61 results about "Feature tracking" patented technology

The invention discloses an equipment health feature extraction method and system based on a depth auto-encoder, and relates to the technical field of equipment feature extraction. According to the equipment health feature extraction method and system based on the depth auto-encoder, S1, equipment operation state data, environment interference data and channel attribute data are collected and preprocessed to construct a standardized stable feature expression set; s2, channel disturbance sensitivity is evaluated, and a signal retention strategy is dynamically selected; s3, analyzing signal disturbance components and switching a channel input form; s4, comprehensively evaluating a channel health fluctuation level, and controlling a stability identifier and a feature tracking strategy; and S5, jointly monitoring a feature offset and attribution relationship, correcting the feature mapping path and triggering abnormal attribution analysis. The problems that in an extreme environment, operation signals are prone to being disturbed, health feature extraction lacks a structural stability judgment mechanism, and abnormal states and environmental inducements are difficult to distinguish are solved.

The invention relates to image processing, in particular to a complex environment-oriented power transmission line inspection image processing method, which comprises the following steps of: screening key frames with high geometric information content by quantitatively evaluating inter-frame motion amplitude and feature tracking quality; extracting features of each key frame by using a pre-trained visual model; performing intra-frame feature aggregation on the features of each key frame by using an intra-frame self-attention mechanism to obtain corresponding intra-frame feature representation; performing global feature aggregation on the intra-frame feature representations of all the key frames by using a global self-attention layer to obtain inter-frame feature representations; constructing a multi-task learning network, and carrying out end-to-end collaborative optimization on depth estimation, image defogging and high-level semantic segmentation; inputting the inter-frame feature representation into a multi-task learning network, and obtaining a restored clear image through deep fusion of scene depth information and image degradation priori; according to the method, the defect that the dual requirements of quality improvement and feature retention of the power transmission line inspection image in a complex environment cannot be met can be overcome.

The invention discloses a state estimation method based on mixed features and adaptive key frame selection, and belongs to the field of autonomous navigation of unmanned systems. The implementation method comprises the following steps of: acquiring a plurality of images, and synchronously acquiring inertial measurement data of a six-axis IMU (Inertial Measurement Unit); the method comprises the following steps: performing key frame decision on a current frame image according to inertial measurement data, and classifying the image into a key frame image and a non-key frame image: acquiring a key frame feature matching set through an entropy maximization self-supervision feature extractor and an image attention matcher, and constructing a visual reprojection error term based on triangulation map points; non-key frame feature points are extracted by adopting an improved corner detector, feature displacement is obtained through IMU constraint optical flow tracking, and a visual re-projection error and an IMU pre-integration residual term are generated; and based on a vision-inertia tight coupling optimization module solution system, combining a vision reprojection error, an IMU pre-integration residual error and a pneumatic constraint residual error to construct a tight coupling optimization problem, and solving an optimal pose state. The method is beneficial to navigation of the small-sized high-speed unmanned aerial vehicle.

The invention discloses a multi-modal feature tracking method based on a traffic scene, and relates to the technical field of traffic scene feature tracking, and the method comprises the steps: collecting traffic scene RGB image data and event camera data, and constructing a multi-modal time series data set; constructing a multi-modal feature tracking model; extracting a time sequence motion feature of the event camera data, and generating a displacement prediction vector; state prediction updating is carried out, and a tracking prediction result is generated; and continuously tracking the traffic target to generate motion tracking trajectory data. According to the invention, the technical problems of poor image quality, low tracking accuracy, weak single-mode tracking anti-interference capability and difficulty in adapting to high-frequency tracking requirements caused by the influence of factors such as illumination, weather, exposure and the like during traffic scene feature tracking in the prior art are solved; the technical effects of realizing collaborative feature tracking of the RGB image and the event camera multi-modal data in the traffic scene and improving the frequency, accuracy and anti-interference capability of feature tracking are achieved.

