1461 results about "Point pair" patented technology

Disclosed in the invention is an improved method of a RGB-D-based simultaneously localization and mapping (SLAM) algorithm. The method comprises two parts: a front-end part and a rear-end part. The front-end part is as follows: feature detection and descriptor extraction, feature matching, motion conversion estimation, and motion conversion optimization. And the rear-end part is as follows: a 6-D motion conversion relation initialization pose graph obtained by the front-end part is used for carrying out closed-loop detection to add a closed-loop constraint condition; a non-linear error function optimization method is used for carrying out pose graph optimization to obtain a global optimal camera pose and a camera motion track; and three-dimensional environment reconstruction is carried out. According to the invention, the feature detection and descriptor extraction are carried out by using an ORB method and feature points with illegal depth information are filtered; bidirectional feature matching is carried out by using a FLANN-based KNN method and a matching result is optimized by using homography matrix conversion; a precise inliners matching point pair is obtained by using an improved RANSAC motion conversion estimation method; and the speed and precision of point cloud registration are improved by using a GICP-based motion conversion optimization method.

The invention provides a remote sensing image registration method of a multi-source sensor, relating to an image processing technology. The remote sensing image registration method comprises the following steps of: respectively carrying out scale-invariant feature transform (SIFT) on a reference image and a registration image, extracting feature points, calculating the nearest Euclidean distances and the nearer Euclidean distances of the feature points in the image to be registered and the reference image, and screening an optimal matching point pair according to a ratio; rejecting error registration points through a random consistency sampling algorithm, and screening an original registration point pair; calculating distribution quality parameters of feature point pairs and selecting effective control point parts with uniform distribution according to a feature point weight coefficient; searching an optimal registration point in control points of the image to be registered according to a mutual information assimilation judging criteria, thus obtaining an optimal registration point pair of the control points; and acquiring a geometric deformation parameter of the image to be registered by polynomial parameter transformation, thus realizing the accurate registration of the image to be registered and the reference image. The remote sensing image registration method provided by the invention has the advantages of high calculation speed and high registration precision, and can meet the registration requirements of a multi-sensor, multi-temporal and multi-view remote sensing image.

The invention relates to a combined three-dimensional digital imaging method of digital speckle projection and phase measuring profilometry. The method comprises the following steps: (1), a random digital speckle pattern is projected onto an object and acquired through two cameras on left and right sides; (2), a sinusoidal fringe pattern is projected onto the object, and a corresponding sinusoidal fringe pattern is acquired; (3), the folded phase value of a coding fringe pattern is calculated; (4), a series of points (Pr1 to Prn) on the right camera corresponding to a point P1 on the left camera are searched according to the folded phase value; (5), the point Pr on the right camera corresponding to the point P1 is fixed in the position of P1 according to the acquired digital speckle pattern; (6), processes of the step (4) and (5) are performed to all points on the left and right cameras, to find all the corresponding point pairs; (7), the corresponding point pairs are re-established into three-dimensional point cloud data. The method has the advantages that the imaging speed is fast, small calculated amount, accurate result, and the method is suitable for the fields of on-line three-dimensional detection and dynamic three-dimensional data acquisition with higher accuracy and speed requirements.

The invention discloses an accurate part positioning method based on binocular microscopy stereo vision, which belongs to the technical field of computer visual measuring and relates to an accurate precision part positioning method based on the binocular microscopy stereo vision. A binocular microscopy stereo vision system is adopted, two CCD (charge coupled device) cameras are adopted to acquire the images of the measured parts, the image information in the to-be-measured area on the measured part is amplified by a stereo microscope, a checkerboard calibrating board is adopted to calibrate the two CCD cameras, and a Harris corner point detecting algorithm and a sub-pixel extracting algorithm are adopted to extract feature points. The extracted feature points are subjected to the primary matching and correcting of matching point pairs, and the feature point image coordinates are inputted to a calibrated system to obtain the space actual coordinates of the feature points. The accurate part positioning method based on the binocular microscopy stereo vision solves the measuring difficult problems generated by the small size of the to-be-measured area, high positioning demand, non-contact and the like. The accurate positioning of the precision part is well finished by adopting the non-contact measuring method of the binocular microscopy stereo vision.

