99 results about "Optical flow computation" patented technology

The invention provides an in-tunnel inrush water pressure monitoring method and system based on dual-acquisition multi-modal data, and belongs to the technical field of inrush water pressure monitoring. Comprising the steps of collecting video data by adopting a high-definition camera; sequentially performing water flow area intelligent segmentation, optical flow calculation and ConvLSTM time sequence prediction on the video data to predict the water flow velocity; an infrared camera is adopted to collect temperature data; self-adaptive temperature gradient calculation and temperature anomaly detection are sequentially carried out on the temperature data to predict the water flow temperature; aligning and fusing the predicted data of the water flow velocity and the water flow temperature, and constructing a water pressure calculation model to obtain a water pressure predicted value; and determining an early warning level according to the water pressure predicted value so as to adjust the monitoring mode of the inrush water pressure in the tunnel. According to the method, the infrared technology and the vision field technology are combined, accurate prediction of the inrush water pressure in the tunnel can be achieved, and a corresponding early warning monitoring mode is provided according to the prediction result.

The invention discloses a multi-frame photoacoustic image reconstruction method based on optical flow alignment and depth feature fusion. The method comprises the following steps: S1, obtaining photoacoustic signal data; s2, reconstructing a photoacoustic cross-sectional image; s3, performing optical flow calculation and image alignment; and S4, training the deep feature fusion network. According to the method, through optical flow motion correction and a potential space learning mechanism, space-time information and complementary features in the aligned multiple frames of images are dynamically integrated, complementary information in the multiple frames of images is adaptively fused, noise is suppressed, and finally reconstruction of high signal-to-noise ratio and high spatial resolution images of biological tissues is achieved. Experimental results show that the method can significantly improve image quality, recover image distortion and detail loss caused by motion and noise, and provide a new effective scheme for promoting robust clinical application of a photoacoustic imaging technology.

The invention discloses a high-efficiency time sequence optical flow method, system and device for fusing event and image information and a medium. The method comprises the following steps: constructing a time sequence feature extraction module; constructing a double-branch collaborative iteration optical flow module of the event and the image; constructing a fusion modeling and iterative optimization optical flow module based on cross attention; training a dual-branch collaborative iteration optical flow module of the event and the image and a fusion modeling and iterative optimization optical flow module based on cross attention on a training set of an MVSEC public data set to obtain weights, and testing on a test set of the MVSEC public data set to obtain an optical flow calculation result; the system, the equipment and the medium are used for implementing the method. The method has the advantages of high time resolution, high dynamic modeling capability, fine multi-modal fusion, high estimation precision and the like, the motion sensing performance in a complex scene is remarkably improved, and powerful technical support is provided for an intelligent visual system.

The invention relates to the field of image processing, in particular to an industrial feeding anomaly recognition method based on visual flow detection, and the method comprises the steps: obtaining an image sequence for optical flow calculation through the collection and preprocessing of an image sequence of an industrial feeding port; performing nonlinear fusion on the pixel brightness value and the time brightness change rate to obtain a highlight suppression factor; carrying out distribution analysis on the local gradient structure tensor characteristic value to obtain a gradient structure confidence factor; performing joint weighting on the highlight suppression factor and the gradient structure confidence factor to obtain a dynamic space-time confidence weight; weighted optical flow calculation and variance threshold judgment are carried out on the dynamic space-time confidence coefficient weight, a feeding abnormal state recognition result is obtained, and therefore the problems of optical flow calculation distortion and abnormal misinformation caused by mirror surface highlight in the SMT micro component feeding process are solved.

The invention provides a real-time processing method for endoscopic image blood vessel enhancement and microcirculation evaluation, which comprises the following steps of: processing an original endoscopic image, and locally enhancing a brightness channel through a contrast-limited adaptive histogram equalization technology; self-adaptive frequency domain-space domain decomposition is realized based on local texture complexity analysis; performing multi-scale Hessian matrix blood vessel detection on the low-frequency component; applying a directional Gabor filter bank to the high-frequency component; spatial-temporal feature fusion is realized through multi-resolution pyramid optical flow calculation; synchronously completing blood vessel probability prediction, blood vessel diameter estimation and blood flow direction prediction by using a lightweight multi-task deep learning network; the blood flow velocity is analyzed and calculated based on a speckle mode, the perfusion density is subjected to accelerated statistics through an integrogram, and vascular morphological parameters are extracted by adopting an improved skeleton algorithm. According to the invention, an enhanced blood vessel visualization effect and a real-time microcirculation quantitative evaluation function can be provided, and the overall improvement of the endoscope image processing quality and efficiency is realized.

