165 results about "Optical flow computation" patented technology

The invention discloses a target automatically recognizing and tracking method based on affine invariant points and optical flow calculation, which comprises the following steps: firstly, carrying out image pretreatment on a target image and video frames and extracting affine invariant feature points; then, carrying out feature point matching, eliminating mismatching points; determining the target recognition success when the feature point matching pairs reach certain number and affine conversion matrixes can be generated; then, utilizing the affine invariant points collected in the former step for feature optical flow calculation to realize the real-time target tracking; and immediately returning to the first step for carrying out the target recognition again if the tracking of middle targets fails. The feature point operator used by the invention belongs to an image local feature description operator which is based on the metric space and maintains the unchanged image zooming and rotation or even affine conversion. In addition, the adopted optical flow calculation method has the advantages of small calculation amount and high accuracy, and can realize the real-time tracking. The invention is widely applied to the fields of video monitoring, image searching, computer aided driving systems, robots and the like.

The invention relates to a method for detecting vehicle motion information based on integration of binocular stereoscopic vision and optical flow, and the method comprises the following steps of: analyzing and processing left and right image sequences taken by a binocular stereoscopic camera, and acquiring motion information of the three-dimensional space velocity and angle velocity of the car body through feature matching, three-dimensional reconstruction, feature extraction and optical flow estimation, in order to estimate motion information of the car body. As a new method for detecting car motion parameters, it is simply installed, applied to the actual operating environment of the car, high in precision and strong in real time. By the method, inaccurate reading numbers and measuring errors occurred by environmental factors can be effectively reduced, and the visual sensor has the advantages of simple structure and much information and can provide a mass of other information. The system selects ground points as interested feature points to effectively reduce interference of the motion object and improve estimation precision. At the same time, by the method, steadiness and robustness of the algorithm can be enhanced by fully utilizing depth information.

The invention discloses a human body motion feature extraction method based on a global remarkable edge area, comprising steps of using a contrast between an area and a whole image to calculate the significance, reducing the color quantity of the color space, smoothing the significance of the color space, calculating the significance area according to the space relation of the neighboring areas, performing morphology gradient changing on the foreground area segmented by a binarized threshold to generate a global remarkable edge area, traversing strong corner points of all grids of the video frames under various sizes, collecting key characteristic points, the light stream amplitude value of which is not 0, in the remarkable edge area, solving the displacement of the strong corner point according to the corrected light stream field, and forming the human body motion local time space characteristic by using the strong corner point continuous multi-frame displacement locus and the neighbourhood gradient vector. The invention extracts the motion characteristics through global remarkable edge area, eliminates the background noise points irrelevant to the human body motion, removes the affect on the light stream calculation by the camera motion, improves the accuracy of the human body motion local time space characteristic description and improves the human body motion recognition rate.

A motion estimation method based on the angle-point matching and the optical flow method under violent light change, relating to the computer vision technology field, comprises the following steps of: first, carrying out angle-point detection to a present frame; next, carrying out normalized angle-point matching of the present frame and a last frame rotation-invariantly; then, dividing the matched frame into blocks and calculating the affine transformation parameter in each block, carrying out block global motion vector estimating of the present frame according to the affine transformation parameter and performing motion compensation of the last frame using the vector estimation; again, carrying out block linear illumination compensation of the last frame after the motion compensation; finally, estimating the global motion vector of the next frame via the optical flow computation of the linear illumination compensated image and the present image frame. With overcoming the advantages of the invalidation of the optical flow method caused by bad conditions, violent light change and great rotation, etc, the invention obtains higher accuracy.

In techniques for video denoising using optical flow, image frames of video content include noise that corrupts the video content. A reference frame is selected, and matching patches to an image patch in the reference frame are determined from within the reference frame. A noise estimate is computed for previous and subsequent image frames relative to the reference frame. The noise estimate for an image frame is computed based on optical flow, and is usable to determine a contribution of similar motion patches to denoise the image patch in the reference frame. The similar motion patches from the previous and subsequent image frames that correspond to the image patch in the reference frame are determined based on the optical flow computations. The image patch is denoised based on an average of the matching patches from reference frame and the similar motion patches determined from the previous and subsequent image frames.