In an image recognition apparatus (1), feature point extraction sections (10a) and (10b) extract feature points from a model image and an object image. Feature quantity retention sections (11a) and (11b) extract a feature quantity for each of the feature points and retain them along with positional information of the feature points. A feature quantity comparison section (12) compares the feature quantities with each other to calculate the similarity or the dissimilarity and generates a candidate-associated feature point pair having a high possibility of correspondence. A model attitude estimation section (13) repeats an operation of projecting an affine transformation parameter determined by three pairs randomly selected from the candidate-associated feature point pair group onto a parameter space. The model attitude estimation section (13) assumes each member in a cluster having the largest number of members formed in the parameter space to be an inlier. The model attitude estimation section (13) finds the affine transformation parameter according to the least squares estimation using the inlier and outputs a model attitude determined by this affine transformation parameter.

The invention provides an SAR image registration method based on SIFT and normalized mutual information. The method includes the steps that firstly, a standard image I1 and an image to be registered I2 are input and are respectively pre-processed; secondly, features of the pre-processed image I1 and features of the pre-processed image I2 are extracted according to the MM-SIFT method to acquire initial feature point pairs Fc and SIFT feature vectors Fv1 and Fv2; thirdly, initial matching is carried out through the Fv1 and the Fv2; fourthly, the Fc is screened for the second time according to the RANSAC strategy of a homography matrix model, final correct matching point pairs Fm are acquired, and a registration parameter pr is worked out according to the least square method; fifthly, I2 is subjected to space conversion through affine transformation, and a roughly-registered image I3 is acquired through interpolation and resampling; sixthly, pr serves as the initial value of normalization information registration, I1 and I2 are subjected to fine registration through the normalized mutual information method, a final registration parameter pr1 is worked out, and a registered image I4 is output. The method can be quickly, effectively and stably carried out, and SAR image registration precision and robustness are improved.

The invention belongs to the crossing field of computer vision and intelligent robots, relates to an indoor three-dimensional scene rebuilding method based on a double-layer rectification method and solves the problems that an existing indoor scene rebuilding method is expensive in required equipment, high in computation complexity and poor in real-time performance. The indoor three-dimensional scene rebuilding method based on the double-layer rectification method comprises Kinect calibration, SURF feature point extraction and matching, mapping from a feature point pair to a three-dimensional space point pair, three dimensional space point double-layer rectification based on random sample consensus (RANSAC) and inductively coupled plasma (ICP) methods and scene updating. According to the indoor three-dimensional scene rebuilding method based on the double-layer rectification method, the Kinect is adopted to obtain environmental data, and the double-layer rectification method is provided based on the RANSAC and the ICP. Indoor three-dimensional scene rebuilding which is economical and rapid is achieved, and the real-time performance of rebuilding algorithm and the rebuilding precision are effectively improved. The indoor three-dimensional scene rebuilding method based on the double-layer rectification method is applicable to the service robot field and other computer vision fields which are relative to the three-dimensional scene rebuilding.

The invention provides a method for barrier perception based on airborne binocular vision. The method comprises the following steps of setting a coordinate system of an airborne binocular vision camera, calculating a formula for conversion between the coordinate system and computer image coordinates of an image obtained by the airborne binocular vision camera according to the coordinate system, wherein the airborne binocular vision camera comprises a left camera and a right camera, extracting characteristic points of an image obtained by the airborne binocular vision camera, carrying out a characteristic vector description process on the characteristic points, carrying out stereo matching of left and right images according to characteristic vectors of the characteristic points to obtain preliminary matching point pairs, eliminating error matching in the preliminary matching point pairs to obtain final matching point pairs, creating a disparity map according to the final matching point pairs, and carrying out barrier perception according to the disparity map. The method for barrier perception based on airborne binocular vision has the advantages of strong adaptability, good instantaneity and good concealment performance.

The invention discloses a depth camera-based visual mileometer design method. The method comprises the following steps of acquiring the color image information and the depth image information in the environment by a depth camera; extracting feature points in an initial key frame and in all the rest image frames; tracking the position of each feature point in the current frame based on the optical flow method so as to find out feature point pairs; according to the number of actual feature points and the region size of the overlapped regions of feature points in two successive frames, selectively adopting the sparse direct method or the feature point method to figure out relative positions and postures between two frames; based on the depth information of a depth image, figuring out the 3D point coordinates of the feature points of the key frame in a world coordinate system based on the combination of relative positions and postures between two frames; conducting the point cloud splicing on the key frame during another process, and constructing a map. The method combines the sparse direct method and the feature point method, so that the real-time performance and the robustness of the visual mileometer are improved.