The invention discloses a dual-light imaging automatic registration method of a near-infrared and visible light shared sensor, relates to the field of dual-light imaging image registration, and aims to solve the problem of difficult dual-light camera image registration in the prior art. An infrared cut-off filter and a single near-infrared pass filter alternately pass through a lens through a filter circulation switching mechanism, an image formed by interweaving a visible light image and a near-infrared image is shot, optical flow calculation is carried out through the image which is well imaged, pixel motion of an image frame with low imaging quality is estimated, and the image quality is improved. And finally, registering and outputting the high-imaging-quality image and the estimated image. According to the method, accurate matching of the visible light image and the near-infrared image in the aspect of spatial parameters is achieved, meanwhile, a mapping transformation matrix with complex calculation is converted into simple optical flow estimation, rapid, efficient and accurate registration of the visible light image and the near-infrared image is achieved, and therefore a registration image with clear pictures and unified spatial parameters is obtained.

The invention relates to the technical field of image enhancement, in particular to a rhizosphere microflora distribution analysis method based on image enhancement, which comprises the following steps: acquiring a continuous microscopic image sequence of a rhizosphere microenvironment, converting the continuous microscopic image sequence into a gray data space, and constructing a standard input frame set for optical flow calculation. According to the method, the stability score of each feature is calculated by using the signal-to-noise ratio stability measurement function, and the long-life-cycle feature is screened and reserved, so that the random imaging noise of the short life cycle can be filtered fundamentally, and the topological structure of the target community is obtained; and mapping the topological structure back to an original image space by using an inverse persistence mapping algorithm to be fused with a continuous microscopic image sequence, and finally, remarkably enhancing weak biological signals while keeping original form details, thereby realizing high signal-to-noise ratio analysis of microflora spatial and temporal distribution in a complex rhizosphere environment.

The invention discloses a 3D RS and pyramid model fused optical flow calculation method for real-time rendering frame insertion, and the method comprises the steps: 1, obtaining the information of each texture in a rendering frame, so as to construct a texture pyramid; step 2, selecting a grey-scale map corresponding to the lowest resolution to estimate an optical flow initial value; step 3, generating an optimal up-sampling optical flow according to the optical flow initial value, and updating the optimal up-sampling optical flow in a neighborhood to obtain an updated optical flow; 4, repeating the step 3 to obtain the up-sampling optical flow with the highest resolution; 5, performing fine adjustment on the up-sampling optical flow under the highest resolution; step 6, calculating motion vectors of target pixels in the previous rendering frame and the current rendering frame based on the rendering pipeline, and generating an optimal motion vector in combination with the up-sampling optical flow after fine tuning of the target pixel block; and 7, reconstructing a final intermediate frame image. The invention also provides an optical flow calculation system. The method provided by the invention can effectively solve the problem of wrong sampling of real-time rendering frame insertion.

The invention discloses an optical flow estimation method, device, equipment and medium, and the method comprises the steps: extracting the multi-scale features of an image through a pyramid network structure, and obtaining the initial optical flow of each level through feature warping processing and cost body calculation; calculating an uncertainty map of each level of optical flow based on the initial optical flow; calculating the fusion weight of the current layer by adopting the uncertainty map of the current layer and the fused uncertainty map of the previous layer, and carrying out weighted fusion to obtain the fusion optical flow of the current layer; calculating the motion gradient consistency loss based on the optical flow gradient difference of the adjacent pyramid layers; constructing a total loss function based on the endpoint error loss, the uncertainty loss and the motion gradient consistency loss of the initial optical flow; and training the optical flow estimation network by using the total loss function to obtain an optimized optical flow estimation model, and performing optical flow prediction on the to-be-estimated image. According to the invention, the overall precision, robustness and interpretability of optical flow estimation in a complex scene are improved.