The invention discloses a method for tracking deformable hand gesture based on video streaming, comprising the steps of: obtaining a frame image, and extracting a sub-image containing a human hand from the obtained frame image; selecting feature tracking points from the sub-image containing the human hand, and initializing a continuously self-adaptive mean shift tracker by the sub-image containing the human hand; performing optical flow calculation on the selected feature tracking points to serve as a local tracking result, and synchronously overall tracking the human hand by the continuouslyself-adaptive mean shift tracker to obtain a global tracking result; updating the feature tracking points; and adopting the result of the optical flow tracking as the final output result of the deformable hand gesture. The method for tracking the deformable hand gesture based on video streaming can be used for tracking the human hand with randomly deformable hand gesture and enabling human-computer gesture interaction to operate in a more comfortable manner. According to the invention, the tracking can be performed aiming at the randomly deformable hand gesture, the interference from change of a background and a large area of complexion is eliminated, and the real-time robust hand gesture tracking is achieved.

Apparatus and methods for estimating the optical flow velocity components using an upwind discretization are presented. In prior art, the knowledge of the signs of the optical flow velocity components is needed for using upwind discretization. But in optical flow computations the velocity components are precisely the unknowns and thus their sign information is not available. This invention discloses apparatus and methods in which upwinding is done based on the sign of the local time derivative instead of the signs of the optical flow velocity components.

The present invention discloses an optical flow computation method using a time domain visual sensor, and specific implementation steps are given out. Different from a current method for performing optical flow computation by adopting "a frame sequence image (video)", the method disclosed by the present invention uses a visual information acquisition device-time domain visual sensor to perform optical flow computation of a field of view. Furthermore, because each pixel autonomously performs change detection and asynchronous output, motion and change of an object in a scenario can be sensed in real time, a continuity assumption of "no change in brightness and no change in local speed" in a differential optical flow computation method is well satisfied, and the real-time property of optical flow computation is greatly improved while the precision of optical flow computation is significantly improved. Therefore, the method is very suitable for optical flow computation and subsequent tracking and speed measurement of the high-speed moving object.

An aspect of the present invention provides a moving obstacle detecting device installed in or on a moving body for detecting moving obstacles in the vicinity of the moving body, the moving obstacle detecting device that includes an image capturing unit configured and arranged to capture an image of the surroundings of the moving body, an optical flow calculating unit configured to find an optical flow corresponding to characteristic points in the image captured with the camera, a focus of expansion calculating unit configured to calculate focus or foci of expansion of the motion vectors in the optical flow found by the optical flow calculating unit, a histogram creating unit configured to create a histogram of the distribution of the focus or foci of expansion calculated by the focus of expansion calculating unit, a peak test unit configured to detect at least one moving obstacle to test peaks in the histogram, the test determining whether or not the peaks correspond to the same object if two or more peaks exist in the histogram created by the histogram creating unit.

The invention discloses a pedestrian counting method based on multiple features. A background lattice model and a capture prospect lattice are set, so the position of the light spot of the background lattice model and the position of the light spot of the capture prospect lattice are compared, a prospect light spot offset model is set, then a light spot offset local maximum value point area in the prospect light spot offset model is extracted to serve as a head area of a pedestrian and is clustered to a head block mass, a motion trail is traced finally, and the number of people is determined and recorded based on motion features. The pedestrian counting method is based on light stream computing, renewal of a pedestrian individual track is achieved by means of intra-frame block mass bilateral matching, and statistics of the number of people is accurate and reliable. The pedestrian counting method is further based on sparse depth data, and improves the execution speed of an algorithm compared with a method based on a dense depth. At last, relative depth acquisition equipment adopted in the pedestrian counting method is low in cost, so the pedestrian counting method has a greater advantage in cost.

