151 results about "Video tracking" patented technology

The invention provides an accident liability intelligent determination method and system fused with video tracking, and the method comprises the steps: S1, carrying out the alignment evidence collection of a video, an IMU, a GPS, and an OBD, and forming an evidence chain through framing Hash, chain Hash, and a timestamp; s2, completing detection segmentation on the shared trunk multi-branch network; s3, in combination with appearance re-identification and a motion model, realizing cross-frame association, and outputting track, speed and shielding recovery; s4, mapping the trajectory to BEV according to depth and calibration, and constructing a space-time scene graph by fusing lanes, signals, speed limit and the like; s5, extracting events such as lane changing, line merging and signal rushing, and positioning a conflict point and a way giving relation; s6, generating error factors according to a rule base and priority judgment; s7, quantifying the responsibility proportion according to the weight, the collision participation time and the reaction time; and S8, generating a report evidence of the key frame and the rule list. The system is composed of an acquisition evidence obtaining module, a video understanding tracking module, a space-time reconstruction module, an event extraction module, a rule reasoning module, a responsibility quantification module and a report evidence storage module.

The invention provides an open vocabulary 3D segmentation method based on a three-dimensional Gaussian sputtering technology, and the method comprises the steps: S1, obtaining a multi-view 2D image set, and carrying out the initialization three-dimensional reconstruction; s2, segmenting the multiple views by using SAM to obtain a 2D mask; s3, associating masks of the same instance by using a video tracking segmentation model (SAM2); s4, extracting open vocabulary features of each mask by using a visual language model (CLIP); s5, performing average pooling and dimension reduction on the CLIP feature of the mask of each instance to form a representative open vocabulary feature of the instance; s7, training the model, and supervising the embedding of a low-dimensional embedding feature # imgabs0 #; and S8, comparing the CLIP feature of the query text with the semantic feature of the Gaussian body to obtain the Gaussian body with high correlation, and generating a mask of the query object through rasterization rendering. According to the method, only a multi-view 2D image is needed, and the three-dimensional object and the scene are segmented by combining the three-dimensional Gaussian sputtering technology and the open vocabulary technology, so that the problems of difficulty in 3D data acquisition and high cost are solved.

The invention discloses a highway tunnel lighting intelligent control method and system based on the Internet of Things, a sensing module is composed of a millimeter wave radar vehicle detector, a tunnel interior brightness sensor, a tunnel exterior brightness sensor and a vehicle speed detector, and a transmission module is a dual-mode communication module composed of a Zigbee wireless network module and a power line carrier communication module. A fuzzy control algorithm library is built in an edge computing gateway of the control module and is used for fusing data of the sensing module in real time and generating a dimming instruction, and each Internet of Things segmentation controller of the control module manages a plurality of lamps and is internally provided with a fault isolation circuit; and an intelligent management platform is arranged in the application module and is used for receiving an instruction of the control module and generating a dynamic dimming strategy. According to the invention, the millimeter-wave radar and video tracking data fusion, Zigbee + PLC dual-mode transmission and edge gateway self-learning algorithm are combined; according to the invention, the defects of the traditional scheme in the aspects of real-time performance, reliability, energy efficiency ratio and operation and maintenance cost can be systematically solved.

The invention discloses an RGBT target tracking method based on convolution attention fusion, and belongs to the technical field of computer vision, and the method comprises the following steps: 1, video preprocessing; step 2, feature extraction: encoding the input image information into a feature sequence through linear mapping and a Transform encoder; step 3, local feature enhancement and fusion: selecting a local region through a sliding window, and performing cross attention calculation on the local feature of the current modal and the global feature of another modal so as to realize local feature enhancement and effectively fuse cross-modal information; and step 4, target tracking prediction. According to the method, more stable RGBT target tracking is realized through full extraction and fusion of local feature information, and the performance of the method on a plurality of video tracking data sets is superior to that of other advanced RGBT tracking algorithms.

