, an MS-KCF human face detection model is provided. The method comprises steps as follows: S1, establishing an MS detection network; S2, detecting a target by use of the MS network; S3, updating the tracking model to predict the position of the next frame of human face target; S4, updating an MS detection network after tracking several frames, and performing redetection and positioning on the human face target; S5, comparing and analyzing experiment results. Experiments prove that the MS-KFC model guarantees the detection stability of the humanface with larger angle change and very severe shielding in the image sequence, and besides, the detection speed is greatly increased.

The invention provides a model ship based autonomous navigation control simulation system and method of an under-actuated unmanned ship. The whole system is arranged on a model ship, and the model ship is arranged in a simulated navigation channel; the system comprises an environment sensing sub system, a path programming sub system and a motion control sub system; the environment sensing sub system collects the navigation state of the model ship and simulates environment factor information; according to the navigation state of the model ship and the simulated environment factor information, the path programming sub system carries out navigation programming to obtain the tracking route, the set speed and the set course; and the motion control system combines a course tracking model, a path tracking model, the route deviation, the course deviation and the speed deviation to calculate the rudder angle and propeller rotating speed instruction needed in next navigation of the model ship and control a steering engine and a propeller. According to the invention, simulation experiments are carried out on the model ship, an instance is provided for operation and control experiments of large ships, important guarantee is provided for safe navigation of large ships in inland rivers, and the system reduces the difficult and cost of experiments of the large ships.

The invention relates to a method for predicting the life of a solid tantalum electrolytic capacitor, which comprises the following steps of: 1, collecting current degradation data; 2, determining a degradation track model and a degradation accelerating model; 3, extrapolating the degradation track model to obtain the pseudo-failure life of each sample; 4, performing hypothesis test of pseudo-life distribution and the estimation of unknown parameters; 5, determining the relation of population parameters of the pseudo-life distribution and a stress level; 6 estimating the population parameters of the life distribution of a tantalum capacitor under the normal stress by extrapolating; and 7, determining the average life and reliability curve of the tantalum capacitor. The method has novel concept and simple programs and does not need life tests, the test time can be shorted, the test cost can be saved, and the problem of disharmony between the prediction of the conventional life and engineering is solved, so the method has wide application prospect in the technical field of life prediction.

The invention discloses a target tracking method and a related device. The target tracking method and the related device are used for achieving real-time tracking and processing of a target. The target tracking method comprises the steps of according to a position of the target in a video monitoring image of a previous frame, using a motion predicting model built for the target for predicting positions of the target in a video monitoring image of a current frame, obtaining predicted positions of the target, conducting target detecting in an image area where the predicted positions are distributed according to a detecting model built for the target, obtaining detected positions of the target, conducting target tracking in the image area where the predicted positions are distributed according to a tracking model built for the target, obtaining tracked positions of the target, conducting weight fusion on the detected positions and the tracked positions according to an image position overlapping ratio, obtaining the tracked positions of the target in the video monitoring image of the current frame, adjusting an intelligent video monitoring device according to the tracked positions of the target in the video monitoring image of the current frame, and achieving tracking of the target.

The invention relates to a method of controlling a holder camera to automatically track a target. The method comprises the following steps. A tracking model is automatically chosen. A video frequency processing module automatically chooses an automatic tracking model and a global search model according to the tracking state of the target. The processing process of the automatic tracking model is that the video frequency processing module judges the state of the holder camera and after the tracked target is matched, control orders are sent according to the position and size of the target to regulate the position and the focal distance of the camera. The processing process of the global search model is that the video frequency processing module detects and matches the target through a taking turn manner in a plurality of search areas of a monitored area until the tracked target is found. The invention has reasonable design, realizes the automatic tracking function of the camera on the target, ensures that the position and size of the target are suitable for monitoring, has the advantages of high tracking precision, strong real-time property, low cost, flexible expansion and the like, facilitates the use of monitoring staff and improves the work efficiency of the monitoring staff.