The application discloses a video instance segmentation method and system based on global feature enhancement, and belongs to the technical field of computer vision and deep learning, which comprises the following steps: inputting continuous video frames into a double-channel decoupling feature extraction module to obtain appearance and position features; inputting the two types of features into a tracker and a global feature memory updating module to dynamically update and obtain long-time sequence global context features; completing instance detection matching by the tracker, judging target state to obtain disappeared and newly appeared target queries; combining long-time sequence features to model current frame time sequence to generate enhanced features; finally, the tracker completes segmentation and tracking based on the enhanced features and the target queries, and outputs a segmentation mask sequence, an instance ID, a category ID and a tracking state. Through double-channel decoupling, double-memory block updating, multi-cost matching and multi-weight feature fusion, the application strengthens appearance feature representation and long-time sequence information utilization, and significantly improves segmentation accuracy and tracking robustness in complex scenes.

This invention relates to a global feature tracking method for time-varying data based on unsupervised segmentation. It constructs a volume data segmentation network to segment the input volume data and achieve global tracking. This invention introduces deep learning to segment volume data, separating feature domains from the background. This enables automatic and accurate tracking of specific features in complex data without any manual annotation, reducing tracking complexity and improving tracking accuracy. It introduces a global feature tracking method; by selecting a target feature, it can track the trajectories of features similar to the target feature across all time steps. Users can select any feature at any time step to track features of interest in time-varying data. It can track extracted features from a global perspective, avoiding tracking errors and defects caused by local tracking methods. Furthermore, it adds the tracking of spatiotemporally similar features, simultaneously tracking the complete paths of spatially similar features of the target feature, improving feature tracking accuracy.

A surgical navigation system includes an extended reality head-mountable display (HMD). The surgical navigation system further includes a computer system comprising at least one processor, wherein the computer system is configured for performing operations. The operations include receiving camera images while a surgeon is wearing the HMD and performing a surgical procedure on a subject. The operations further include identifying at least one image feature depicted in the images. The operations further include tracking a location of at least one surgical object using the image feature. The operations further include sending one or more visual indicators to the HMD to display, to the surgeon, a target location for the surgical object for the surgical procedure. The target location includes a location external to the subject for the surgeon to use the surgical object to perform the surgical procedure on the subject.

This invention discloses a terahertz photoelectric image registration method using video feature following, comprising the following steps: image preprocessing; SURF-based first-frame image registration: calculating feature points of the matching first-frame radar image and first-frame optical image based on SURF features; video following feature extraction; and sequence image registration. This invention utilizes a video feature following method to transform the complex problem of heterogeneous image registration into a homogeneous image registration problem, reducing the problem complexity and improving registration accuracy.

The invention belongs to the technical field of industrial automation and safety monitoring, and particularly relates to an industrial door anti-pinch safety early warning method based on video stream moving target detection, which comprises the following steps of: S1, acquiring a real-time video stream of an industrial door area, performing preprocessing and feature tracking on the real-time video stream, mapping an image coordinate to a world coordinate system, and acquiring a real-time video stream of the industrial door area; acquiring a real-time position parameter and a movement speed parameter of the moving target; and S2, according to the motion speed parameter of the motion target and the direction relative to the door body, a radial approximation equivalent threat speed index is calculated, and the radial approximation equivalent threat speed index is used for representing the equivalent threat speed of sprint of the motion target to the door body. By establishing the multi-dimensional dynamic collision risk model, differentiated accurate response to high-speed forklifts and low-speed pedestrians is achieved, and the passing efficiency of the industrial door is maximized on the premise that safety is guaranteed.

Aspects of this technical solution can increase speed of processing and lower computational hardware complexity in motion detection, while maintaining integrity of motion detection across image frames. For example, in image-processing environments associated with autonomous or semi-autonomous navigation (e.g., driving, robotic navigation, etc.), a large volume of image data is to be rapidly and accurately processed to maintain reliable and up-to-date models of a physical environment. Thus, embodiments in accordance with this disclosure can provide high-speed and accurate motion detection of input frame data.

The invention relates to an image data sampling architecture for image feature tracking. According to the technical scheme, the processing speed of motion detection can be improved, the complexity of calculation hardware is reduced, and meanwhile the integrity of motion detection across image frames is kept. For example, in an image processing environment associated with autonomous or semi-autonomous navigation (e.g., driving, robotic navigation, etc.), there is a need to quickly and accurately process large amounts of image data to maintain a reliable and latest model of the physical environment. Therefore, high-speed and accurate motion detection of input frame data can be provided according to embodiments of the present disclosure.

