79 results about "Image tracking" patented technology

Disclosed are approaches that may provide image-guidance to interventionalists by providing true 3D visualization and quantitative feedback in real-time. A guidance system may allow a physician to manipulate a medical device and see a 3D rendering with quantitative feedback floating in mixed reality, next to standard monitors. Image tracking may detect and co-register the medical device's 3D position using, for example, bi-plane C-arm X-ray fluoroscopy and provide a 3D trajectory as quantitative feedback. Patterns in a fluoroscopic image may be used to accurately determine an object's z-position from a single angle projection.

The invention discloses a non-motor vehicle passenger-vehicle identification method and system in a crowded scene, and belongs to the technical field of intelligent traffic management. In order to solve the problem of accurate recognition of non-motor vehicles, persons and vehicles in a crowded scene, image data of a traffic intersection is collected through a traffic camera, target detection processing is performed on the image data by using a target detection model, and a person target and a non-motor vehicle target are detected to obtain a person target frame and a non-motor vehicle target frame; judging whether the scene category to which the image data of the traffic intersection belongs is a crowded scene or not to obtain crowded scene image data; personnel key points and non-motor vehicle key points in the crowded scene image data are extracted; associating the personnel key points with the non-motor vehicle key points to determine personnel key points associated with the non-motor vehicle key points to obtain a personnel-vehicle association target; and identifying whether the person-vehicle associated target has traffic violation behaviors or not through an image tracking technology, and if so, identifying the identity information of the person target for traffic violation early warning and processing.

The invention relates to the technical field of image tracking, in particular to an abnormal person tracking method and system based on a face image, and the method comprises the steps: carrying out the static identity feature extraction and dynamic behavior feature extraction of a stable face feature basic image, and obtaining a static identity feature set and a dynamic behavior feature set; performing space-time alignment processing on the static identity feature set and the dynamic behavior feature set to obtain a space-time aligned feature set; performing feature decoupling processing based on the space-time alignment feature set to obtain independent identity features and independent behavior features; and carrying out abnormal mode identification on the independent behavior characteristics to obtain an abnormal behavior type classification result. According to the invention, cross-scene and cross-time-frame personnel tracking can be realized by combining the static identity features and the abnormal behavior classification result; an abnormal person moving track is drawn according to the timestamp and the space coordinate of the abnormal behavior, track coherence is improved based on continuity scores, and tracking interruption caused by view angle switching or time intervals is avoided.

The invention relates to a robot path planning method and system based on optical vision. The robot path planning method comprises the following steps: acquiring an environment image of an environment where a robot is located and robot motion information; based on the environment image, a dense three-dimensional point cloud map is obtained through contrast enhancement, signal extraction, decoding and interpolation optimization; based on the dense three-dimensional point cloud map and the color RGB visible light environment image, generating a three-dimensional point cloud map with semantic tags through background elimination, image tracking, semantic recognition and clustering; estimating the motion state of the robot based on the three-dimensional point cloud map with the semantic tag and the motion information of the robot, carrying out global and local path planning, and obtaining a local path through track smoothing; and converting the local path into the actual action of the robot in combination with the motion state of the robot. According to the invention, through the collaborative design of the active optical visual perception architecture and multiple algorithms, the balance between cost control and performance improvement is effectively realized.

This application proposes a human body recognition system management method, apparatus, storage medium, and electronic device. First, the human body registration identity of the corresponding user is determined based on the target facial image. If no corresponding human body registration image exists for the registered identity, any target human body image from the target human body image tracking set is selected as the corresponding human body registration image. By automatically using any target human body image from the human body image tracking set as the human body registration image when no corresponding human body registration image exists, this overcomes the technical problem of requiring the user to actively provide a human body registration image, reduces user intervention, saves manpower and time, and is more convenient.

The invention discloses an adaptive multispectral target image tracking system and method. The system comprises an input layer, a processing layer and a transmission layer, the input layer is used for acquiring at least two paths of video signals; the processing layer is used for performing signal conversion, intelligent identification and fusion output on the video signals; and the transmission layer is used for respectively executing local video transmission and network RTSP stream distribution on the fused and output video signals. Decoding and format conversion are carried out on an access layer through a GS2971 / GM7150 chip, so that the RV1126 chip can receive different types of videos; in a processing layer, RKMedia is adopted to schedule video streams, YOLOv5 + NPU is adopted to perform intelligent identification, and a fusion module is adopted to perform picture enhancement; and in the output layer, low delay is ensured through local SDI transmission, and remote access is realized through RTSP flow. The system is compatible with complex scenes by adopting a layered decoupling thought.

