Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

44results about How to "Tracking results are accurate" patented technology

Improved target tracking method

The invention discloses an improved target tracking method, and solves the technical problems in the prior art that a Mean-shift algorithm tracking window is fixed and Kalman prediction is sensitive in target adhesion. The specific method includes: using Kalman filtering to predict a target position, reducing a research range, then matching a target through binary image features, correcting kernel function bandwidth of a Mean-shift algorithm, updating the Kalman model, searching a matched target in the vicinity of the predicted position, and completing target tracking by Mean-shift vector iteration. The improved target tracking method provided by the invention combines the advantages of Kalman prediction and the Mean-shift algorithm, is small in calculated amount and has good real-time performance, and effectively guarantees tracking accuracy under a circumstance of target adhesion.
Owner:HOHAI UNIV

Real-time tracking method of nonspecific target based on partitioning

The invention relates to a real-time tracking method of a nonspecific target based on partitioning, comprising three steps of classifier updating, target detection and weight updating. In the method, the target region is divided into multiple blocks; a classifier is used for maintaining each block, and updating is conducted frame by frame; the detection result of each classifier is comprehensively considered to determine the position of the target in the new video frame. In the method, an automatic weight updating mechanism is designed, so as to enable the blocks which are relatively stable to have greater decision-making power over the judgment of results, thus reducing the influence of various interferences; and the tracking performance is better than multiple international published algorithms recently. In the method, changes of appearances of objects caused by various interferences can be captured and accurate tracking can be conducted; the method has universality on target objects in various shapes and types; the calculation has low complexity and can be processed at real time. The invention has wide application prospect in various occasions needing tracking techniques, such as video monitoring, automatic driving, man-machine interaction, intelligent traffic, robot, airborne early warning and the like.
Owner:UNIV OF SCI & TECH OF CHINA

Covariance matching-based active contour tracking method

The invention relates to a covariance matching-based active contour tracking method and belongs to the technical field of visual tracking. In the covariance matching-based active contour tracking method, an image area energy term is modeled by using non-Euclidean geometry. The method comprises the following steps of: manually initializing a curve surrounding an objective and establishing a covariance matrix as a template of an objective contour for an area surrounded by the curve in a first frame; after the contour of the objective is obtained, recording a level set function value of the template to make preparation for a prior shape and calculating a symbolized distance function of the template; from the image of the next frame, deducing a gradient descent flow from a result of the previous frame according to the established energy functional and updating the level set function; and checking whether iteration stops or not. In the method, the tracking result is more accurate; meanwhile, the covariance matrix is used as an area descriptor and all kinds of information in an image sequence and the correlation between all kinds of information are considered comprehensively, and the method does not depend on foreground and background information distribution, so that the tracking method has universality.
Owner:BEIJING INSTITUTE OF TECHNOLOGYGY

Multi-human-body tracking method based on human body part model

ActiveCN108416258AExact match trackingOvercome scaleCharacter and pattern recognitionVideo monitoringPattern recognition
The invention provides a multi-human-body tracking method based on a human body part model and relates to the video monitoring technology field. The method comprises the following steps of acquiring image data, extracting the human body part model of each person and calculating a color characteristic, acquiring the part characteristic set of each person, and gathering and acquiring a multi-human-body part characteristic set list; then, calculating the similarity of the part characteristic set of each person acquired from a current frame and each person in the multi-human-body part characteristic set list acquired from a previous frame and acquiring a matching matrix; and according to the matching matrix, calculating the similar confidence of each person acquired from the current frame to each person acquired from the previous frame, according to the similarity and the similar confidence, matching a target acquired from the current frame with the target of the previous frame, and acquiring a tracking result. By using the method, the detection target of the current frame can be accurately matched to the position of the previous frame and a new target is discovered, and the problem oftracking failure caused by a characteristic difference generated because of human body deformation during a multi-human-body tracking process is effectively solved.
Owner:HUAQIAO UNIVERSITY

Simple and universal fish school video image trajectory tracking method

The invention discloses a simple and universal fish school video image trajectory tracking method, and relates to the field of video image trajectory tracking. Two characteristics that when most fishbodies swim in a two-dimensional plane, the area of an overlook angle is almost kept unchanged, and sudden change cannot occur in the swimming direction of the fish bodies in a short time are utilized; a nearest neighbor method is combined for fish swarm tracking; when a certain fish is tracked, firstly, fish bodies with large area difference are eliminated according to historical information of the fish, then, a connecting line of centers of mass of two continuous frames of fish bodies is adopted as the swimming direction of the fish bodies, and a target with too large direction angle difference is eliminated; finally, inter-frame correlation between front and back frames is performed on the coordinate position of the fish body by adopting a nearest neighbor method to complete tracking. An image corrosion method is adopted, and therefore the problem of mutual shielding of fish bodies is solved. Tracking results are more accurate. The method is easy to popularize and has a wide application prospect. The method can be used for researching fish behavioristics in various laboratories so as to obtain accurate action tracks of biological populations, construct a biological activity model or assist in verification hypothesis.
Owner:XIAMEN UNIV