The invention discloses an in-situ global foreign matter checking and grabbing system for reactor internals on the lower portion of a reactor. The in-situ global foreign matter checking and grabbing system for the reactor internals on the lower portion of the reactor comprises a base device, an adjusting device and a foreign matter grabbing device. The foreign matter grabbing device is arranged on the base device through the adjusting device. The foreign matter grabbing device comprises a flexible arm assembly, a flexible arm driving assembly, a quick change assembly, at least one clamping jaw used for grabbing foreign matter and a video tracking and checking assembly used for searching the foreign matter and monitoring the grabbing condition. The flexible arm assembly is arranged on the adjusting device; the at least one clamping jaw is detachably arranged on the flexible arm assembly through the quick-change assembly; the video tracking and checking assembly is arranged on the adjusting device; according to the invention, the flexible arm assembly carries various types of clamping jaws to carry out automatic grabbing operation, so that the in-service nuclear power plant reactor internals can accurately, efficiently and safely grab and clean foreign matters in the in-situ interior of a pressure vessel in a whole region.

The invention relates to the technical field of intelligent city management and computer vision, and discloses an intelligent street lamp inspection management system based on deep learning. The system comprises an image recognition module, an algorithm processing module, an environment module, a data analysis module, an authority management module, an intelligent path planning module, a video tracking module, a multi-scale detection module, an image conversion module and a detection storage module. And the image recognition module is used for carrying out image data acquisition on the street lamps and realizing image recognition, extracting features of the street lamps by using a convolutional neural network, and carrying out target detection by using a Yolov4 algorithm model. Through automatic image recognition and detection functions, the system can replace manual visual inspection and monitor the state of the street lamp in real time, the time required for inspection is greatly shortened by using a deep learning target detection algorithm and an efficient intelligent path planning technology, the system can complete inspection of a large number of street lamps in a short time, and the inspection efficiency is improved. And the inspection efficiency is obviously improved.

A single-target tracking method and device based on a state space model and attention and a medium are characterized in that firstly, hierarchical feature extraction is carried out, hierarchical modeling is carried out on a template and a search area through a mixed attention mechanism and the state space model, and a time sequence token is introduced into each frame at the final stage of hierarchical feature extraction to carry out target information aggregation; and aggregating the time sequence tokens by using a state space model to realize cross-frame information propagation, multiplying the obtained time sequence tokens by corresponding frame feature maps to serve as feature enhancement, and finally predicting a bounding box of a target object by a prediction head to realize target representation. According to the method, the state space model and the attention mechanism are combined for single-target tracking, especially target tracking under a large-resolution video, the state space model is introduced into the field of single-target tracking, the advantage of linear complexity of the state space model is brought into full play, meanwhile, the disadvantage of limited retrieval capacity of the state space model is made up, and high-resolution video tracking is achieved. And finally, the consumption of computing resources is reduced while accurate tracking is realized.

The invention discloses a ship intelligent cargo tracking and monitoring method and system, and the method comprises the steps: building a ship intelligent cargo tracking and monitoring frame based on a GPS positioning system and an Internet system, building a cargo tracking and monitoring APP, registering and building account information, registering cargoes, receiving and verifying the cargoes, and tracking the cargoes. Real-time monitoring and cargo inquiry are carried out on ship cargo transportation conditions; a consigner inquires cargo transportation information and receives the cargo through a cargo tracking and monitoring APP system; the cargo arrives at a transportation destination and receives cargo information with a consignee; and the consignee confirms cargo receiving and carries out transportation evaluation. According to the intelligent ship cargo tracking and monitoring method, actual transportation of cargoes is known through GPS positioning tracking, temperature tracking detection, humidity tracking detection, video tracking detection, vibration tracking detection, inclination tracking detection, pressure tracking detection and time tracking detection of ship cargoes.

The invention discloses a multi-granularity video target tracking method and device based on a unified architecture and a medium, and the method comprises the steps: firstly designing a task specific prompt and a unified decoder, and unifying target state expressions of different granularities of masks, bounding boxes and key points through a task adaptive memory mechanism; secondly, constructing a multi-granularity video tracking data set, and fusing a memory matching core principle to realize compatibility of a single model to a multi-granularity tracking task; and finally, based on large-scale video data training, enabling the model to synchronously optimize the cross-granularity tracking capability. The invention provides a unified tracking architecture, solves the problems of insufficient generalization and parameter redundancy caused by task module independence in the existing method, dynamically adapts to feature expression requirements of targets with different granularities through a task adaptive memory mechanism, and realizes full-granularity continuous frame tracking by taking an initial state as input. And a model does not need to be independently designed for each task. According to the method, an efficient solution is provided for cross-granularity video analysis, and meanwhile standardized development of a general target tracking technology is promoted.