The invention relates to an unmanned aerial vehicle aerial photography sequence image-based slope three-dimension reconstruction method. The method includes the following steps that: feature region matching and feature point pair extraction are performed on un-calibrated unmanned aerial vehicle multi-view aerial photography sequence images through adopting a feature matching-based algorithm; the geometric structure of a slope and the motion parameters of a camera are calculated through adopting bundle adjustment structure from motion and based on disorder matching feature points, and therefore, a sparse slope three-dimensional point cloud model can be obtained; the sparse slope three-dimensional point cloud model is processed through adopting a patch-based multi-view stereo vision algorithm, so that the sparse slope three-dimensional point cloud model can be diffused to a dense slope three-dimensional point cloud model; and the surface mesh of the slope is reconstructed through adopting Poisson reconstruction algorithm, and the texture information of the surface of the slop is mapped onto a mesh model, and therefore, a vivid three-dimensional slope model with high resolution can be constructed. The unmanned aerial vehicle aerial photography sequence image-based slope three-dimension reconstruction method of the invention has the advantages of low cost, flexibility, portability, high imaging resolution, short operating period, suitability for survey of high-risk areas and the like. With the method adopted, the application of low-altitude photogrammetry and computer vision technology to the geological engineering disaster prevention and reduction field can be greatly prompted.

A fast geometric correction method for a satellite remote sensing image and based on a control point image database comprises the following steps: the geographical scope of an image to be corrected is determined; all control points which meet requirements are retrieved in the control point image database according to the determined geographical scope of the image to be corrected; automatic matching and point selection are conducted, and control point pairs used for geometric correction are determined; whether the number and the distribution of the matched control point pairs meet the requirements of the geometric correction is judged, if the judging result is yes, the next step is conducted, and if the judging result is no, the retrieval step is conducted again; a geometric correction model is selected according to the type and imaging features of a sensor; a transformational relation between a pixel coordinate and a geodetic coordinate is established based on the matched control point pairs and the selected geometric correction model; and the geometric correction is conducted based on a digital elevation model and the transformational relation so as to obtain a corrected digital orthoimage. According to the fast geometric correction method for the satellite remote sensing image and based on the control point image database, the automatic image matching technique is introduced into automatic selection of the control points, and automatic geometric correction processing of the satellite remote sensing image can be realized.

The invention discloses a non-overlapping field-of-view camera gesture calibration method based on point cloud feature map registration. The method comprises the following steps that: (1) carrying outbasic calibration on a plurality of cameras of a non-overlapping field of view to obtain an internal reference; (2) utilizing the plurality of cameras to carry out environment detection and synchronous positioning and mapping, constructing a point cloud map, and extracting a key frame to solve the pose matrix of the camera; (3) extracting an image frame from the key frame of one camera, carryingout similarity detection on the key frames of other cameras, constructing a matching frame point set and a matching point pair set, and carrying out projection error minimization on the projection ofthe point cloud map point on the image frame and the practical pixel coordinate; and (4) carrying out feature matching on a frame near the matched frame, blending all feature points, carrying out global optimization and iterative solution on a relative pose matrix, selecting a correction parameter according to a practical situation, and carrying out final gesture calibration on the camera. By useof the method, the problems of high calibration work intensity, low work efficiency and low accuracy of a traditional calibration method are solved.

Disclosed are embodiments of systems and methods to stitch two or more images together into a composite image. By finding matching point pairs for a pair of images, a homography transform may be obtained for the pair of images. The homography transform may be used to generate a composite image of the image pair. In an embodiment, the process of identifying a homography transform may be iterated. In an embodiment, when forming the composite image, the transformed foreground regions may be selected such that there is no intersection of foreground pixel regions. In an embodiment, foreground pixel regions on the border of an image may be removed. The resulting composite image is a larger image generated from the selected regions from the input images. In embodiments, the process may be repeated for sets of images with more than two images.