The invention discloses a deep-learning-based multi-target pedestrian detection and tracking method. The method comprises the following steps: step one, carrying out multi-target pedestrian detectionand joint point extraction on an inputted video and storing obtained position information and joint point information as inputs of a next stage; step two, selecting a key frame at an interval with thecertain number of frames and carrying out apparent characteristic extraction in a pedestrian in the key frame; to be specific, according to the obtained position information and joint point information, extracting upper body part attitude characteristics and color histogram characteristics respectively for pedestrian association between key frames; and step three, carrying out continuous trackingon the pedestrians in the key frames with a threshold slow starting strategy, a block matching rate model detection algorithm, a historical state keeping voting algorithm and a shielding detection method for tracking effect improvement, returning to the step one after tracking ending, and detecting the key frames again to ensure stability of the method.

The invention discloses an intelligent man-machine conversation system in a closed domain. The system comprises a first modeling module, a second modeling module and a third modeling module, wherein the first modeling module is used for constructing a multi-feature fusion deep intention recognition model on the basis of a bidirectional long / short-term memory network and a convolutional neural network; the second modeling module is used for constructing an MC-BLST-MSCNN-based conversation state tracking model by adoption of a man-machine conversation system current state input and context statement combined modeling manner; and the third modeling module is used for constructing a Bi-LSTM matching model on the basis of an extraterritorial recover mechanism of a displacement attention mechanism, and inputting recognized user intentions and user slot values into a displacement network to carry out weight distribution the attention mechanism, so as to realize conversation state encoding andconversation control matching. The system has higher intention recognition accuracy, conversation state tracking correctness and conversation control stability, so as to enhance the cognition intelligence capability of man-machine systems.

The invention discloses a multi-target pedestrian detecting and tracking method in monitoring video, comprising the steps that a target detection network based on deep learning is adopted for detecting a first frame of pedestrian image, and an initial rectangular area having one or a plurality of corresponding pedestrian targets can be obtained; based on the initial target area information, the Histogram of oriented gradients feature of a target can be extracted, and kernel function autocorrelation calculating of Fourier expansion domain can be conducted, and the tracking model is initializedbased on the calculating result; based on the target area information of the tracking model, a multi-dimensional construction of a pyramid will be carried out from the second frame of pedestrian image, and the extracting of the Histogram of oriented gradients feature matrix and the kernel function autocorrelation calculating of Fourier expansion domain can be conducted on each scale of the pedestrian rectangular area; the returned check condition is determined, and the identity re-verification and the updating of the tracking model can be conducted on the pedestrian target having returned check. The invention is advantageous in that the problem of drifting models can be resolved; a more accurate pedestrian moving track can be obtained; real-time performance is good.

The invention discloses a lane change track optimization and visualization achievement method based on vehicle models and vehicle speeds. The method includes the following steps of (1) calling whole vehicle models of the four vehicle models in ADAMS software according to the following steps: (1-1) obtaining a whole vehicle structure parameter, and (1-2) obtaining vehicle dynamic characteristics; (2) analyzing tracking errors of an existing lane change track model and extracting lane change track models suitable for different vehicle models and different vehicle speeds; (3) optimizing lane change tracks on the basis of safety and comfort according to the following steps: (3-1) building a driving control DCD file on the basis of a lane change track constraint condition, and (3-2) creating a driver control model and a double-lane three-dimensional road model, creating a driver, road and vehicle simulation system, and obtaining an optimal lane change track after simulation. According to the method, lane change behaviors of a vehicle can be observed dynamically in real time, an intelligent vehicle lane change system is optimized, comfort of drivers is improved, and road driving safety is guaranteed.

The invention relates to a drilling engineering technology in petroleum engineering, in particular to a method for monitoring the track of an actually drilled wellbore. The method comprises the following steps: acquisition of inclination survey data and naturalization treatment; determination of characteristic parameters of a wellbore track model; determination of a hole inclination equation and an azimuth equation; calculation of coordinate increment; calculation of surveying point coordinates; and so on. As for large-displacement wells containing catenary or parabola tracks and various directional wells and horizontal wells, the method can accurately monitor the coordinates of various surveying points, so as to make the monitoring result accord with the practice of the drilling engineering, thereby effectively improving the construction precision and the reliability of catenary drilling and parabola drilling.