The invention discloses a security risk intelligent response method and system. The response method comprises the following steps: acquiring video streams, equipment states and environment sensor data in real time through a multi-source interface; processing through a feature fusion engine to generate a threat matrix containing threat type judgment and confidence; matching an emergency plan library based on a dynamic weighting algorithm; an equipment control instruction queue is generated according to a matching result, conflict detection is executed, and multiple instructions of the same equipment are sequenced according to fire control, access control locking and video tracking; and finally, acquiring an equipment execution state in real time and feeding back the state to a plan library to form closed-loop control, thereby solving the problems of high false alarm rate and response lag in traditional security and protection. Historical and real-time data are fused through a dynamic decision model, the limitation of a static plan is broken through, and the disposal efficiency is improved; instruction conflicts are eliminated through a conflict resolution mechanism; and a self-healing system and model correction are combined to form double-closed-loop optimization.

The invention discloses a video tracking control method and device for a non-linear moving target based on adaptive sliding mode control, which are suitable for accurate tracking of a dynamic target. The system is composed of a visual sensing module, a main control module and an execution mechanism, the visual sensing module collects position information of a target object in real time and transmits the position information to a main control chip through serial port communication, the main control chip calculates the error between the target position and the current position of the system and the error change rate, and a sliding mode surface is designed based on sliding mode control. Meanwhile, sliding mode control parameters are dynamically adjusted in combination with a self-adaptive control algorithm, so that the robustness of the system to uncertainty and external interference is enhanced, and meanwhile the chattering phenomenon is restrained. A stepping motor control signal is optimized by introducing a dynamic acceleration and deceleration algorithm, so that control smoothness and response precision are ensured, and real-time accurate tracking of a target object is realized. The method has strong robustness, high dynamic performance and control precision on the premise of not depending on a target kinematic model, can be widely applied to the fields of unmanned aerial vehicle tracking, intelligent monitoring, robot vision and the like, and effectively solves the problems of easy buffeting, response hysteresis, insufficient control precision and the like during nonlinear target tracking in the prior art.

The invention relates to the technical field of intelligent linkage control systems for video monitoring, access control and fire alarm, particularly discloses an intelligent linkage control system for video monitoring, access control and fire alarm, and aims to solve the problems of low integration level, linkage logic solidification and high misjudgment rate of an existing security and protection system. The system comprises a multi-source data sensing fusion module, a fire situation intelligent evaluation module, a dynamic evacuation path planning module and a multi-system cooperative execution control module, and intelligent evaluation and dynamic evacuation path planning of the fire situation are realized by fusing video, access control and fire alarm data. And video tracking, entrance guard opening and closing and fire-fighting broadcasting are cooperatively controlled, so that the accuracy and efficiency of emergency response are improved.

The invention discloses a video relay tracking method and platform based on track positioning, and the method comprises the steps: calculating a road section where a target vehicle travels in real time according to the positioning data of the target vehicle on the basis of the existing traffic guarantee equipment and facility conditions, and carrying out the real-time tracking of the target vehicle based on the information of the road section where the target vehicle travels; and crossing video monitoring corresponding to the crossing that the target vehicle leaves the crossing, is about to arrive at the crossing and the next crossing of the crossing is about to arrive at is automatically tracked and played in a relay manner, so that the aim of monitoring key information such as the target vehicle and traffic road conditions in a task execution process in real time is fulfilled. The technical scheme can be applied to various tasks such as traffic police department service guarantee needing to track and monitor the target vehicle, and the problems that in the prior art, video tracking has high requirements for business personnel, and intersection video monitoring equipment is poor in playing effect are solved.