The application provides an image tracking and collecting method, an imaging system and a computer device. The image tracking and collecting method comprises: initially collecting N pieces of image data of an imaging part, recording a position parameter of the imaging part in each piece of the image data, fitting a motion curve according to the position parameter, wherein N is greater than or equal to 3; predicting a predetermined position of the imaging part when collecting a next frame of the image data according to the motion curve; and controlling the movement of a ray generating device and a signal collecting device according to the predetermined position. Through the image tracking and collecting method, the system can predict the predetermined position of the imaging part when collecting the next frame of the image, and control the signal collecting device to move according to the predetermined position, so that the imaging part is always located in the signal receiving range of the signal collecting device. In the process of photographing the human body, human intervention is not required, the quality of the dynamic image is ensured, and the efficiency of the examination is improved.

The present embodiments provide systems and methods for tracking objects using refined bounding boxes and bounding box association. The system can include image sensors, a detector, a tracker, and a server.

The present application relates to a human image tracking method and apparatus, and a non-volatile storage medium and an electronic device. The method comprises: acquiring a current image frame, and determining tracking information of the current image frame; determining a target face recognition region and a target head and shoulder recognition region from the current image frame; on the basis of a degree-of-overlap indicator between the target face recognition region and the target head and shoulder recognition region, and a first preset condition, respectively determining a first face recognition region and a first head and shoulder recognition region, and a second face recognition region and a second head and shoulder recognition region; acquiring a historical image frame corresponding to the current image frame; on the basis of a historical recognition region, determining a prediction region of a tracking object in the current image frame; matching a second region with the prediction region, so as to obtain a similarity indicator; and determining, from the second region, a target region with a similarity indicator between same and the prediction region meeting a second preset condition, and on the basis of the target region, updating tracking information of the prediction region.

The embodiment of the invention provides a target image tracking control method and system for a photoelectric turret, computer equipment and a medium, and belongs to the technical field of automatic control, and the method comprises the steps: S1, obtaining the altitude of a task area; s2, calculating an angular velocity compensation value; s3, selecting an interest target; s4, calculating an azimuth tracking miss distance and a pitching tracking miss distance; s5, calculating an initial angular velocity; s6, calculating the compensated angular velocity; and S7, the angular velocity difference value is input into a stable controller, the stable controller outputs a motor driving amount after receiving the angular velocity difference value, and a stable platform torque motor is driven to rotate. According to the method, the relative movement of the aircraft and the target is overcome, the visual axis translation caused by the linear displacement angular velocity in the flight state is counteracted, the target capturing and tracking difficulty in the flight is greatly reduced, the tracking performance of the target in the dynamic state is ensured, and the target capturing efficiency is improved. The method is small in calculation amount, low in cost, high in portability and suitable for being used on the photoelectric turret.

The invention provides an image acquisition system for tracking imaging of a rotor blade. The image acquisition system comprises a light source tracking system, a high-speed imaging system and a synchronous control system, the light source tracking system comprises a light source tracking high-precision turntable, a light source tracking reflector group, a light source tracking protective cover and an excitation light source; the high-speed imaging system comprises an imaging high-precision turntable, an imaging reflector group, a high-speed imaging protective cover, a high-speed camera and a coaxial light splitting device; the synchronous control system comprises a photoelectric coded disc, a delay signal emitter and an image tracking correction plate. The device is reasonable in structural design, all-phase 2D temperature and pressure dynamic measurement of the surface of the rotor blade can be realized by combining the pressure-sensitive (PSP) / temperature-sensitive (TSP) coating, and image blurring can be eliminated, the signal-to-noise ratio of the image can be improved and high-precision temperature and pressure measurement results can be obtained under the condition of long exposure time based on the light source and the image tracking system of the high-speed rotating mirror.

The present disclosure relates to systems, non-transitory computer-readable media, and methods for detecting and tracking edges in raster images using an advanced edge detection algorithm. For example, the disclosed systems generate a local histogram ranking of pixels within a sliding pixel window in a raster image, based on pixel values ​​located within the sliding pixel window. In some embodiments, the disclosed systems determine an edge thickness for an edge depicted in a region of the raster image enclosed by the sliding pixel window by comparing pixel ranks specified by the local histogram ranking of the pixels.In certain embodiments, the disclosed systems also provide the edge for display based on determining the edge thickness for the area of ​​the raster image enclosed by the sliding pixel window.