The invention relates to an abnormal behavior identification method in an error BP Adaboost network based on video motion information feature extraction and an adaptive boost algorithm. The method comprises: light streams are calculated according to adjacent image frames of a video; a light stream direction is calculated by light streams at a horizontal direction and a vertical direction, a light stream direction histogram is calculated by using the intensity of the light stream as a weight, and then a histogram feature is converted into a feature attribute having a probability attribute; an error BP Adaboost network based on an adaptive boost algorithm is trained based on normal and abnormal training samples, thereby obtaining a classifier; at a testing stage, before the classification model obtained by training is used, a light stream direction histogram of a testing sample is obtained according to a calculation method with the same light stream histogram of the adjacent frames; and then abnormal behavior identification in the testing sample is carried out according to the classification model obtained by training and learning. According to the invention, the method has characteristics of high identification rate and low calculation complexity and can be widely applied to abnormal behavior identification and action analysis fields.

An autonomous navigation, guidance, and control process for docking and formation flying by utilizing the laser dynamic range imager (LDRI) and key technologies, including fuzzy logic based guidance and control, optical flow calculation (including cross-correlation, phase-correlation, and image differential), software design and implementation, and verification methods. The autonomous navigation, guidance, and control process includes the steps of providing an expected reference position relative to a target; generating a carrier position and attitude relative to said target by a Range and Intensity Images Provider; producing a relative position and attitude error; and producing control commands from said relative position and attitude error for relative motion dynamics.

The invention relates to the field of video processing, and discloses a method and a device for detecting a strenuous motion in a video. According to the method and the device, a motion corner point light stream in a video frame is detected and calculated; a motion entropy is calculated according to the motion corner point light stream; and whether the strenuous motion exists is judged. By adopting a motion corner point, a motion can be effectively captured without being influenced by a corner point on the background, and noise interference can be reduced; furthermore, only the motion corner point light stream is required to be calculated, and light streams of non-motion pixels are not required to be calculated, so that the calculation amount is greatly reduced; for each motion communication region, all the motion corner point light streams in the motion communication region are counted, so that the motion entropy of the motion communication region can be calculated; and therefore, whether the strenuous motion exists is judged.

Adaptive system and method for interacting with and navigating 2D and 3D audiovisual content by interpreting the side to side and / or up and down physical gestures and sounds of a multi-person audience. The system may be used in a movie theatre, stadium, music arena, or other venue where an audience views the same view screen. A camera pointed at the audience captures audience motion and a microphone captures sounds. Optical flow is determined by comparing successive video frames. A Motion Index between −100 and 100 is calculated using Optical flow, comparing more recent Optical Flow vectors with previous values. Methods are provided to give equal weight to audience members closer to and further from the camera. Motion Index Values are used to control and interact with content scenarios.

The invention discloses a template matching and optical flow method-based detection method for a power transmission line foreign matter of an unmanned aerial vehicle image, and belongs to the technical field of picture processing. The method comprises the steps of power transmission line video image collection, linear edge contrast enhancement, mean filtering, edge straight line feature extraction, power transmission line region setting, template matching, optical flow calculation, optical flow grayscale image binarization, feature point detection tracking and foreign matter identification, and the like. According to a feature that a power transmission line is approximate to a horizontal straight line, the linear edge contrast enhancement method and the edge straight line feature extraction method are proposed, so that an edge straight line of the power transmission line can be accurately extracted, and the adaptability and the edge straight line feature extraction are strong. According to a principle of large moving displacement of corresponding pixel points of an object relatively close to a camera in the image, a foreign matter region determination method is proposed by integrating template matching, optical flow calculation and image binarization, so that the accuracy is high and the robustness is strong. A foreign matter is tracked and identified through feature point matching of the foreign matter, so that the accuracy is high and the adaptability is strong.