The invention discloses a target tracking system based on video perception. The target tracking system comprises a monitoring shooting unit, a data acquisition unit, a data transmission unit, a data storage unit, a data processing unit, a remote monitoring unit, an image correction unit, a feature extraction unit, a matching tracking unit and a UPS power supply unit. The monitoring shooting unit is used for carrying out video monitoring shooting through video shooting equipment; and the data acquisition unit is used for performing real-time acquisition on data generated in the process of performing video sensing tracking on the target. The invention belongs to the technical field of video tracking. By integrating the visibility, illuminance, light inlet amount, noise value and luminous flux multi-dimensional information acquisition module, the system can evaluate the shooting environment in real time, automatically adjust image parameters, effectively reduce noise interference and enhance image definition, so that the system can accurately identify target features, and the target features can be accurately identified even in a complex and changeable low-illumination environment. And rapid and accurate target tracking can be realized.

To provide a balance function management system for balance function state information generation and a balance function rehabilitation program.SOLUTION: A memory configured to store at least one processor and at least one artificial neural network model that stores instructions executable by the processor and is executed on a computing device, wherein the at least one processor is configured to input frame images of n (where n is a natural number) videos obtained by photographing a subject through n cameras to the at least one artificial neural network model, At least one of information related to the coordinates of the head, the coordinates of the center of the pupil, and the phase change of the eyeball of the subject according to the order of the frame images of the m-th (m is a natural number from 1 to n) moving image may be acquired, and information related to the movement of the head and the movement of the eyeball for generating the balance state information or performing the balance rehabilitation program may be generated using the acquired information.SELECTED DRAWING: Figure 1

The invention provides a crossing control method and system based on video analysis and a magnetic steel sensor, and belongs to the technical field of crossing control. By obtaining a train route state in front of a crossing, when a train triggers a first route position, a first position signal is obtained; tracking the train according to the first position signal video to obtain the position of the train, and giving an alarm through an audible and visual alarm; when the train triggers the second position of the access road, a second position signal is obtained, and the barrier machine is controlled to close the crossing according to the second position signal; and judging whether the current train passes through the crossing area or not according to a target identification result of video tracking, and determining to start a barrier machine to open the crossing and turn off an audible and visual alarm according to a judgment result. Therefore, the automation degree and the accuracy of barrier control are improved through video tracking and identification of the train route by judging the network, and the intelligent degree and convenience are improved.

This invention discloses a method for establishing a network for accurate target tracking in noisy environments, comprising the following steps: setting the number of training iterations for adversarial noise training, the type and magnitude of noise generated by the noise generator, and the balance coefficient of the tracker's loss function; feeding a pair of images into the noise generator, which adds noise to the image pairs according to the set parameters, and then feeding the processed image pairs into the backbone network for feature extraction; classifying and regressing the response map obtained after feature extraction to obtain the tracking result of the current frame, and then iteratively running this process until all video frames of all video sequences in the test dataset are traversed; recording and saving the tracking result of each frame, and quantitatively analyzing the tracking accuracy and tracking success rate. This invention, while maximizing the BCE loss value, minimizes the final tracker loss value through training, thereby enabling the tracker to handle tracking in various noisy environments while maintaining tracking performance on clean videos.

The present invention discloses a visible light and infrared target tracking method and device based on a modal competition and cooperation network. The method includes: generating a first candidate sample set centered on the predicted position of the target in the previous frame of the current frame to be tracked; inputting the first candidate sample set into a trained tracking model to obtain the target tracking result of the current frame; the tracking model includes a feature extraction network and a target classifier; the feature extraction network includes a first extraction network, a second extraction network, and a weight allocation network. Both the first extraction network and the second extraction network include the first three convolutional layers of VGG-M, and each convolutional layer contains N convolutional kernels of the same size to form a convolutional kernel group. By introducing the convolutional kernel group and multiple weight allocation networks to generate a multi-core dynamic cooperative convolutional kernel, and using several convolutional kernels included in the convolutional kernel group, more diverse features can be extracted, achieving a good multi-modal video tracking effect.