To accurately track an object to be tracked.SOLUTION: An image processing apparatus comprising: an acquisition unit configured to acquire images captured in time series and distance information in a depth direction in a plurality of regions of the images; a detection unit configured to detect a candidate region of an object to be tracked from the images based on image features of the images; With respect to the candidate area detected from the image, the image tracking apparatus includes an estimation unit which estimates a shielding state indicating whether or not the subject to be tracked is shielded by another subject different from the subject to be tracked, and a determination unit which determines the candidate area of the subject to be tracked from the candidate area detected from the image based on an estimation result of the shielding state.SELECTED DRAWING: Figure 6

The invention relates to a material in-situ strain DIC analysis method and system based on CT imaging, and the method comprises the steps: preparing speckles having significant X-ray absorption difference with a matrix, directly tracking speckle displacement through a CT image while carrying out the defect change of a sample in a deformation process through CT, and replacing a conventional camera to achieve speckle collection for DIC strain analysis. According to the method, synchronous analysis of defect expansion and a strain field in a material loading process is realized through staged CT scanning and image post-processing, and an integrated solution is provided for research on mechanical properties of materials. The space limitation of CT equipment is broken through, internal defect evolution and surface full-field strain data of the material are synchronously obtained, and the method is suitable for multi-scale mechanical behavior research of metal, ceramic, polymer and composite materials thereof.

The application discloses a kind of human shadow ammunition equipment warehouse-in and warehouse-out intelligent identification method and system, it is related to intelligent warehousing management technical field, by the inventory registration of each human shadow ammunition equipment in equipment warehouse, based on warehouse-out demand table, inventory information, inventory state and warehouse-in application information, carry out warehouse-out, warehouse-in operation, the visual image tracking of several human shadow ammunition equipment of warehouse-in and warehouse-out, based on image tracking result obtains several frames equipment directional picture, obtains trace information and analyzes according to the equipment directional picture of total frame number, extracts and obtains warehouse-out feature set and warehouse-in feature set, and compares with the warehouse-in and warehouse-out template configured in advance, based on the identification of abnormal warehouse-in and warehouse-out behavior according to comparison result, and information extraction is carried out to warehouse-in and warehouse-out behavior, establishes topological information tree and carries out linkage information push, realizes the safe, efficient, accurate ammunition equipment warehouse-in and warehouse-out management.

The present disclosure provides a method and apparatus for determining a pose of a tracked object in an image tracking process. In this method, a current frame of image and first point cloud information corresponding to the current frame of image are acquired, the first point cloud information being based on a relative relationship between an image capture device and an ambient space: and a pose of the tracked object in the current frame of image relative to the image capture device is determined according to the first point cloud information and object feature information of the tracked object if the tracked object is in a static state, the object feature information being based on a relative relationship between the tracked object and the ambient space. In this method, by using the first point cloud information to convert the relative relationship of the tracked object, it facilitates determining the pose of the tracked object.

The invention relates to the technical field of space debris orbit forecasting, in particular to a space debris orbit forecasting correction method based on machine learning and triangular characteristic decomposition. The azimuth angle and the pitch angle are converted into six-dimensional feature vectors through triangular feature decomposition, and a real-time mapping model of orbit forecast deviation is established in combination with a neural network. The two bottleneck problems on which traditional image tracking depends are solved: 1) in a strong laser interference scene, tracking failure caused by background light saturation of an optical camera is avoided, and real-time image feedback is replaced by deviation prediction driven by historical echo data; 2) when space debris in an earth shadow area is invisible, constructing a training set by using deviation data captured at the moment of successful laser ranging, and realizing autonomous correction of forecast errors under a non-illumination condition; the method gets rid of dependence on space debris optical visibility, realizes all-day and all-arc telescope pointing deviation correction, and improves the reliability and continuity of space debris orbit prediction under complex working conditions.

This application discloses an image tracking processing method and apparatus. The method includes: obtaining feature points and attributes of each frame image in a real-time video; performing a geometric transformation on a reference frame image according to preset transformation conditions to obtain a relation table; determining a lookup index for at least one point pair based on the attribute differences between the first and second feature points of multiple point pairs; determining an evaluation value for each lookup index based on the template image corresponding to each lookup index in the relation table; obtaining the target rotation angle and target scaling scale based on the evaluation value of each lookup index; and determining the target transformation matrix based on the target rotation angle and target scaling scale. This method can establish a relation table by performing a geometric transformation on the reference frame image, quickly calculate the evaluation values ​​of multiple lookup indices through the relation table, and use multiple evaluation values ​​to quickly find the target rotation angle and target scaling scale, effectively improving the processing efficiency of image tracking.