The invention discloses a light stream calculation system and a method. The light stream calculation system comprises a characteristic extraction module, a characteristic matching module and a characteristic monitoring module, wherein the characteristic extraction module sequentially calculates the characteristic values corresponding to the pixel points of an input image; the characteristic matching module sequentially calculates the light stream information corresponding to the pixel points of the input image; and the characteristic monitoring module synchronizes the characteristic values corresponding to the pixel points and the light stream information corresponding to the pixel points, and determines whether to output the light stream information of the pixel points based on the characteristic value corresponding to a pixel point. In the invention, a stream type hardware calculation mode is adopted to realize characteristic matching and characteristic extraction of the pixel points of the input image; following the input of each pixel point of the image, the characteristic value and light stream information corresponding to each pixel point can be finished in a stream manner; and therefore, dense light stream information can be provided. Moreover, due to the indication of the characteristic value, only the light stream of the pixel points with good characteristics can be output, and thus the real motion conditions of the target can be reflected relatively well.

The invention in particular relates to a street traffic volume statistics and sped measurement method based on binocular stereo vision. The method comprises the following steps: (1) performing gray-scale treatment on the image sequence acquired by a left camera of a binocular camera, and performing optical flow computation on the processed image sequence; performing connected component analysis on the previous optical flow computation result, filtering false targets by virtue of area characteristics, thereby obtaining the area and centroid coordinates of the moving vehicles; counting the street traffic volume in the mode of setting the statistical area and comparing the centroid coordinates of the vehicles; and converting the centroid coordinates of the vehicles to a camera coordinate system by utilizing the parallax error information of the moving vehicles and the internal and external parameters of the binocular camera, and calculating the vehicle speed. The method disclosed by the invention is suitable for statistics and speed measurement of multiple vehicles in the street and has the characteristics of measurement accuracy, high reliability and low implementation cost.

The invention discloses a multi-scale feature optical flow learning calculation method based on a self-attention mechanism, and the method comprises the steps: firstly selecting any two continuous frames of images in an input image sequence, carrying out the pyramid feature extraction of the selected two frames of images, and solving a sequence initial optical flow field; secondly, performing feature fusion on the initial optical flow field and the corresponding features thereof, and respectively capturing attention dependence relationships by superposing the fusion features and the features of each layer of the pyramid corresponding to the fusion features and utilizing a self-attention mechanism; after superposition of channel layers is carried out, carrying out feature extraction calculation to solve a residual optical flow field; therefore, the optical flow calculation precision of the model at the image boundary or the moving edge in the large-displacement motion state is further improved. The boundary blur phenomenon caused by large displacement motion in image sequence optical flow calculation is improved, and the method has higher calculation precision and better applicability for complex scenes and large displacement image sequences.

A technique for estimating the optical flow between images of a scene and a segmentation of the images is presented. This involves first establishing an initial segmentation of the images and an initial optical flow estimate for each segment of each images and its neighboring image or images. A refined optical flow estimate is computed for each segment of each image from the initial segmentation of that image and the initial optical flow of the segments of that image. Next, the segmentation of each image is refined from the last-computed optical flow estimates for each segment of the image. This process can continue in an iterative manner by further refining the optical flow estimates for the images using their respective last-computed segmentation, followed by further refining the segmentation of each image using their respective last-computed optical flow estimates, until a prescribed number of iterations have been completed.

Point matches between images within an image sequence are identified by sparse optical flow computation and employed to compute a fundamental matrix for the epipolar geometry, which in turn is employed to derive an epipolar geometry constraint for computing dense optical flow for the image sequence. The epipolar geometry constraint may further be combined with local, heuristic constraints or robust statistical methods. Improvements in both accuracy and performance in computing optical flow are achieved utilizing the epipolar geometry constraint.

The invention relates to the field of indoor visual positioning and navigation and in particular relates to a TOF (Time of Flight)-based optical flow positioning system. The system comprises an image and inertia data acquisition module, an optical flow calculation module, a depth estimation module, a motion estimation module and an output module. According to the system disclosed by the invention, the image information as well as acceleration and angular velocity are acquired by the image and inertia data acquisition module, characteristic point information and depth estimation information of the image can be acquired by the optical flow calculation module and the depth estimation module, and then the motion estimation module combines the characteristic point information and depth estimation information and calculates to obtain the final positioning position, so that accurate and rapid three-dimensional positioning is realized.