The invention discloses a target detection method and system under a dynamic background, and the method comprises the steps: S1, obtaining a time image, and detecting a background region and each target region in the image; s2, sparse optical flow calculation is carried out on the background area and the image of the moment, and the offset of each feature point in the background area is obtained; meanwhile, sparse optical flow calculation is carried out on each target area and the image, and the offset of each target area is obtained; s3, performing sub-grid division on the image, and calculating the average offset of each sub-grid; s4, performing motion compensation on the offset of the target area according to the average offset of the sub-grid where the central point of the target area is located, and separating out the autonomous motion offset of the target; and S5, obtaining a target motion state or / and a target motion speed. According to the method, the target state under the dynamic background can be accurately estimated while the calculation efficiency can be ensured.

The invention is suitable for the technical field of hydrographic survey, and provides a millimeter wave radar speed measurement calibration method based on visual flow measurement, which comprises the following steps: acquiring data, constructing a database and preprocessing an image, performing target detection on the preprocessed image based on YOLO v8, performing region segmentation on a water surface region after target extraction is completed, and obtaining a millimeter wave radar speed measurement calibration result. Visual flow measurement calculation is carried out based on an optical flow method, when optical flow calculation is carried out, calculation should be carried out in each area, space-time synchronization is carried out on a visual flow measurement result and a flow velocity measured by a millimeter wave radar, credible areas are screened, abnormal data are processed, and linear coefficients are fitted for each credible area k through a least square method; according to the method, under the condition that millimeter wave radar speed measurement single-point measurement is limited, area speed measurement information is provided through visual flow measurement, and the problem that the millimeter wave measurement range is insufficient is solved.

The invention relates to the technical field of computer vision and neuromorphic calculation, in particular to a memristor-based motion information extraction and fast optical flow calculation system, which comprises a front-end imaging module, a rear-end imaging module, a memory module and a signal processing module, and is characterized in that the front-end imaging module is used for acquiring a visual scene and converting the visual scene into an analog voltage signal; the voltage conversion module is used for extracting a light intensity change condition based on the analog voltage signal; the neuromorphic module is used for coding time dynamic information of a visual scene according to the light intensity change condition; and the signal processing module is used for extracting a time motion clue based on the time dynamic information and carrying out optical flow calculation on a corresponding region of the time motion clue. According to the invention, efficient motion analysis can be realized, and the optical flow calculation speed is improved.

An image processing method, electronic device and storage medium are provided. The method includes generating luminance gradient data from input frames including a first input frame and a second input frame; obtaining an optical flow vector field by performing multiscale optical flow calculation based on the luminance gradient data; scaling optical flow vectors in the optical flow vector field; and performing warping based on the second input frame and the scaled optical flow vectors to obtain an interpolated or extrapolated frame. This method can improve video contents or stream contents.

The invention relates to the technical field of biomedical engineering. The invention discloses a video-based Moire reflection recognition method and device, terminal equipment and a storage medium, which can automatically recognize Moire reflection and improve the accuracy of Moire reflection recognition. The method comprises the following steps: acquiring a target video of a target object; performing human body part segmentation recognition processing on each frame of image of the target video by adopting an image segmentation module of a Moire reflection recognition model to obtain a target segmentation result corresponding to each human body part in each frame of image; performing optical flow calculation processing on the target video and all target segmentation results by adopting a motion feature module of a Moire reflection recognition model to obtain a plurality of pieces of optical flow time sequence data corresponding to each human body part; and adopting a classification and discrimination module of the Moire reflection recognition model to carry out infant motion classification and recognition processing on all the optical flow time sequence data to obtain a Moire reflection recognition result of the target object.

The invention provides a vision-based auxiliary-mark-free port machinery equipment movement distance detection method and system, and relates to the field of computers.The method comprises the steps that monocular cameras are installed on lifting appliances on the two sides of a gantry crane, the gantry crane is controlled to move along a preset track, and a calibration image sequence containing ground linear features is obtained; a radial distortion parameter of the camera is calculated, a nonlinear mapping model from a pixel coordinate system to a world coordinate system is established, a ground unshielded area is selected as a dynamic monitoring area, and a texture richness thermodynamic diagram is generated; after receiving a measurement instruction of a scheduling system, initializing an optical flow accumulator, loading a current feature point set, synchronously collecting monocular video streams, and distributing the monocular video streams to at least three preprocessing threads to execute differential image enhancement; optical flow calculation is executed on all preprocessing results in parallel, a multi-mode optical flow vector field is generated, and three-level optical flow screening is implemented. The precision and the sensitivity of small-range movement measurement of the gantry crane are improved through an optical flow method, and the influence of environmental factors on a measurement result is reduced.