The invention disclosed here relates to a method and system for real-time target tracking based on deep learning. The method for real-time target tracking according to an embodiment is performed by a computing device including a processor, and includes pre-training a target tracking model for detecting a tracking target from an image using pre-inputted training data, receiving an image with a plurality of frames, and detecting the tracking target for each of the plurality of frames by applying the target tracking model to the image. According to an embodiment, there is a remarkable reduction in the time required to detect the target from the image, thereby allowing real-time visual tracking, and improvement of the hierarchical structure and introduction of a new loss function make it possible to achieve more precise localization and distinguish different targets of similar shapes.

The invention relates to a table tennis target tracking and trajectory predicting method and device, a storage medium, and computer equipment. The method includes the following steps: acquiring images of a tracking target taken by two cameras at the same time, and extracting a candidate region corresponding to the tracking target from each image; inputting the candidate regions to a preset tracking model and processing the candidate regions to get a bounding box corresponding to the tracking target; acquiring the 2D coordinates of the center of the bounding box corresponding to the tracking target taken by the two cameras at the same time, and calculating the 3D coordinates of the center of the bounding box according to the projection matrix of the cameras; and acquiring the 3D coordinates of the bounding box at continuous time to constitute a continuous coordinate sequence, inputting the continuous coordinate sequence to a recurrent neural network LSTM and generating a subsequent coordinate sequence through calculation, and getting the trajectory of the tracking target according to the coordinate sequence. The location of the target is tracked using the preset tracking model. Based on the location and LSTM, the advantages of timing features can be analyzed, and the trajectory of the tracking target can be accurately predicted.

The invention relates to a front multi-vehicle tracking method integrating a millimeter radar and depth learning vision. The millimeter radar is used for obtaining front data information, according tothe echo reflection strength and width information, invalid information is removed, and only front vehicle information remains. According to the method integrating the millimeter radar and a camera,by filtering radar information and generating a motion trajectory through an online tracking model, and trajectory correlation is performed. The front vehicles related to the trajectory are recorded and numbered. According to the generated trajectory and the numbered front vehicles, the steps just need to be repeatedly executed on data of the next period, consistency checking is performed, and thedata is added into the numbered trajectory. For newly appearing vehicles, trajectory generation, trajectory collection and numbering are performed according to the beginning steps. By combining advantages of the millimeter radar and the vision depth learning, and the target tracking accuracy and robustness for front multiple vehicles can be effectively improved.

The invention provides a self-driving car lane changing control method based on monocular vision and belongs to the technical field of car electronic control. A camera is installed on a roof of an self-driving car and used for collecting a lane line image; an image processing module is used for processing and recognizing the lane line image to obtain a fit lane line; an upper computer module is used for calculating turn angle increment of a steering wheel, and outputting a motor control signal to an executing unit. According to the method, lane changing reference angles are defined, the optimal lane changing track model based on the lane changing reference angles is provided, the turn angle increment of the steering wheel is obtained by means of a preview control theory to control the steering wheel of the car, and the left lane changing reference angle or the right lane changing reference angle is selected according to needs. Lane changing closed-loop control can be achieved only by using the camera for path recognizing, the lane changing process is stable and reliable, cost is reduced, and the complex degree of the system is reduced.

The invention discloses an image guiding and tracking method based on prediction, which comprises the following steps: 1, inputting image sequences and a breathing status feature set of a tumor anatomical structure; 2, establishing a status association database; 3, acquiring real-time images and breathing status features; 4, predicting the delayed breathing status features, and determining the image sequences in a corresponding tumor shape varying section; 5, registering the real-time images with the image sequences in the tumor shape varying section one by one in an interested area; 6, selecting an optimal tumor shape image from registration results; and 7, outputting the image of the tumor anatomical structure. The image guiding and tracking method based on the prediction establishes the correlation between breath and tumor movement by the quantification and analysis on the breathing statuses, predicts the breath and tracks the movement according to the correlation, and improves the efficiency and the accuracy of tracking a dynamic tumor in real time.