A method for stabilizing an image during suspension of a video on an air-based platform comprises the steps as follows: first, selecting a frame image in a video series as a reference frame; extracting characteristic points of a reference image and a current image of the video series by using such a characteristic extraction method as invariant scale and feature transformation; preliminarily matching the characteristics by taking a euclidean space distance as a characteristic match evidence so as to form characteristic match point pairs; further selecting the characteristic match points according to invariability of relative positions of characteristic points of an image background and removing wrong matched characteristic match point pairs and the characteristic match point pairs positioned on a movement target; performing least square calculation by using the characteristic match point pairs in a six-parameter affine transformation module so as to obtain module parameters; and performing correction compensation to the current image so as to obtain stable video series output with fixed visual field. During the process, the invention also provides an idea of changing a new reference frame with an interval of a certain number of frames, thereby reducing errors and improving the stability accuracy; and the invention can be applied to traffic monitoring, target track and other fields and has wide market prospect and application value.

The invention discloses a fast image splicing method based on an improved SURF algorithm. An existing corner extraction method and an existing corner feature description method are improved, mismatching of extracted corners is eliminated, and multiple images can be spliced fast. At first, an improved FAST algorithm is adopted for increasing the extracted corners, the operation speed of the FAST algorithm for extracting the corners is high, and after the improvement is carried out, the stability is good; secondly, the combination of SURF description and LBP description is adopted for describing the corner feature, and in this way, the speed for matching the corners can also be increased; then an RANSAC method is adopted for eliminating mismatching, accuracy is improved, and a more accurate transformation matrix is obtained, so that fast splicing is carried out; finally, according to obtained matching point pairs, parameters for transforming images to be spliced into reference images are calculated, and a slow-in and slow-out method is adopted for finishing image splicing.

The invention provides a three-dimensional reconstruction-based unmanned aerial vehicle image stitching method. The method comprises the following steps of: obtaining unordered images shot by an unmanned aerial vehicle camera in a shooting area; carrying out feature point extraction on the obtained unordered images, carrying out feature matching on the extracted feature points to obtain matched feature point pairs; and sorting the unordered images according to the matched feature points and a sequence; analyzing the matched feature points so as to obtain camera attitude parameters, distortionparameters and space coordinates of the feature points; carrying out image clustering on the arranged images according to the camera attitude parameters and the space coordinates of the feature pointsso as to form a plurality of image clusters, respectively extracting dense point clouds of the image clusters, and completing dense matching; and carrying out ortho-rectification on the dense matchedimages according to the dense point clouds, the camera attitude parameters and the distortion parameters, and stitching the ortho-rectified images to obtain a stitched image. According to the method,the stitching speed is improved and projection distortion is not caused. The invention furthermore provides an unmanned aerial vehicle image stitching system.

The present invention discloses a method for registering synthetic aperture radar image with a change area based on a point pair constraint and Delaunay, and the problem of SAR image registration with a change area without using ground control point data is mainly solved. The method comprises a step of filtering inputted floating image and reference image respectively and then using a SURF algorithm to extract feature points, a step of carrying out similarity measure matching of a normalized descriptor matrix on obtained reference image feature point and floating image feature point and then carrying out distance and wiring direction angle constraint to obtain an initial candidate matching point pair set, a step of using matching feature points in the floating image and the reference image to construct Delaunay, searching a homonymous triangle pairs in the Delaunay of the two images to update the candidate matching point pair set, and taking non-repeated vertices in all homonymous triangle pairs corresponding to a minimum affine error as a final matching point pair set, and calculating an affine change matrix, carrying out affine conversion and bicubic interpolation on the floating image, and obtaining a final registration result image.

The invention proposes a reconstruction method and system for processing three-dimensional point cloud containing a main plane scene. The method comprises the following steps of obtaining a multi-angle image of a static scene by using a camera with known internal parameters; detecting characteristic points of the image, and matching characteristic points of any two images to obtain a matched point pairs and obtaining a matched point sequence by projecting the same scene point; for image pairs containing the preset number of matched point pairs, obtaining a basic array between the image pairs according to the matched points, and storing corresponding space plane point sets; determining the corresponding position relationship between the image pairs according to the basic array; realizing camera fusion and three-dimensional point reconstruction in a standard coordinate frame according to the corresponding position relationship between the image pairs; and optimizing the reconstruction result of the three-dimensional point cloud. The reconstruction method for processing three-dimensional point cloud containing main plane scene of the invention can overcome defects of the existing reconstruction method for processing three-dimensional point cloud and can realize the three-dimensional reconstruction not depending on the scene.