The mental illness recognition system based on the optical flow features collected by a visual sensor comprises a mental illness expert consultation video data preprocessing module, a facial stress unit extraction module, an optical flow change feature unit construction module, a classification model construction module and a patient mental illness category recognition module connected in sequence. The mental illness expert consultation video data preprocessing module feeds the patient facial picture to the facial stress unit extraction module. The facial stress unit extraction module feeds the optical flow calculation method and the patient facial picture sequence to the optical flow change feature unit construction module. The optical flow change feature unit construction module feeds the optical flow change feature unit to the classification model construction module and the patient mental illness category recognition module respectively. The classification model construction module feeds the classification model to the patient mental illness category recognition module. The mental illness and normal sample classification recognition is realized under the condition that the user facial video has good clarity.

The invention discloses an optical flow calculation method, device and equipment based on comparative learning and a readable storage medium, and relates to the technical field of optical flow calculation. Comprising the following steps: firstly, performing feature extraction on an original image and a target image through a convolutional network to obtain an original feature map of the original image, background features of the original image and a target feature map of the target image; constructing a related body based on the original feature map and the target feature map; the related body is used for recording the pixel matching degree; then attention calculation is carried out on the background features to obtain background attention, and a 2D motion vector is extracted based on a correlation body; and finally, inputting the background attention and the 2D motion vector into an online network to obtain an optical flow calculation result. According to the method, the accuracy of an optical flow calculation result is improved.

The invention relates to the technical field of computer vision, and discloses a competition people stream monitoring and early warning method based on multi-dimensional feature fusion video analysis, which comprises the following steps: acquiring an original video stream, receiving the acquired multi-path high-definition original video stream, and carrying out parallel processing on the original video stream through a first path and a second path respectively; based on the obtained crowd kinetic energy field, crowd entropy field and high-precision density map, constructing a quantitative risk model fusing multi-dimensional features to realize risk decision; and explainable crowd fluid state semantics and risk levels are output to a user through a four-level early warning decision-making mechanism based on a comprehensive risk index given by the quantitative risk model fusing the multi-dimensional features. According to the method, core vision algorithms such as dense optical flow calculation, crowd density estimation of a deep convolutional neural network and crowd behavior modeling based on statistical fluid mechanics are comprehensively utilized, and real-time and quantitative perception and dynamic risk early warning of the microscopic state of dense crowds in scenes such as large stadiums and competition sites are achieved.

The invention belongs to the technical field of target tracking, and particularly relates to a human body gait data extraction method and system in a multi-target scene. The method comprises the following steps: correcting a mahalanobis distance formula for calculating spatial feature similarity in an original DeepSORT algorithm by utilizing a displacement field obtained by calculating dense optical streams of adjacent frames to obtain an improved DeepSORT algorithm, performing target tracking in a multi-target scene by utilizing the improved DeepSORT algorithm, and when a target is temporarily shielded or illumination suddenly changes, performing target tracking by utilizing the improved DeepSORT algorithm. The optical flow field can still speculate the target position through background pixel displacement, correct prediction deviation caused by target shielding or sudden illumination change, avoid tracking target loss or target ID jump, and realize stable tracking of multiple targets of pedestrians in a multi-target complex scene, so that the quality of gait data extraction is improved, and the precision of a gait recognition result is improved.

According to the invention, the processing load during optical flow calculation is reduced. To this end, the information processing apparatus includes an optical flow calculation unit that calculates an optical flow by using pixel signals of each of a plurality of pixel groups having different exposure periods.