Multi-sensor target fusion and tracking method and system

The invention discloses a multi-sensor target fusion and tracking method and system. The fusion method comprises the steps of firstly judging whether data transmission time of two sensors is valid ornot, then performing time and space registration on the two pieces of data which are judged to be valid, and removing objects which should not participate in target matching fusion; calculating the fusion coefficient of the data transmitted by the two sensors, and calculating the optimal matching according to the obtained fusion coefficient. According to the tracking method, on the basis of the fusion method, the tracked target state is updated according to the optimal matching result, the tracking result with the large error is screened and processed according to the tracking historical result and prior information, and finally high-precision detection and tracking of the dynamic obstacle are achieved.
Owner:MOMENTA SUZHOU TECH CO LTD

Multi-target tracking method based on multi-agent deep enhancement learning

The invention discloses a multi-target tracking method based on multi-agent deep enhancement learning. The method comprises: detecting multiple targets through a target detector, regarding the detected multiple targets as multiple agents, and then using a deep enhancement learning method to obtain a combined action set of the multiple targets, thereby completing multi-target tracking. The method applies a multi-agent deep enhancement learning technology to the multi-target tracking method for the first time, and can overcome technical disadvantages that artificial design features are not comprehensive and not accurate enough, and can improve calculation speed, realize real-time tracking. The method has high multi-target tracking accuracy rate and precision, and the number of misinformationand missing reports is low. The method is less affected by interference factors in a multi-target tracking scene, and tracking results are more accurate.
Owner:NANJING QIANHE INTERNET OF THINGS TECH CO LTD

Target object tracking method and device

The invention provides a target object tracking method and device. The method comprises the steps: extracting marked parts of a target object detected in a current frame parallax image; carrying out the clustering of parallax points in a next frame parallax image, so as to form a plurality of cluster; calculating the similarity of each cluster in the next frame parallax image with each marked part, so as to determined candidate clusters similar to the marked parts; carrying out the grouping of all candidate clusters according to the position relation among all marked parts, so as to enable the candidate clusters belonging to the same object to be divided into the same group; determining the similarity of each group with the target object based on the similarity of all candidate clusters in each group; selecting the group with the highest similarity, and enabling the position where the group is located to serve as the position where the target object is located in the next frame parallax image. Through the technology of target object tracking, the method also can obtain an accurate tracking result under the conditions that the posture of the target object is frequency changed and the target object is partly shaded.
Owner:RICOH KK

Two-dimension track tracking controller of underwater robot

The invention belongs to the object tracking control field, and discloses a two-dimension track tracking controller of the underwater robot. The two-dimension track tracking controller is realized through steps of utilizing a least square method to perform modeling on a movement track under a motion coordinate system, designing a self-adapting wide domain prediction control two-dimension track accuracy tracking controller of the underwater self-adapting robot on the basis of that, and realizing the two-dimension track tracking control of the underwater robot. The invention is applicable to the object tracking and identification.
Owner:TIANJIN POLYTECHNIC UNIV

Channel tracking method and device

Embodiments of the invention provide a channel tracking method and device, and relates to the field of communication channel processing. The method and system can perform channel tracking on communication systems provided with uniform rectangular arrays. The method includes obtaining an initial emission horizontal angle, an initial emission pitch angle, an initial receiving horizontal angle, an initial receiving pitch angle and preset gain information; calculating the desired value of a channel matrix of a current time block on the basis of a preset channel matrix formula and channel time-varying formula according to the obtained information; calculating the estimated values of a horizontal and vertical array of the current time block according to the obtained information, a preset angle time-varying formula and the preset channel matrix formula; obtaining the optimal value of the horizontal and vertical array of the current time block on the basis of a matrix decomposition algorithm according to the estimated values of the horizontal and vertical array of the current time block and the desired value of the channel matrix of the current time block; and calculating the optimal valueof the channel matrix of the current time block according to the optimal value of the horizontal and vertical array of the current time block.
Owner:CHINA UNITED NETWORK COMM GRP CO LTD