The present invention discloses a video tracking method, device, terminal, and medium. The present invention combines the semantic features of the target tracking point itself with the surrounding contextual features, improving the ability to query spatial information. This allows the correct spatial features of the target tracking point to be found even in long videos or when the tracking target undergoes significant changes. Point tracking tasks in long videos are accomplished by forming a point query based on the semantic features, contextual features, and point positions of the target tracking point, and updating the point query frame by frame. The method is applicable to fields such as video editing, augmented reality, 3D reconstruction, and optical flow estimation.

This invention discloses a video tracking and recognition method for the spatial trajectory of hoisted objects based on multi-source data fusion, belonging to the field of machine vision monitoring technology. The method includes acquiring pre-loaded physical boundary constants of the tower crane and simultaneously collecting multi-source data during tower crane operation. Based on the physical boundary constants, kinematic calculations are performed on the multi-source data to obtain the relative vertical displacement, the three-dimensional absolute spatial coordinates of the luffing trolley center point, and the absolute coordinates and three-dimensional rotation matrix of the camera's optical center in the global three-dimensional coordinate system. Edge detection is performed on the temporally aligned color video frame matrix to extract candidate wire rope contours and generate a binary edge pixel matrix. Collinear edges are mapped to polar coordinate space using Hough transform. This invention reduces the uncertainty of monocular vision depth scale and improves stability and accuracy under conditions of large-amplitude lifting and complex swinging.

The invention provides a multi-species wild animal video tracking method based on deep learning, and belongs to the technical field of wild animal recognition. The problem that tracking task modeling is feature matching between a template and a current frame image is solved, a motion model does not need to be preset, and the method can more flexibly adapt to field scenes with irregular motion and complex environment. The system is mainly composed of a feature extraction and encoder module, a detection branch, a tracking branch, an ID matching and updating module, a trajectory reconstruction module, a tracking trajectory compensation detection mechanism, a triple association strategy and a trajectory unification strategy, by adopting the tracking method based on time sequence correlation modeling, the interference problem of a complex environment and an animal movement mode in wild animal tracking can be effectively solved, and the method is more suitable for target tracking in a field dynamic scene.

The invention relates to a photoelectric radar integrated detection method and system based on target autonomous identification, and relates to the technical field of photoelectric system and radar system integrated detection. The method comprises the following steps: step 1, area selection; 2, preset photographing point position conversion is carried out; step 3, carrying out azimuth assembly sector scanning; step 4, splicing module processing; 5, processing by an information fusion identification module; step 6, processing by a display and decision module; and step 7, photoelectric video tracking mode switching. According to the photoelectric radar integrated detection method and system based on autonomous target identification, the authenticity and tracking priority of the target are judged through combination of the target detected by the radar and the target identification result of the photoelectric system on the wide-width wide-angle image shot by the photoelectric system, the tracking control precision of the photoelectric system is improved, the radar false alarm rate is reduced, and the detection accuracy of the photoelectric system is improved. The guiding efficiency is improved.

The embodiment of the application discloses a kind of normative checking method and device based on artificial intelligence, first receive and respond target detection instruction, select one or more target detection objects based on the target detection instruction, in order to better detect the operation article or operation action corresponding to target detection object, the target detection area corresponding to these target detection objects is acquired in advance, and the two are bound to form a normative check target;Image data is collected by image acquisition equipment, including static picture data and dynamic video tracking data, the data obtained after image data is preprocessed is more accurate, which helps to better extract feature points and then accurately identify targets;Embodiment greatly improves the monitoring feasibility and flexibility of student experiment process based on artificial intelligence, helps to accurately judge the operability of student experiment, and improves efficiency.

The invention discloses a target intelligent tracking method, device and equipment based on multifunctional loop fusion, a storage medium and an unmanned aerial vehicle, and the method comprises the steps: determining image motion feature data of a motion process according to a tracking video collected by a video tracking device of the unmanned aerial vehicle for the motion process of a target object, determining signal motion characteristic data of the motion process according to a tracking signal acquired by a signal tracking device of the unmanned aerial vehicle for the motion process; determining adjustment motion feature data of the motion process through reciprocal enhancement of the signal motion feature data to the image motion feature data; and controlling the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjustment motion feature data and the signal motion feature data in the annular attractor network. Through information fusion of a reciprocal enhancement technology and an annular attractor network, accurate target tracking in a high-speed motion scene is realized, and the problems of high delay and large error when an unmanned aerial vehicle tracks a target object are solved.