The invention relates to the technical field of video image enhancement, and discloses a video image enhancement method and system based on reinforcement learning, and the method comprises the steps: obtaining an original video frame sequence, and decomposing the original video frame sequence into an illumination component and a reflectivity component; performing optical flow calculation on adjacent frames to obtain an optical flow vector field, tracking the spatial position of each pixel point and extracting a brightness value to form a motion trail tube bundle; on the basis of the motion trail tube bundle, abnormal pixel judgment is carried out by utilizing reinforcement learning, and a pixel-level abnormal judgment graph is constructed; performing structure recovery processing on the reflectivity component, and replacing an abnormal pixel according to the pixel-level anomaly judgment graph to obtain an optimized reflectivity component; and calculating a gradient field of the illumination component, carrying out regionalization adjustment on the gradient field according to the brightness distribution of the illumination component to obtain a target gradient field, reconstructing the target illumination component and carrying out synthesis to obtain an enhanced video image. The availability of low-quality monitoring videos is improved, and the visual quality of video images is improved.

This invention relates to a bidirectional unsupervised method and system for calculating optical flow in an event camera. The method includes: acquiring a discrete event stream and converting it into forward and reverse event volumes through discrete sampling; extracting features using an optical flow computation convolutional network and enhancing features using a shift-window attention network; constructing a total loss function by combining bidirectional consistency loss, temporal consistency loss, image consistency loss, and smoothing loss; pre-training the optical flow computation convolutional network in an unsupervised manner using the total loss function, and then performing supervised fine-tuning training on a dataset containing real optical flow labels to obtain the final optical flow computation convolutional network. Based on the bidirectional losses for forward and reverse optical flow computation, unsupervised pre-training is achieved, and fine-tuning training is performed on existing real-world scene datasets, enabling the final optical flow computation convolutional network to achieve high-precision and robust event optical flow calculation.

This application provides a vehicle control method, device, storage medium, and vehicle terminal. The method includes: performing optical flow calculations on a first target image and a second target image using an optical flow algorithm to generate optical flow data; extracting at least one optical flow feature from the optical flow data; inputting the at least one optical flow feature into a deep learning model to generate a recognition result; generating a collision risk assessment based on the recognition result; generating a target driving strategy based on the collision risk assessment; and controlling vehicle driving based on the target driving strategy. This allows for the processing of image frames through optical flow calculations, enabling accurate calculation and evaluation of obstacle movement in continuous image data, effectively identifying and tracking dynamic obstacles in image sequences, providing crucial information for subsequent obstacle avoidance decisions, and quickly determining whether to take obstacle avoidance measures by combining the collision risk assessment of the obstacle and the current vehicle. This timely obstacle avoidance decision-making effectively avoids potential collisions and improves driving safety.

The embodiment of the invention discloses a gesture recognition method and electronic equipment. The method comprises the following steps: performing image acquisition on a user gesture through an image acquisition device to obtain a continuous video sequence; performing gesture region extraction on each image frame in the continuous video sequence to obtain a gesture image, and performing feature extraction on the gesture image to obtain spatial stream features; performing optical flow calculation on two continuous frames of images to obtain optical flow images, and performing feature extraction on the optical flow images to obtain time flow features; and fusing the spatial flow features and the time flow features to obtain fused features, and performing gesture classification on the fused features based on a classifier to determine a gesture recognition result. According to the scheme, the space features and the time features can be subjected to cross-modal correlation fusion, so that the two features are complementary in gesture recognition, the motion features can know that the apparent features pay attention to a dynamic area, the border features can correct misjudgment of the operation features, and the accuracy of gesture recognition is improved.

The application discloses a visual-based unmanned aerial vehicle positioning method and device and a storage medium, which comprises the following steps: collecting high-altitude images through a camera carried by an unmanned aerial vehicle, and preprocessing the high-altitude images to obtain preprocessed high-altitude images; performing feature point selection on the preprocessed high-altitude images to obtain a target feature point array; performing optical flow calculation and displacement vector secondary processing based on the target feature point array to obtain an overall displacement vector of a current frame; performing self-rotation state judgment on the unmanned aerial vehicle; performing emergency stop state judgment on the unmanned aerial vehicle under the condition that the unmanned aerial vehicle is in a non-self-rotation state; and performing yaw correction and accumulation based on the overall displacement vector of the current frame to obtain the current position of the unmanned aerial vehicle. By introducing the self-rotation and emergency stop judgment mechanisms, errors are avoided, and the errors caused by the displacement vectors of the feature points in different regions are corrected, thereby further improving the positioning accuracy of the unmanned aerial vehicle.