The invention discloses an image matching method comprising the steps of performing global template matching of a template image in a source image, controlling the template image to slide in the source image to find out an optimal matching region; calculating a feature point and a feature vector of the template image and the source image; based on the feature point and the feature vector, calculating the visual similarity between the optimal matching region and the template image; if the visual similarity is zero, determining a mismatch between the optimal matching region and the template image; if the visual similarity is not zero, obtaining feature matching point pairs between the template image and the source image; and calculating a positioning coordinate of the optimal matching image based on the feature matching point pairs. The invention further provides a mobile phone application test platform based on the image matching. The invention can reduce the complexity of a matching algorithm, thereby improving the image matching efficiency and the mobile phone application test efficiency based on the image matching.

A method of aligning a pair of images with (101) a first image and a second image, wherein said images comprise a plurality of intensities corresponding to a domain of points in a D-dimensional space includes identifying (102) feature points on both images using the same criteria, computing (103) a feature vector for each feature point, measuring a feature dissimilarity (104) for each pair of feature vectors, wherein a first feature vector of each pair is associated with a first feature point on the first image, and a second feature vector of each pair is associated with a second feature point on the second image. A correspondence mapping (105) for each pair of feature points is determined using the feature dissimilarity associated with each feature point pair, and an image transformation (106) is defined to align (108) the second image with the first image using one or more pairs of feature points that are least dissimilar.

The invention provides a method for detecting gestures. The method includes steps of detecting a predefined motion mode to determine a region-of-interest to be detected; detecting in the region-of-interest according to a scheme of a multi-scale sliding window, extracting local mean features from skin color membership images based on window images at first, classifying the features through a pre-trained Ada-Boost classifier, further extracting feature points on the basis of a gray-scale image of the window images processed via the Ada-Boost classifier, and then classifying the point-pair features through a random forest classifier; and clustering different types of target gesture windows and outputting accurate positions and shapes of the gestures. The invention further provides a system for detecting the gestures. The method and the system for detecting the gestures are simple, rapid and stable in implementing, and have the advantages of real-time performance, interference resistance, high identification precision and the like.

An image matching method includes setting first and second lattices to first and second images respectively, computing potential force to each second lattice point of the second lattice by a gradient of an image correlation potential energy based on a position of each first lattice point and pixel thereof and a position of the second lattice point and pixel thereof, computing elasticity force to the second lattice point from elasticity energy between the second and adjacent lattice points, computing frictional force occurring at the second lattice point, performing a numerical analysis of an equation of motion regarding the second lattice point and based on the potential force, elasticity force and frictional force to obtain a convergence state of the second lattice points, and adding a new lattice point between an adjacent lattice point pair of second lattice points according to a distance between the adjacent lattice point pair.

The invention discloses a negative obstacle detection method based on the local structure feature of a laser radar point cloud. In collected laser radar point cloud data, three structure features including local point cloud distance hopping, local point cloud dense distribution and local point cloud height decreasing are detected line by line; on the basis of the three structure features, candidate point pairs which may belong to a negative obstacle are obtained through extraction and screening in single line laser point cloud; all candidate point pairs obtained by the laser point cloud of each line laser are clustered on the basis of point pair length consistency and spatial position consistency to obtain a negative obstacle candidate area; and through area filtering and point pair amount filtering, obtaining a negative obstacle area. By use of the method, the negative obstacle in environment can be effectively detected, and the method has the advantages of good detection success rate, small calculation cost and high instantaneity.

The invention discloses a living fingerprint verifying method based on a thin-plate spline deformable model, which collects and preprocesses the fingerprint sequence of a finger for verifying, registers the fingerprints before and after deformation, selects corresponding minutiae point pair to serve as the criterion point pair for estimating the thin-plate spline model, employs improved thin-plate spline model to describe deformation of fingerprints, computes the deformable energy of the fingerprint for verifying in each direction on the basis of model parameter to an overall bending energy vector, trains respectively on a true and a false fingerprint deformation base, obtains space distribution of the bending energy vector, respectively computes similarity from the bending energy vector for measuring to an eigen fuzzy set of true and false bending energies, and selects proper threshold value to obtain the verifying result. The invention can identify false fingerprint from the living fingerprint, effectively realizes living fingerprint verification in real-time, increases safety of automatic fingerprint verifying systems, enlarges the application range of automatic fingerprint verifying systems and has important application value in the field of biological character identification.