Computerized methods and systems for locating barcodes applied to objects are provided. A method can receive a first image of a first barcode fragment applied to a first object captured at a first time and identify a first position of the first barcode fragment. The method can also receive a second image of a second barcode fragment captured at a second time and identify a second position of the second barcode fragment. The method can also predict a range of possible positions of the first barcode fragment at the second time based on a tracking model that tracks the first barcode fragment based on the first position, and determine that the first barcode fragment and the second barcode fragment correspond to the same barcode, if the second position is within the range of possible positions of the first barcode fragment at the second time.

The invention discloses a target tracking method and device. The target tracking method comprises the following steps of: establishing a network tracking model according to the current frame of image; determining the current frame of image comprising a tracked target; obtaining a tracked target area in the current frame of image; obtaining the next frame of image of the current frame of image; obtaining a plurality of target candidate areas in the next frame of image; calculating the similarity between the tracked target area and each target candidate area; determining a target tracking area in the plurality of target candidate areas according to the similarities; determining the tracked target in the target tracking area; and obtaining the current state of the tracked target, and updating the network tracking model according to the current state. According to the method and device disclosed by the invention, the network tracking model is established, so that the individual differences between same objects of the tracked targets can be recognized, and the appearance changes of the target objects under various conditions can be effectively expressed when sample objects are updated.

The invention discloses a deep sea multibeam sound ray accurate tracking method. The method mainly comprises the steps of (1) establishing a spatial and temporal variation ocean temperature and salinity field model on the basis of the comprehensive analysis of influence factors of sea surface temperature and salinity and with the integration of marine satellites and Argo buoy multiple-source marine physical hydrological observation data, (2) calculating each beam initial incident angle through analyzing the influence of a ship instantaneous attitude on a beam initial incident angle and considering a ship attitude, (3) calculating a sound speed and an inversion sound speed based on an empirical orthogonal function through the spatial and temporal temperature and salinity field at a sound speed profile no-measured area, obtaining the average value of sound speeds obtained by two methods at a corresponding point as the sound speed of the point, and thus calculating a three-dimensional sound speed profile model, (4) constructing an efficient constant gradient sound ray accurate tracking model, and (5) providing a sound ray tracking precision evaluation method. According to the method, the influence of the ship attitude on the beam initial incident angle is considered, the precision of beam footprint coordinates can be greatly improved.

The invention relates to a hand-writing and gesture recognize method used in mobile device, wherein via comparing the nearby two image frames obtained by camera device, calculating out the relative motion vector of two image frames; and connecting the relative motion vectors of all nearby image frames in one whole input process, end-to-end, to obtain the accurate plane motion track of cam; using said motion track to match the built stroke or gesture track model; when the match is successful, the match result is output to the display screen of mobile device. The invention also discloses a relative device, without adding special hardware on the mobile device, to save the cost.

The invention provides a traffic accident detection method and device. The method comprises the following steps: obtaining a traffic monitoring video image which is acquired at a pre-set position by a camera; establishing a vehicle track model and / or a vehicle surrounding environment model according to the traffic monitoring video image; and detecting a traffic accident according to the vehicle track model and a pre-established road surface structure model, and / or detecting a traffic accident according to the vehicle surrounding environment model and the pre-established road surface structure model, wherein the road surface structure model is a model which is established according to a monitoring video image without the traffic accident on the road surface. The automatic traffic accident detection method adopting a traffic accident detection system saves a lot of manpower and material resources, and the working intensity of monitoring personnel is alleviated. The risks of incorrect report and missed report of the traffic accident, caused by the fact that the effort is not enough, do not occur, so that the accuracy of detecting the traffic accident is improved by the detection method.