Alpha beta filtering based ship target tracking processing method

An embodiment of the invention provides an alpha beta filtering based ship target tracking processing method. The alpha beta filtering based ship target tracking processing method comprises the following steps of enabling a processor to establish a predication port door according to a plotting target position which is corresponding to an original plot, the course and the speed; identifying track targets inside the predication port door; judging whether the number of the track targets is larger than 0 or not, if so, determining that the original plot is a tracking target exists and judging whether the the number of the track targets is larger than 1, if so, selecting an optimal track target to relate to the original plot if the number of the track targets is larger than 0, adopting an alpha beta filtering to relate to the original plot if the number of the track targets is equal to 1, if the number of the track targets is not larger than 0, determining that the original plot is that a tracking target does not exist, performing straight-line extrapolation on the track targets and relating to the original plot. The alpha beta filtering based ship target tracking processing method achieves comprehensive analysis and processing of various conditions and enables a track target tracking result to be accurate.
Owner:大连海大船舶导航国家工程研究中心有限责任公司

K sparseness based rapid robust target tracking method

The invention relates to a K sparseness based rapid robust target tracking method. The K sparseness based rapid robust target tracking method comprises the following steps: solving a sparse matrix byutilizing a positive and negative template dictionary; solving an optimization problem of the positive and negative template dictionary l1 by utilizing the sparse matrix, namely K sparseness, whereinunder the condition that the sparseness is relatively small and the signal-to-noise ratio is the same, the speed of solving a sparse coefficient is 10 times or more than 10 times of an LASSO (Least Absolute Shrinkage and Selection Operator) algorithm; dividing a current-frame image into a plurality of image samples and calculating a reconstruction error by utilizing a normalized positive templatematrix and a normalized negative template matrix; judging error parameters through reconstruction and comparing samples obtained by histograms, so as to realize target tracking in the current-frame image. Therefore, the target tracking accuracy is also improved, a more accurate tracking result can be obtained and the instantaneity of tracking is improved.
Owner:SHAOGUAN COLLEGE

Moving target tracking method based on cascade detector

The invention discloses a moving target tracking method based on a cascade detector. The moving target tracking method comprises the following steps of (1) initializing a related filtering tracker; (2) initializing a cascade detector; (3) enabling the correlation filtering tracker to obtain a first region of interest; (4) detecting a cascade classifier target; (5) synthesizing the first region ofinterest and the detection region; (6) updating the tracking model and the cascade detector, wherein the first region of interest is obtained through the correlation filtering tracker, the cascade detector comprises a variance classifier, a multi-channel random fern classifier and a related consistent classifier, the detection area is acquired through the three classifiers, the first region of interest and the detection area are subjected to weighted correction to acquire the second region of interest, so that the oscillation of a tracking result can be effectively prevented. Meanwhile, the accuracy of the first region of interest acquired by the tracking model can be ensured by adopting the cascade detector.
Owner:HUAZHONG UNIV OF SCI & TECH

Object tracking method and device

The invention discloses an object tracking method and device, and relates to the field of image processing. According to the specific implementation scheme, multiple frames of first images shot by a first camera device and a first shooting moment of each frame of first image are acquired, wherein the first images comprise a first object; multiple frames of second images shot by a second camera device and a second shooting moment of each frame of second image are obtained, wherein the second image comprises a second object; a relative relationship between the first camera device and the secondcamera device is obtained; and whether the first object and the second object are the same object or not is judged according to the multiple frames of first images, the first shooting moment of each frame of first image, the multiple frames of second images, the second shooting moment of each frame of second image and the relative relationship. According to the object tracking method and device provided by the embodiment of the invention, the problem that the robustness of the existing scheme is poor due to the fact that the existing scheme is greatly influenced by the image shooting angle canbe solved, and the tracking result is more accurate.
Owner:BEIJING BAIDU NETCOM SCI & TECH CO LTD

Echo block tracking method and device of radar echo image and storage medium

The invention discloses an echo block tracking method and device of a radar echo image and a storage medium, and relates to the technical field of weather forecast, and the echo block tracking method of the radar echo image comprises the following steps: obtaining a plurality of radar echo images to form a radar echo image sequence, wherein a plurality of radar echo images in the radar echo image sequence come from the same radar and are arranged according to a time sequence; performing hierarchical division on each radar echo image in the radar echo image sequence to obtain a plurality of binarized hierarchical matrixes; determining all echo blocks in the hierarchical matrix according to a communication rule; calculating the similarity between echo blocks in the hierarchy matrixes of the same hierarchy of two adjacent radar echo images; and judging the two echo blocks with the similarity greater than a threshold as the same cloud cluster. By comparing the similarity of the echo blocks hierarchically, the echo blocks in the radar echo image can be accurately tracked.
Owner:GUANGDONG UNIV OF TECH +1