The invention provides an image registration system and a method thereof. The method comprises the following steps of: firstly calculating scale invariant feature transform (SIFT transform) of a real-time image and a reference image to obtain feature points of the real-time image and the reference image; selecting a plurality of pairs of feature points with maximum similarity from the feature points of the real-time image and the reference image; carrying out matching verification of the plurality of pairs of feature points with maximum similarity selected from the feature points of the real-time image and the reference image and rejecting a mismatching point pair therein to obtain a plurality of pairs of feature points which are verified; according to the plurality of pairs of feature points which are verified, determining the parameter of an affine transformation function of the real-time image relative to the reference image; and carrying out radioactive transform of the real-time image according to the affine transformation parameter to obtain the registered image. The system and the method are difficult to influence and high in registration accuracy.

The invention relates to an empty bottle wall defect detection method, which comprises the following steps: shooting two images by using a camera before and after the same empty bottle is rotated for 90 degrees during advancing, then transmitting the images to a CPU of an industrial personal computer to perform analysis on the images, and judging whether the bottle wall has defect. Image processing comprises the following steps: A, scanning edge point pairs on a bottle neck, positioning the bottle wall, and dividing a detection area of the bottle wall to perform subarea processing; B, pre-processing image data in the positioned area by adopting a grey stretching method; C, partitioning the images by adopting a maximum between-cluster variance method to acquire target information; and D, performing connectivity analysis on the partitioned bottle wall images, extracting characteristic data of each defect, and judging whether each detected connected domain has real defect according to the mass center position, posture ratio and area characteristic of the connected domain. The invention discloses a detection device used for the detection method at the same time. The detection method and the detection device are easily applied in detection equipment of an industrial flow line so as to realize automatic high-speed accurate detection on the empty bottle wall defect.

The invention relates to an uncalibrated multi-viewpoint image correction method for parallel camera array. The method comprises the steps of: at first, extracting a set of characteristic points in viewpoint images and determining matching point pairs of every two adjacent images; then introducing RANSAC (Random Sample Consensus) algorithm to enhance the matching precision of SIFT (Scale Invariant Feature Transform) characteristic points, and providing a blocking characteristic extraction method to take the fined positional information of the characteristic points as the input in the subsequent correction processes so as to calculate a correction matrix of uncalibrated stereoscopic image pairs; then projecting a plurality of non-coplanar correction planes onto the same common correction plane and calculating the horizontal distance between the adjacent viewpoints on the common correction plane; and finally, adjusting the positions of the viewpoints horizontally until parallaxes are uniform, namely completing the correction. The composite stereoscopic image after the multi-viewpoint uncalibrated correction of the invention has quite strong sense of width and breadth, prominently enhanced stereoscopic effect compared with the image before the correction, and can be applied to front-end signal processing of a great many of 3DTV application devices.

A method of using fisheye image to generate stereoscopic panorama includes shooting out a numbers of fisheye image by fisheye camera according to specified sequence, confirming character point air at overlapped region of adjacent images and setting-up stereoscopic space model as per obtained matching point pair, calculating out position of original image on stereoscopic space and resampling original fisheye image according to calculated out parameter in order to generate out stereoscopic panorama.

The invention provides an urban intersection automatic identification method based on low precision space-time trajectory data, which reduces the cost for obtaining urban intersection and provides basic data and a reference method for future auto-drive and intelligent traffic. The method comprises following steps: first, performing pretreatment to the data according to the spatial density and time interval of the trajectory data; secondly, tracking and identifying turning processes of the trajectory and extracting turning point pairs where the change of course angle exceeds a turning angle threshold value and the time interval is less than a time threshold value in one turning process; then, obtaining clusters and cluster centers of the turning point pairs by means of growing cluster method based on distance and space; finally, classifying the turning point pair clusters based on a cluster method of local point connection and completing intersection automatic identification through analyzing the turning property and the number of center points of the cluster center points. The intersection identification accuracy by means of the method is 94.3%.