67results about How to "Guaranteed registration accuracy" patented technology

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 discloses a boundary feature point registering method for point cloud splicing in a three-dimensional scanning system. The method comprises the following steps that (1) a three-dimensional laser scanner acquires space sampling points on the surfaces of a real object in different view angles; (2) a boundary detecting method of a point cloud gravity center distance feature is used for extracting point cloud boundary feature points in different viewing angles; (3) an improved iterative closest point (ICP) algorithm is used for registering point cloud according to the extracted point cloud boundary feature points; (4) the registering precision is evaluated according to a registering error standard, and whether a registering result meets a registering request or not is verified. According to the boundary feature point registering method for point cloud splicing in the three-dimensional scanning system, boundary feature point extraction is conducted on the point cloud to be registered, the defect that all points in point cloud data need to be traversed to search for the corresponding points in a traditional ICP algorithm is overcome, on the basis that the registering precision is guaranteed, the complexity of the algorithm is effectively lowered, and meanwhile the efficiency of point cloud registering is obviously improved.

A non-rigid registering method of a mouse three-dimensional CT image comprises the following steps: A, a direction and a position of a source image of a mouse three-dimensional CT image are calibrated; B, a threshold segmentation algorithm is used to segment the mouse image to obtain a mouse bone of the source image and a mouse bone of a target image, a center of mass of a bone slice connected domain is extracted to serve as a characteristic point, so that a source image characteristic point and a target image characteristic point are obtained; C, the source image characteristic point and the target image characteristic point in the step B are matched, the matched characteristic point is used to calculate a thin plate batten transformation matrix, and a hierarchical transformation mode is used to carry out the thin plate batten transformation to the mouse source image; and D, the mouse source image after the thin plate batten transformation and the target image carry out a fine registering. According to the invention, when the characteristic points are extracted, the number of characteristic points is reduced; the hierarchical transformation mode is used to overcome the defect of long registering time; an automatic registering in the image registering is basically realized; and registering the three-dimensional mouse image is easier.

The present invention discloses a quick image registration method based on a visual remarkable area, which belongs to the field of image processing. The quick image registration method is used for matching an image A to be matched with a reference image B and respectively acquiring remarkable areas of A and B. The mass centers of three remarkable areas of A are vertexes of a triangle, and the triangle is used as a first characteristic triangle. Similarly a second characteristic triangle is acquired. The first characteristic triangle and the second characteristic triangle which are similar form a similar triangle pair, wherein the similarity of (a,b) is maximal. An affine transformation model for transforming from A to B is established based on (a,b), wherein the related initial matching parameters comprise horizontal translation amount, vertical translation amount, rotation angle and scale parameter. From the initial matching parameter, Powell searching for setting step length is performed for acquiring searching values. Each searching value is used for performing multiple registration tests on the A and B, and the searching value which corresponds with an optical registration test is the optimal matching parameter as a result. The optimal matching parameter is used for performing affine transformation on A for acquiring a final registration result.

The invention discloses a medical image registration algorithm based on a convolutional neural network. The algorithm specifically comprises the following steps: data acquisition and preprocessing; Making a neural network training data set; Training a neural network; Performing coarse registration on the neural network model; And fine registration is carried out based on a gray value registrationalgorithm. According to the invention, the convolutional neural network can be skillfully combined with the traditional gray value-based registration algorithm; The invention discloses a convolutionalneural network elimination method in medical image processing. the complete labeled input sample is lacked; Training sample limitation and other problems, the spatial difference between a group of input fixed images and floating images is quickly calculated; Meanwhile, a neural network model rapidly obtains the spatial difference between the input images to be registered, the large spatial difference is reduced, the situation that a local optimal solution can be obtained in the optimization iteration process based on a gray value registration algorithm is avoided, and therefore the spatial difference between the two images to be registered can be rapidly and accurately calculated.