Tracking type lifting sonar and sonar control method

The invention relates to a tracking type lifting sonar and a sonar control method. The sonar comprises a wet end and an array plane arranged around the periphery of the wet end. The number of the array surfaces is a plurality and is an even number; each array surface comprises at least one transducer array and a hydrophone array, each array surface is connected with the wet end through a connecting rod, the two ends of the connecting rod are respectively provided with a swing device, and the swing devices can drive the connecting rod or the array surfaces to rotate. According to the sonar, thetransducer array and the hydrophone array are both arranged on the array surfaces, each array surface can rotate, and when the position (or angle) of the tracking target changes, the array surfaces can also rotate to a more appropriate position, so that relevant data information of the tracking target can be collected conveniently, and the tracking result is more accurate.
Owner:HAINAN UNIVERSITY

Single-target tracking method, device and system

The invention provides a single-target tracking method, device and system, and relates to the technical field of computer vision, and the method is applied to equipment configured with a tracking network. The tracking network comprises a backbone network, an STN and a similarity measurement layer. The method comprises the following steps: acquiring a frame image to be tracked and a template imagecontaining a target object; extracting the template feature map of the template image and the first feature map of the frame image through a backbone network; performing feature migration on the firstfeature map through the STN to obtain a second feature map; and calculating a first similarity score map between the template feature map and the second feature map through a similarity measurement layer, and determining the regression box of the target object in the frame image based on the first similarity score map. According to the invention, the accuracy of the regression box of the target object and the accuracy of target tracking can be effectively improved.
Owner:MEGVII BEIJINGTECH CO LTD

Real-time real-person virtual hair try-on method based on 3D face tracking

The invention relates to a real-time real-person virtual hair try-on method based on 3D face tracking. The method comprises the following steps: firstly, carrying out real-time 3D face tracking for virtual hair try-on; then three-dimensional hair model wearing based on orientation consistency is carried out; and finally, re-coloring the three-dimensional hair model keeping the color difference ofadjacent pixels. According to the method, through the lightweight model and the 3D face feature points, the problems of calculation time consumption and unstable tracking result generation caused by association of the 2D face feature points and the vertexes of the three-dimensional face model are avoided, so that a rapid and accurate tracking result is realized. And the registration of the three-dimensional hair model enables the try-on hair to be more accurately fit with the real face, so that the authenticity of the virtual try-on hair is improved. In addition, a method for changing the texture color of the three-dimensional hair model is added, so that the hair try-on experience of a user and the functionality of a hair try-on system are enhanced.
Owner:ZHEJIANG GONGSHANG UNIVERSITY

Multilayer restriction multi-target tracking algorithm

The present invention provides a multi-target tracking algorithm based on multilayer restriction. The multi-target tracking algorithm comprises: employing a simple strategy to rapidly obtain tracking segments, employing vision information to perform correction and segmentation of all the tracking segments, converting a result to a graph model about the tracking segments for solution, employing high-level semantic information of the result obtained through solution to perform further correction, and obtaining a final tracking target track. The multilayer restriction multi-target tracking algorithm fully employs the vision feature of a tracking target to perform restriction layer by layer from shallow to deep so as to solve the problem that an accurate result is difficult to obtain because a current method is difficult to fully employ the vision feature and allow the tracking result to be accurate.
Owner:NORTHWESTERN POLYTECHNICAL UNIV

A channel tracking method and device

Embodiments of the invention provide a channel tracking method and device, and relates to the field of communication channel processing. The method and system can perform channel tracking on communication systems provided with uniform rectangular arrays. The method includes obtaining an initial emission horizontal angle, an initial emission pitch angle, an initial receiving horizontal angle, an initial receiving pitch angle and preset gain information; calculating the desired value of a channel matrix of a current time block on the basis of a preset channel matrix formula and channel time-varying formula according to the obtained information; calculating the estimated values of a horizontal and vertical array of the current time block according to the obtained information, a preset angle time-varying formula and the preset channel matrix formula; obtaining the optimal value of the horizontal and vertical array of the current time block on the basis of a matrix decomposition algorithm according to the estimated values of the horizontal and vertical array of the current time block and the desired value of the channel matrix of the current time block; and calculating the optimal valueof the channel matrix of the current time block according to the optimal value of the horizontal and vertical array of the current time block.
Owner:CHINA UNITED NETWORK COMM GRP CO LTD