The invention provides a multi-target tracking detection method and device based on Darkflow-DeepSort, and a storage medium, and relates to the technical field of intelligent decision making. The multi-target tracking detection method comprises the following steps: S110, carrying out the training through employing a YOLOv3 algorithm, and obtaining a Darkflow-based target detection model; S120, inputting the detection image into a trained Darkflow-based target detection model to obtain apparent characteristics of a plurality of targets, wherein the detection image is obtained by decoding a monitoring video; S130, inputting the apparent characteristics of the plurality of targets into a trained DeepSort-based target tracking model, wherein the target tracking model is obtained by training adata set MOT16Challenge of multi-target detection; and S140, performing frame-by-frame data association processing on the monitoring video by using a Kalman filter of a target tracking model to realize multi-target tracking in the monitoring video. By utilizing the multi-target tracking detection method, the multi-target tracking detection speed can be increased, and multi-target tracking is completed under the condition that the detection accuracy is not lost.

The invention relates to a time-varying rotational inertia considered adaptive fault tolerance control method of spacecraft attitude. The method comprises that a dynamic spacecraft attitude tracking model which does not take faults of a performer into consideration but includes external disturbance and a time-varying rotational inertia is established, and a time-varying rotational inertia model isestablished; and on the basis of the dynamic spacecraft attitude tracking model, a nominal controller and corresponding adaptive laws are designed, and the system stability and tracking error convergence without faults are ensured; and the performance fault is added to the dynamic spacecraft attitude tracking model, an auxiliary controller is designed on the basis of the nominal controller, the problem that a fault occurs in the performer, the time-varying rotational inertia exists and the spacecraft is influenced by an external disturbance moment during on-orbit work of the spacecraft is solved, and the fault tolerant capability and robustness of the system are ensured.

The invention belongs to the field of computer graphics and image processing, and specifically discloses a real-time multi-class and multi-target tracking method of a video. The tracking method comprises the following steps: s1: carrying out video frame preprocessing, such as superpixel segmentation; s2: on the basis of a superpixel block, designing a target detector, carrying out off-line training, and fully utilizing movement features so as to detect all moving targets in the video; s3: utilizing the trained detector to carry out target detection on a given video; s4: designing a target tracking model, and tracking the targets in the video; and s5: carrying out trajectory visualization. The tracking method has the following beneficial effects: 1: the complexity and the time consumption of an algorithm are greatly lowered since the video is subjected to the target detection on the basis of the superpixel block; and 2: all multiple classes of moving targets in the video can be detected, so that the tracking of all moving objects in the video can be realized under a situation of one camera, and therefore, hardware cost is greatly lowered.

The invention discloses a wheel type moving robot track tracking method based on a fast terminal sliding mode. The wheel type moving robot track tracking method comprises steps of (1) establishing a kinematic model of the wheel type moving robot and an expectation track model and establishing an error model according to the kinematic model and the expectation track model, (2) introducing an appropriate sliding mode surfaces s1, s2, designing a virtual feedback amount according to the error model <~ > theta e<->, (3) obtaining a linear speed deviation signal <~ > Upsilon and a angular velocity deviation signal <~ >Omega, (4) constructing the wheel type moving robot kinematic model and substituting the linear speed deviation signal <~ > Upsilon and the angular velocity deviation signal <~ >Omega into the kinematic model, and designing a left-and-right-wheel torque controller Tau, an unknown parameter estimator <^>Phi and an external interference disturbance estimator. The wheel type moving robot track tracking method based on a fast terminal sliding mode can track the expectation track in the limited time under the disturbance complicated working condition of the unknown parameters and the external disturbance, has a good tracking effect and has a strong robustness for the unknown parameters and the external disturbance.

Platform independent distributed software that constructs a security management infrastructure for different locations is described. The software includes a control manager module regulating access to critical business assets. The control manager interfaces with and bridges various type of biometric software and hardware systems. The software further includes a tracking model consisting of a custom report generation tool for monitoring people, business critical assets, and events in any particular location. In addition to tracking and access control, the software provides an event notification service for registering, monitoring, and storing primary / composite events into a modeled event history database. The software is integrated into three components. One component is the company headquarters integrated with platform independent legacy software comprising multi-tiered ERP and business packages for cross domain security management and monitoring. Another component is the managed location where the access control regions are held and the main tracking software is installed. The final component is the client software integrated with the biometric devices, regulating authentication for people, assets, and / or events.