The invention provides a motor fault detection method and system based on vibration and current double-domain fusion, and relates to the technical field of motor fault detection.The method comprises the steps that an electric vehicle motor serves as a fault traceability root node, hierarchical progressive structured association topology pruning is conducted, and a minimum complete diagnosis topology is output; backtracking motor time sequence fault records to carry out fault association mining, and constructing an association fault probability matrix; probability graph model parameterization is carried out to obtain a fault probability association topology; executing multi-thread cross-modal feature tracking identification, and outputting a real-time motor fault prediction result and K associated fault prediction results; performing Bayesian inverse probability derivation, and outputting a real-time motor fault probability; and binding as a fusion diagnosis conclusion and outputting. According to the invention, the technical problem of inaccurate fault diagnosis caused by the fact that a single signal monitoring method in the prior art cannot accurately identify multiple fault modes when facing complex motor faults is solved.

This invention discloses a real-time target tracking method for unmanned aerial vehicles (UAVs) based on asynchronous semantic correction. It employs a spatiotemporally decoupled architecture of a high-frequency geometric main thread and an asynchronous semantic correction thread. The main thread estimates the target's geometric state frame-by-frame using sparse feature tracking and filtering models, outputting a target prediction bounding box. The asynchronous thread calls a lightweight semantic segmentation model at low frequency according to trigger conditions to generate a target semantic mask and uses the motion trajectory within the delay period for spatiotemporal alignment compensation. The system performs dual reliability checks on semantic input and output, updating the target state through a hierarchical write-back mechanism. When the main thread's geometric propagation fails, an emergency fallback tracker takes over the output to ensure tracking continuity. This invention enables category-independent, real-time, robust target tracking in open scenes on edge platforms.

The embodiment of the invention provides a driving scene image generation method and device based on Mama, electronic equipment and a storage medium, and relates to the technical field of image processing. The method comprises the following steps: performing coding quantization processing on an image sequence in an acquired real driving scene to obtain a discrete potential representation sequence, performing optical flow estimation and feature tracking on the image sequence to generate a sparse motion trail diagram of the image sequence, selecting multiple frames of images from the image sequence as key frames, and obtaining a sparse motion trail diagram of the image sequence; extracting guide pixels from each key frame based on the sparse motion trail diagram, inputting the sparse motion trail diagram and the guide pixels into a double-branch conditional encoder to obtain a fused conditional vector, inputting the potential representation sequence and the conditional vector into a Mamba model to generate a predicted potential representation sequence with a target length, and decoding the predicted potential representation sequence to obtain a predicted potential representation sequence with a target length. And obtaining a predicted image sequence. Therefore, the sense of reality and the dynamic continuity of the predicted image sequence can be obviously improved.

Techniques and systems are provided for pose prediction. For instance, a process can include predicting a future pose of the apparatus; identifying a set of tracked features; and selecting a subset of cameras from a plurality of cameras for feature tracking based on the identified set of tracked features and the future pose of the apparatus.

The application discloses a line feature tracking method and device based on LK optical flow, and belongs to the field of computer vision.The application solves the problem that the existing method cannot guarantee real-time performance due to large operation amount required by interframe line feature extraction and matching.The method of the application first extracts line features from a collected first frame image by using an improved EDlines algorithm; then, line feature tracking is performed on a next frame image based on the line features in the first frame image and the LK optical flow method, and the image pyramid form is adopted during the line feature tracking.The application solves the problem that the traditional method requires large operation amount for descriptor extraction and interframe matching in each frame image, reduces the time consumption of line feature extraction and interframe matching, and guarantees real-time performance.The method of the application can be applied to line feature tracking in images.

A method including capturing at least one image using a plurality of cameras, identifying a plurality of facial features associated with the at least one image, stabilizing a location of at least one facial feature of the plurality of facial features, converting the location of the at least one facial feature to a three-dimensional (3D) location, predicting a future location of the at least one facial feature, and rendering a 3D image, on a flat panel display, using the future location of the at least one facial feature.