The application provides an accident responsibility intelligent judgment method and system fusing video tracking, comprising: S1 aligning video with IMU, GPS and OBD for evidence collection, adopting frame hashing, chain hashing and timestamp to form an evidence chain; S2 completing detection segmentation on a shared main branch network; S3 realizing cross-frame association by combining appearance re-identification and motion model, and outputting trajectory, speed and occlusion recovery; S4 mapping the trajectory to BEV according to depth and calibration, and constructing a space-time scene graph by fusing lanes, signals, speed limits and the like; S5 extracting events such as lane changing, merging and signal violation, and positioning conflict points and yielding relationships; S6 generating fault factors according to a rule base and priority; S7 quantifying responsibility proportion according to weight, collision participation and reaction time; and S8 generating a report for storing evidence of key frames and rule lists. The system comprises collection and evidence collection, video understanding and tracking, space-time reconstruction, event extraction, rule reasoning, responsibility quantification and report storage modules.

This application provides a method, apparatus, medium, and device for video tracking of the goods circulation process. The method includes: acquiring the barcode scanning time of the goods; distributing tasks to a controller, so that the controller controls the circulation of the goods within a goods sorting system and records the time of the goods at each circulation node; capturing video of the internal workings of the goods circulation system using multiple cameras positioned at different locations; searching for videos corresponding to the goods based on the barcode scanning time and the time of each circulation node; and processing the searched videos to obtain a video tracking file of the goods. The video tracking file allows for quick and convenient visual tracking of the goods circulation system.

The invention provides a signal generator, method and system for a dual-channel video tracker of a Cameralink interface, and relates to the field of image processing and testing. The generator comprises a ZYNQ time sequence control module, a Cameralink coding module, a communication control module and a power supply module. Two paths of hundred-hertz high-frame-frequency image time sequence signals are generated through a ZYNQ time sequence control module, and stable output of two paths of Cameralink video sources is ensured. Through the instruction of the communication control module, the logic of the signal control module in the ZYNQ time sequence control module can be quickly modified, the change of target characteristics and background characteristics is realized, and various complex signal sources are provided, such as target complex motion trail realization, target shielding realization and sea-sky-line simulation. The method comprises the following steps: configuring two chips for a Cameralink coding module; and finally, the generated differential signals are sent to Cameralink output connectors XP1 and XP2, so that stable output of the two paths of Cameralink video sources is further ensured.

The present invention discloses a target-guided cross-camera continuous video tracking and monitoring method, comprising the following steps: obtaining target information, obtaining camera resource capability information of the target area and sensitivity information of the monitoring area; obtaining user priority information and monitoring task importance information; calculating, based on the target information and the camera resource capability information, alternative cameras capable of monitoring the target; calculating the monitoring effect of each alternative camera and the priority number of each alternative camera; and selecting and switching cameras for target tracking based on the monitoring effect and priority number of the alternative camera. The present invention adopts dynamic resource scheduling, incorporates user priority, the importance of the task performed by the camera, the sensitivity of the monitoring area, and the capability of the camera resource into resource scheduling, performs resource scheduling based on the monitoring effect of the camera, dynamically switches cameras, and economically realizes multi-target cross-camera continuous video tracking and monitoring.

The invention discloses a vehicle multi-target tracking method capable of keeping a unique mark number, and belongs to the field of computer vision, aiming at the problem that the target mark number is not unique after the observation angle is changed or shielded during multi-target tracking. Vehicle videos blocked at the same observation angle or at different observation angles are collected to construct a data set, and the data set is processed according to the following process: inputting a video 1, tracking to obtain a vehicle target label and position, and intercepting a target picture frame by frame to form the data set for training a DenseNet121 network; inputting a video 2, tracking to obtain target information, inputting a target picture into the trained DenseNet121 network for re-identification, reserving an old label for an old target, endowing a new label for a new target, intercepting a new target picture to expand a data set, and updating the training network by using the data sets of the video 1 and the video 2; and inputting a video 3 and a subsequent video, and repeating the steps, namely tracking the target, re-identifying the label, expanding the data set and updating network training.