Embodiments of the present disclosure disclose a method, device, equipment and medium for unmanned aerial vehicle (UAV) control based on UAV vision. A specific embodiment of the method comprises: a control signal emitting device emits signals in real time; a control recognition device performs recognition processing on the unknown flying object to generate a recognition result; a monitoring UAV in a UAV group performs image tracking processing on the unknown UAV; the unknown UAV is positioned; a second unknown UAV trajectory is generated, and the first unknown UAV trajectory is corrected according to the second unknown UAV trajectory to obtain a corrected UAV trajectory; the unknown UAV is subjected to trajectory prediction processing to generate a UAV predicted trajectory; and the UAV group or the interception device intercepts the unknown UAV. This embodiment improves the regional security of the detection area.

This invention provides an image processing apparatus and an image processing method capable of reducing the processing load of image processing for identifying multiple objects from an image and improving recognition accuracy. The image processing apparatus (IPA) includes a multi-class recognition unit (122), a tracking processing unit (131), and a recognition unit (132). The multi-class recognition unit (122) performs multi-class recognition processing on an image captured by an imaging device (ID) to identify multiple categories of objects. The tracking processing unit (131) performs image tracking on the objects identified by the multi-class recognition processing, calculating the predicted position of the tracked object in the image at a later time based on the image at a previous time. The recognition unit (132) performs binary classification recognition processing on the predicted position in the image at the later time, corresponding to the category of the tracked object, to identify the category of the tracked object.

This application relates to the field of monitoring equipment control technology, and provides a control method, system, device, medium, and program product for a camera-PTZ linkage system. The method includes: after performing coordinate association calibration on the image pixels of the camera and PTZ camera in the camera-PTZ linkage system, detecting the pitch angle of the camera during operation; when a change in the camera's pitch angle is detected, correcting the initial coordinate mapping relationship between the calibrated camera and PTZ camera based on the change in the camera's pitch angle, and determining the corrected camera-PTZ coordinate mapping relationship; and controlling the PTZ camera to perform linkage tracking of the camera's monitoring image based on the corrected camera-PTZ coordinate mapping relationship. In this embodiment, when the camera's pitch angle changes, the coordinate mapping relationship between the camera and PTZ camera can be corrected according to the change in the camera's pitch angle, reducing the possibility of inaccurate PTZ camera tracking due to changes in the camera's lens angle and improving the accuracy of image tracking.

The invention discloses a vehicle three-dimensional model reconstruction method based on videos acquired by multiple street lamps, and the method comprises the steps: taking a multi-angle video acquired by the street lamps as a target vehicle in an input reconstruction video, obtaining images of the target vehicle at multiple angles through an image tracking and segmentation algorithm, and inputting the segmented images into a 3D generation model, and the obtained 3D rough model of the target vehicle is used as a prior reference for subsequent reconstruction. And using the coarse model obtained in the previous steps as reconstruction guidance, and reconstructing the implicit SDF field representation of the target vehicle end to end by using a miniaturized volume rendering method. And extracting a triangular mesh representation of the target vehicle from the SDF field, and carrying out UV expansion operation on the triangular mesh to obtain a matched texture map. And finally, improving the definition of the texture to obtain a target vehicle model. The method is excellent in performance in complex urban traffic scenes which are difficult to process, the calculation efficiency and precision are also remarkably improved, and the method has wide application prospects and market value.

The present application belongs to the live broadcast technical field, disclose a virtual anchor multi-camera view automatic switching method and system based on image tracking. Initialize and maintain a live broadcast process state machine, get the current live broadcast state of its output, receive the video stream and audio stream of the driver; obtain the time series human key point data from the video stream through the posture estimation model, and obtain the primary action semantic vector through the time series action classification model analysis; identify the speech text stream from the audio stream, and match the speech text stream through the multi-level keyword trigger library to obtain the instruction intensity vector; the intention arbitrator predicts the comprehensive intention confidence; when the comprehensive intention confidence exceeds the dynamic threshold, the current action is determined as an effective shot switching instruction, and the corresponding virtual camera position switching is triggered according to the primary action semantic vector. The present application can accurately distinguish effective instruction actions and unconscious habitual actions based on intelligent intention understanding, thereby effectively reducing the false triggering rate of shot switching.