This invention relates to direct scale level selection for multi-level feature tracking. A method for mitigating motion blur in a visual-inertial tracking system is described. In one aspect, the method includes: accessing a first image generated by an optical sensor of the visual tracking system; accessing a second image generated by the optical sensor of the visual tracking system, the second image following the first image; determining a first motion blur level of the first image; determining a second motion blur level of the second image; identifying a scale change between the first and second images; determining a first optimal scale level for the first image based on the first motion blur level and the scale change; and determining a second optimal scale level for the second image based on the second motion blur level and the scale change.

The application relates to the field of computer vision and relates to a measurement method for the movement of a marine floating body. The measurement method comprises the following steps: acquiring a video of the movement of the marine floating body on a water body; determining a feature tracking point set corresponding to each frame of image in the video, the feature tracking point set being related to the marine floating body; and determining six-degree-of-freedom movement parameters of the marine floating body according to the feature tracking point set. The application realizes the measurement of three-dimensional complete movement information, i.e. the measurement of six-degree-of-freedom movement parameters, and can improve the robustness and completeness of the measurement of the movement information of the marine floating body.

The invention belongs to the field of geological disaster monitoring and early warning, and discloses a landslide deformation monitoring method based on SAR image offset tracking and multistage filtering, which comprises the following steps: dividing a registered SAR image data set into image pairs, calculating an original offset field of each image pair by using a block phase cross-correlation algorithm, and calculating the original offset field of each image pair; based on correlation threshold and signal-to-noise ratio threshold filtering, median filtering, statistical aggregation, amplitude threshold filtering, vector consistency filtering and density clustering analysis, multi-stage filtering is fused to comprehensively process an original offset field, noise is effectively removed, systematic deviation is corrected, a real and coherent landslide deformation signal is refined, and the landslide deformation accuracy is improved. Therefore, a more accurate and more reliable landslide deformation monitoring solution is provided. The method focuses on image pixel feature tracking and post-processing thereof, integrates multiple technologies to filter noise step by step, can realize landslide identification and long-term dynamic monitoring, improves the accuracy and reliability of landslide monitoring, and provides technical support for landslide disaster prevention and control.

Embodiments are disclosed for automatic display alignment using feature tracking. In some embodiments, a method comprises: capturing, with a first camera of a first display, a first image; capturing, with a second camera of a second display, a second image, where the respective fields of view for the first camera and the second camera are at least partially overlapped; transforming the first head pose into a first reference frame for the first display; transforming the second head pose into a second reference frame for the second display; generating a dual-display environment for the first and second displays based on the first and second head poses in the first and second reference frames, respectively; presenting a first portion of the dual-display environment on the first display; and presenting a second portion of the dual-display environment on the second display.

The invention belongs to the technical field of vehicle control, and particularly relates to a four-quadrant identification dynamic visual tracking system based on feature tracking. Comprising an image data acquisition module, a four-quadrant target detection module, an optical flow tracking module, a trajectory prediction module and a steady-state target management module. The image data acquisition module is responsible for acquiring video stream data containing the surrounding environment of the four-quadrant mark and transmitting the data to the four-quadrant mark detection module frame by frame; the four-quadrant identifier detection module is responsible for performing real-time processing on an input source image and converting the input source image into an ROT region matched with a four-quadrant identifier in the source image; the optical flow tracking module is used for converting a source image transmitted by the image data acquisition module into a grayscale image to realize inter-frame tracking of four-quadrant identification feature points; the steady-state target management module is used for judging how to recover and predict a lost target position after tracking loss; and the track prediction module predicts the position of the next frame of source image through the motion model. Compared with the prior art, the technical scheme has the advantages that the tracking robustness is enhanced, the stability recovery speed is high, and the scene adaptability is high.