Blended ore tracking judgment method and system

The invention discloses a blended ore tracking judgment method and system. The method comprises the steps that: whether a material bin starts feeding or not is judged; if yes, the current time is recorded, and the material bin starting feeding is set as a feeding bin; the feeding duration of the feeding bin is calculated; and the time obtained by subtracting the feeding duration from the current time is taken as the actual feeding time of the uniformly mixed ore. The feeding starting time of the material bin is detected and judged, the detected feeding starting time of the material bin subtracts the time after the feeding duration time to serve as the actual feeding time of the blended ore, the tracking result of the new blended ore component feeding time is basically not delayed, the tracking result is more accurate, and material batching is more accurate.
Owner:SHOUGANG JINGTANG IRON & STEEL CO LTD

Object Tracking Method and System Based on Triple Convolutional Network and Perceptual Interference Learning

The invention discloses a target tracking method and system based on triple convolutional network and perceptual interference learning, which belongs to the field of target tracking research in image processing and machine vision. The method includes: inputting the video to be tracked into the triple convolutional network to obtain target tracking Results; the training of the triple convolutional network includes: constructing the triple convolutional network, obtaining the positive sample pair and the negative sample pair from the data set to obtain the training set; using the training set to train the triple convolutional network, two images of each sample pair in the training set Input the template branch and the detection branch respectively, or input the first frame branch and the detection branch respectively; the template branch and the first frame branch respectively extract the appearance model feature map, and compare the two appearance model feature maps with the feature map of the detection branch respectively. Cross-correlate to get two response maps; respectively calculate the loss of the two response maps for backpropagation, and thus obtain a trained triple convolutional network. The method of the invention has higher target tracking accuracy.
Owner:HUAZHONG UNIV OF SCI & TECH

Object Tracking Method Based on Self-Attention Transformation Network

The invention discloses a target tracking method based on a self-attention transformation network, which is realized by using a search image encoding module, a target encoding module, a decoding module, and a supervision module; the search image encoding module is composed of a multi-head self-attention network and a feedforward The network is implemented in series, and the search image block encoding is calculated through the multi-head self-attention network and the feed-forward network; the target encoding module calculates the target feature encoding through the masked multi-head self-attention network; the decoding module is based on the target feature encoding, through the multi-head attention network in the search In the image feature encoding, the query matching is performed to calculate the position coordinates of the target prediction frame; the supervision module calculates the error between the two according to the position information of the target prediction frame and the real target position information, and the neural network parameters obtained when the error is minimized are used for target tracking. The invention has a stable tracking effect, is easier to capture deformed targets in search images, and generates accurate tracking results.
Owner:中国人民解放军32802部队

A multi-cue visual tracking method based on adaptive sub-block selection

The invention belongs to the visual tracking field and relates to an adaptive sub-block screening-based multi-clue visual tracking method. The method comprises the following steps that: (1) saliency detection is performed on a target region, uniform block division is used in combination, so that candidate sub-blocks can be obtained; (2) multi-scale sampling is performed on the candidate sub-blocks, sub-blocks with large frequency-domain response and the corresponding scales of the sub-blocks are determined, and a candidate sub-block set is updated; (3) motion estimation is performed on the sub-blocks in the candidate sub-block set, and the current location of a tracking target is determined through the multi-clue fusion of the sub-blocks; and (4) a Gaussian kernel corresponding to the location of each sub-block is updated through the current location of the target, and sub-blocks which do not satisfy requirements are re-initialized. According to the adaptive sub-block screening-based multi-clue visual tracking method of the invention adopted, the interference of background can be removed, the visual constraints of middle-level features and the priori constraints of high-level languages are fully utilized, so that the locating of the target is more accurate. The adaptive sub-block screening-based multi-clue visual tracking method has the advantages of simple steps and small computation amount, and is suitable for performing visual target tracking under a blocking condition.
Owner:HUAZHONG UNIV OF SCI & TECH

Covariance matching-based active contour tracking method

The invention relates to a covariance matching-based active contour tracking method and belongs to the technical field of visual tracking. In the covariance matching-based active contour tracking method, an image area energy term is modeled by using non-Euclidean geometry. The method comprises the following steps of: manually initializing a curve surrounding an objective and establishing a covariance matrix as a template of an objective contour for an area surrounded by the curve in a first frame; after the contour of the objective is obtained, recording a level set function value of the template to make preparation for a prior shape and calculating a symbolized distance function of the template; from the image of the next frame, deducing a gradient descent flow from a result of the previous frame according to the established energy functional and updating the level set function; and checking whether iteration stops or not. In the method, the tracking result is more accurate; meanwhile, the covariance matrix is used as an area descriptor and all kinds of information in an image sequence and the correlation between all kinds of information are considered comprehensively, and the method does not depend on foreground and background information distribution, so that the tracking method has universality.
Owner:BEIJING INSTITUTE OF TECHNOLOGYGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products