The invention provides a jacket foundation concrete pouring real-time monitoring system and method based on an underwater robot. The method comprises the steps that the underwater robot generates an enhanced image through a binocular optical camera in combination with LED light supplement, calculates the ratio of a gap volume to a CAD theoretical volume based on binocular parallax, quantifies plumpness, and meanwhile detects defects through a model; the underwater robot analyzes the uniformity of the slurry through an HSV space, detects the segregation rate of the clustered slurry through edges, and collects temperature, pH and viscosity parameters; the underwater robot locks an overflow port, calculates the pixel flow rate through optical flow, converts the real speed through binocular depth, filters, and sends a pump frequency adjusting instruction to the pump station through a communication protocol according to a threshold value; and after the underwater robot finishes pouring, cruising, performing secondary scanning to verify plumpness and defects, floating, recovering and generating a report. According to the invention, real-time monitoring and comprehensive evaluation of the jacket foundation concrete pouring quality and the surrounding environment state are realized, and the controllability and engineering safety of underwater pouring operation are improved.

The invention relates to a Lucas-Kanade optical flow distortion detection method based on feature extraction. The method comprises the following steps: when a transparent part to be detected is not placed, putting a grid plate in front, and shooting the grid plate by using a camera to obtain an undistorted grid image; carrying out downsampling on the reference grid and the distortion grid at the same time, then extracting feature angular points from the reference grid and the distortion grid respectively, and carrying out deformation field calculation corresponding to the points; and mapping a calculation result to an original image layer, and carrying out secondary optical flow calculation to obtain an accurate deformation field so as to calculate optical flow distortion. The measurement method is simple, easy to operate and high in measurement speed, and solves the problems that the measurement speed is low and the measurement precision is poor when the optical distortion of the heterogeneous composite transparent part is measured.

This invention discloses a child fall monitoring method based on motion capture, belonging to the field of intelligent protection technology. The method includes: acquiring video of a child's activities and performing posture estimation to form a continuous sequence of activity frames; performing dual-channel parallel processing on the continuous activity frame sequence to obtain skeletal motion features and trunk motion features; performing spectral analysis and instantaneous velocity calculation on the skeletal motion features to obtain the velocity of lower limb bones, and simultaneously performing dense optical flow calculation on the trunk motion features to generate trunk optical flow velocity; based on the child's fall risk value, performing real-time posture tracking and surrounding environment scanning of the child's activity state, and collecting motion trajectory data; and performing a hazard classification response based on the motion trajectory data to output an intelligent protection plan. This invention, through a dual-channel parallel processing architecture and a spatiotemporal behavior analysis model, achieves collaborative perception of multi-dimensional motion features and accurate capture of the imbalance process of fall precursors, improving the timeliness of fall warnings.

The invention provides a pedestrian state attribute recognition method and device, a computer equipment set and a storage medium, and relates to the technical field of image processing, and the method comprises the steps: obtaining continuous frames of pedestrian images collected by a target vehicle; performing region-of-interest extraction on the pedestrian image of the current frame to obtain a foot region map of the current frame; determining a plurality of effective grounding points of the current frame from the foot area map of the current frame; according to the plurality of effective grounding points of the current frame, performing optical flow tracking in the adjacent frame to obtain a plurality of tracking feature points in the adjacent frame; according to the plurality of tracking feature points, performing optical flow calculation on the pedestrian image of the current frame and the pedestrian image of the adjacent frame pair to obtain optical flow vector information of the plurality of tracking feature points; and determining the state attribute of the pedestrian according to the optical flow vector information of the plurality of tracking feature points. According to the invention, the accuracy of determining the state attribute of the pedestrian is improved, so that the obstacle avoidance capability of the vehicle is improved.

An add-on module for an injection device including a dispense indicator visible through a viewing window in a device housing and the housing defines a longitudinal axis for the injection device. The add-on module includes an optical flow sensor fixed to the add-on module and configured to track movement of the dispense indicator relative to the housing through the viewing window. A processor programmed to run an algorithm enabling optical flow computation from optical flow sensor data is provided, and the add-on module is adapted to be i) releasably attached to the housing and ii) axially and rotationally fixed to the housing when attached.