The invention provides a hub city associated feature tracking analysis method and system based on network analysis, and relates to the technical field of data analysis, and the method comprises the steps: collecting multi-source and multi-dimensional data of a hub city, and constructing a multi-dimensional database which comprises traffic flow, economic flow, information flow and population flow; with cities as nodes and inter-city flow data as edges, edge weights are set, and a multi-level network model of the hub cities is constructed; carrying out network index calculation based on a network analysis method, and establishing a city multi-dimensional interaction relationship; through time sequence analysis of historical data and real-time data, time sequence reverse tracking and time sequence forward trend prediction are carried out, and multi-dimensional feature association strength and development direction are analyzed. According to the invention, the technical problem that the prediction of the future city development trend is lack of accuracy and foresight because city association analysis is generally carried out based on a single data source and the interaction relationship of hub cities in multiple dimensions cannot be comprehensively reflected in the prior art is solved.

A steel ball extrusion deformation accurate measurement method for bearing rated static load determination comprises the following steps: selecting a plurality of steel balls with the same batch and consistent size, and assembling the steel balls in series in a special clamp along the axial direction to form a uniform stress system; applying an axial extrusion load through a testing machine, and acquiring an image sequence with a high-contrast characteristic mark by using a clamp observation window in cooperation with an industrial camera; adopting a sub-pixel-level feature tracking algorithm to extract the overall deformation of multiple steel balls, and based on the deformation consistency hypothesis of the steel balls connected in series, distributing the total deformation in equal proportion according to the number of the steel balls to obtain the contact deformation of the single steel ball; and the critical load corresponding to one-ten-thousandth-diameter plastic deformation is positioned in combination with load data, and core parameters are provided for determining the rated static load of the bearing. According to the invention, the measurement resolution is improved by 5-15 times through a series amplification effect, the precision reaches 0.01 [mu] m in cooperation with vision measurement, the deformation consistency is ensured by a special clamp, the bottleneck of micro deformation measurement is solved, and the device is adaptive to a conventional test scene of the bearing steel ball, and is low in cost and convenient to operate.

Systems and methods are provided for implementing adaptive cursor control. One system may include a camera to detect an image data stream including a feature of a user. The system may also include an electronic processor to detect, via a feature tracking model that receives the image data, a change in a position of the feature. The electronic processor may also, in response to detecting the change in the position of the feature, determine a control parameter for a cursor displayed to the user. The electronic processor may also control the cursor displayed to the user using the control parameter.

The invention relates to the technical field of mobile robot positioning and environment perception, and discloses a degraded environment positioning method based on point cloud intensity information assistance and related equipment, and the method comprises the steps: mapping a three-dimensional point cloud intensity image into a first-level projection intensity image through a channel, and carrying out the image processing of the projection intensity image, and obtaining a second-level projection intensity image; performing gradient calculation in the secondary projection intensity image and determining intensity gradient features, performing multi-stage screening on the intensity gradient features to obtain intensity features, and performing feature tracking and feature updating on the intensity features in sequence to obtain a new round of map feature points; inputting a new round of map feature points into the map intensity feature point projection model, and outputting to obtain a map feature current frame projection; and inputting the predicted pose and the current frame projection of the map features into a pre-constructed intensity and geometry combined observation model for fusion calculation to obtain optimal pose estimation. According to the invention, the problem of positioning in a degraded environment is solved without the assistance of an external sensor.

In various examples, a trailer angle may be estimated using one or more machine learning models to predict one or more keypoints on the center axis of the trailer drawbar (e.g., a keypoint representing the drawbar junction around which the drawbar pivots, one or more other keypoints along the center axis), back-projecting the predicted keypoint(s) onto a three-dimensional (3D) representation of the ground, and calculating the angle between the longitudinal axis of the towing vehicle and a line or ray formed by or fitted to the projected keypoints. The trailer angle may be estimated at any frame rate. For each frame, keypoints may be predicted from that frame and / or optical flow or some other type of feature tracking may be used to propagate predicted keypoint(s) from a preceding frame in lieu of predicting keypoint(s), and the resulting keypoint(s) may be used to estimate the trailer angle for that frame.

Aspects of this technical solution can increase processing speed and reduce hardware complexity in motion detection, while maintaining the integrity of motion detection across multiple frames. For example, in image processing environments associated with autonomous or semi-autonomous navigation (e.g., driving, robot navigation, etc.), large amounts of image data must be processed quickly and accurately to obtain reliable and up-to-date models of a physical environment. Thus, embodiments according to this disclosure can provide fast and accurate motion detection of input image data.