The invention relates to a living body capillary dynamic nerve imaging tracking method, system and device and a medium, and the method comprises the steps: collecting the continuous capillary image data of a living body tissue; performing blood vessel segmentation processing on the continuous capillary image data, and identifying a capillary network skeleton for motion analysis; based on the capillary network skeleton, calculating a motion vector of the segmented continuous capillary image data, and obtaining a motion track of red blood cells in the capillary; according to the motion trail, haemodynamic parameters including the flow velocity, the direction and the blood oxygen level are generated; synchronously acquiring local neural activity data of the continuous capillary image data under the same space-time coordinate system; and according to the haemodynamic parameters and the local neural activity data, generating a tracking result of the in-vivo microvascular haemodynamics and neural activity, so as to achieve the purpose of performing synchronous tracking and dynamic correlation analysis on the microvascular haemodynamics and neural activity in the in-vivo environment.

The invention relates to an image tracking loss judgment method, which belongs to the technical field of computer vision and image processing, and comprises the following steps of: 1, initializing a target appearance model; 2, extracting a subsequent frame target appearance model; step 3, performing similarity calculation based on histogram correlation; 4, tracking loss judgment based on time sequence continuity; 5, updating the target appearance template; the method has the advantages that by introducing judgment logic based on time sequence continuity, it is required that the similarity condition for judging target loss needs to continuously meet multiple frames, a more robust judgment model is constructed, the problem that in an existing method, misjudgment or missed judgment is likely to happen to target unmanned aerial vehicle tracking in a complex scene is solved, and the target unmanned aerial vehicle tracking accuracy is improved. Therefore, the reliability of the ground tracking system is improved.

The present application belongs to the technical field of X-ray detection, and specifically relates to an image tracking navigation method for an X-ray machine, which can calculate the range of the translation amount according to the geometric information of the tilt angle of the detector and the rotation angle of the circuit board, and then the desired window position. Through precise geometric calculation and image processing algorithm, the present application effectively solves the problem of missing detected points caused by the tilt of the detector or the rotation of the object table during the X-ray detection process. This method can dynamically adjust the detection parameters to ensure that the region of interest always remains within the detection field of view, thereby significantly improving the accuracy and reliability of PCB detection. Especially for small components such as high-density integrated BGA solder joints, this method avoids the detection blind area caused by background interference or changes in viewing angle, making defect identification more accurate and efficient.

The invention discloses a follow-up tracking grinding machining center for an engine crankshaft connecting rod neck, and relates to the technical field of engine crankshaft machining equipment. A conveying and feeding assembly used for constant-speed conveying of crankshafts, a crankshaft transferring assembly used for automatic clamping, transferring and coaxial centering of the crankshafts, a rotary clamping assembly used for high-precision coaxial clamping and stable rotary driving of the crankshafts and a grinding assembly used for collecting contours of connecting rod necks in real time are sequentially arranged on the grinding workbench. The follow-up grinding assembly is used for dynamic grinding matched with the rotation track of the crankshaft connecting rod neck. The problems that an existing crankshaft connecting rod neck grinding device lacks full-process precision control, and grinding errors are prone to being caused by clamping; the machining automation degree is low, and manual intervention is large; technological parameter regulation and control are not flexible, and equipment adaptation universality is poor; the operation threshold is high, electric control has no redundant backup, and the maintenance difficulty is large.

Disclosed is a visual tracking method of robot, wherein an execution body of the visual tracking method of the robot is a robot fixedly equipped with a camera and an inertial sensor; the visual tracking method of the robot includes: the robot performing image tracking using a window-based matching mode, and when the robot succeeds in performing tracking using the window-based matching mode, the robot stopping performing image tracking using the window-based matching mode, and then the robot performing image tracking using a projection-based matching model; and then, when the robot fails to perform tracking using the projection-based matching mode, the robot stopping performing image tracking using the projection-based matching mode, and then the robot performing image tracking using the window-based matching mode.

This invention relates to a robot path planning method and system based on optical vision, comprising the following steps: acquiring environmental images of the robot's environment and robot motion information; obtaining a dense 3D point cloud map based on the environmental image through contrast enhancement, signal extraction, decoding, and interpolation optimization; generating a semantically labeled 3D point cloud map based on the dense 3D point cloud map and a color RGB visible light environmental image through background culling, image tracking, semantic recognition, and clustering; estimating the robot's motion state based on the semantically labeled 3D point cloud map and robot motion information, performing global and local path planning, and obtaining a local path after trajectory smoothing; and converting the local path into the robot's actual actions by combining the robot's motion state. This invention effectively achieves a balance between cost control and performance improvement through an active optical vision perception architecture and multi-algorithm collaborative design.