System and method for visually tracking a target object silhouette in a plurality of video frames under very general conditions. The tracker does not make any assumption about the object or the scene. The tracker works by approximating, in each frame, a PDF (probability distribution function) of the target's bitmap and then estimating the maximum a posteriori bitmap. The PDF is marginalized over all possible motions per pixel, thus avoiding the stage in which optical flow is determined. This is an advantage over other general-context trackers that do not use the motion cue at all or rely on the error-prone calculation of optical flow. Using a Gibbs distribution with a first order neighborhood system yields a bitmap PDF whose maximization may be transformed into that of a quadratic pseudo-Boolean function, the maximum of which is approximated via a reduction to a maximum-flow problem.

The invention discloses a robust optical flow field estimating method based on a TV-L1 variation model. The robust optical flow field estimating method comprises the following steps: firstly, performing structural texture resolution on an input image, and establishing an optical flow calculating model based on the TV-L1; secondly, establishing an image pyramid, calculating optical flow on the lowest image resolution layer by a discretized alternating iteration method, further calculating with a calculated value as an initial value of a next higher resolution layer till the highest resolution layer, namely the original image resolution, and accelerating the algorithm by using a GPU (graphic processing unit) so as to improve the instantaneity of the algorithm; finally, calculating the error of the algorithm by using an optical flow error evaluating function. In the robust optical flow field estimating method, the input image is processed by a structural texture resolving method and a texture image is applied to optical flow calculation, so that influence of an image shadow caused by illumination variation on the calculation is avoided; by the robust optical flow field estimating method based on the TV-L1 variation model, segmenting smoothness of the image is kept and the optical flow calculating precision and optical flow calculating speed are improved.

A parallax calculator sets a plurality of target points in each standard image and calculates parallaxes of the individual target points. An optical flow calculator calculates two-dimensional optical flows of the individual target points by searching for points corresponding to the target points set in the standard image captured in a frame immediately preceding a current frame. A movement calculation block calculates a forward moving speed of the moving body and angular velocities thereof in both a pitch direction and a pan direction.

The invention relates to a method for quantitatively analyzing a myocardium acoustic contrast image, which belongs to the technical field of medical image analysis and processing. The method comprises the following steps: determining an energy equation in optical flow calculation; constructing pyramid expressions of each frame of a myocardium acoustic contrast image sequence; calculating an optical flow field of the image sequence by adopting a searching policy from roughness to precision; selecting an ROI area of a myocardium segment; determining displacements of the selected ROI area of the myocardium segment in each frame of the myocardium acoustic contrast image sequence; determining positions of the selected ROI area of the myocardium segment in each frame of the myocardium acoustic contrast image sequence and measuring dB values of sampling points of the positions; determining unknown parameters in a quantitative analysis mathematical model of the myocardium acoustic contrast; and drawing a time-intensity curve of the myocardium acoustic contrast. The method achieves the full automation of the quantitative analysis of the myocardium acoustic contrast image and has accurate analysis result and high clinical effectiveness and objectivity.

A method, device, and apparatus for determining optical flow from a plurality of images is described and includes receiving a first image frame from a first plurality of images, where the first plurality of images have a first resolution and a first frame rate. A second image frame may be received from a second plurality of images, where the second plurality of images have a second resolution less than the first resolution and a second frame rate greater than the first frame rate. A first optical flow may be computed from the first image frame to the second image frame. Additionally, the based at least in part on the first optical flow from the first image frame to the second image frame, a third image frame may be output as part of an output stream. The output stream may have a frame rate greater than or equal to the first frame rate, where the third image frame has a resolution greater than or equal to the second resolution.

The position detecting device includes a travel speed detecting unit, an optical flow calculating unit, a distance calculating unit, and a position detecting unit. The travel speed detecting unit detects a travel speed of a mobile object moving on an arbitrary route. The optical flow calculating unit calculates an optical flow of one side of the mobile object, based on a series of frame images of the one side. The one side is any one of a left side and a right side of the mobile object. The distance calculating unit calculates a distance to a subject on the one side based on the travel speed and the optical flow. The position detecting unit detects a current position of the mobile object based on point data of the subject and the distance to the subject.