The invention discloses an automatic registration fusion method of point cloud data and an optical image based on a line feature. The method comprises the steps of filtering point cloud data by means of a mathematical morphology method; determining a depth map of the optical image and the point cloud data by means of an adaptive supporting weight dense three-dimensional algorithm and a Delaunay triangulation algorithm; extracting the line feature of the depth map through a Hough algorithm, and performing coarse matching in a manner that a ratio between an included angle of two lines and a line length of the line feature is used as a similarity measure; eliminating error matching point pairs through a two-step RANSAC algorithm, and obtaining camera position parameter estimation; and performing color texture mapping between the point cloud data and the optical image, thereby obtaining a fused three-dimensional image. The automatic registration fusion method has advantages of no dependence to prior techniques such as GPS and INS, high robustness and high automatic registration degree.

The invention discloses a satellite multispectral image registration method based on object space positioning consistency. The method comprises the following steps: selecting a certain spectrum in a multispectral camera as a reference spectrum, taking the residual spectrums as non-reference spectrums, respectively performing on-track calibration aiming at the relative geometric distortion between each non-reference spectrum and the reference spectrum, and saving the calibration result; establishing a strict geometric imaging model of each spectrum by utilizing the saved calibration result based on the object space positioning consistency, and performing accurate registration on the non-reference spectrums and the reference spectrum. The invention provides an optical satellite multispectral image high-precision automatic registration method of really geometrical significance, the satellite multispectral image can be subjected to high-precision automatic registration without carrying out image matching, the processing efficiency is improved, and the registration quality is irrelevant to the image quality and ground feature types; and meanwhile, a geometric correction model between wave bands is the strict geometric imaging model and has strictness in theory. The practice proves that the method is feasible, effective, stable in registration quality and high in precision.

The invention relates to a solar-altitude-based self-adapting high orbit area-array camera in orbit geometrical calibration method. The solar-altitude-based self-adapting high orbit area-array camera in orbit geometrical calibration method includes the steps: 1) selecting a control point in a control point database, performing automatic measuring of the control point on an image to be calibrated, and acquiring the measuring information of the control point, wherein selection of control point satisfies the condition that at least one control point exists in the 10*10 grid range; 2) constructing an in orbit calibration geometrical model, based on a two-dimensional cubic surface detector directional angle, of a high orbit area-array camera; 3) utilizing the control point information automatically measured in the step 1, utilizing a step-by-step and grading iterative solution strategy to solve the in orbit calibration geometrical model parameter of the two-dimensional cubic surface detector directional angle in the step 2; 4) constructing a solar altitude self-adapting error compensation model and performing parametric solution in different latitude and longitude areas, at different imaging times and at different solar altitudes; and 5) taking the in orbit calibration geometrical model parameter and the solar altitude self-adapting error compensation model parameter as the final strict imaging model construction parameters to reproduce the calibration image and calibrate the internal and external precision of the reproduced image.

The invention relates to the field of scanning electron microscope image registration, particularly relates to a sequential slice oriented real-time registration device and method for a scanning electron microscope image, and aims to improve the instantaneity of the electron microscope image registration. The real-time registration device comprises an FPGA (Field Programmable Gate Array) and a calculation server, wherein the calculation server comprises a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit); the FPGA is used for directly connecting with an electron microscope toobtain sequential slice image data in real time, calculating a corresponding point between adjacent slice images and finally sending the image data obtained from the electron microscope and the information of the corresponding point between adjacent slices to the calculation server; the CPU in the calculation server regulates and optimizes a corresponding point position obtained by matching in thesequential slice; and the GPU in the calculation server carries out image deformation according to the regulated corresponding point position. By use of the device, a high-accuracy low-time-delay long-sequence registration ability for the high-flux image data of an electron microscope system can be formed, and the real-time registration requirement of high-flux electron microscope sequential slice imaging is met.

The invention discloses a low-light video de-noising method based on phase consistency and scene motion estimation. The edge image of an image to be registered and a reference image is obtained according to a phase consistency method; then registering is performed on the edge images according to a scene-based registering method; difference of the two images after registering is obtained, and corresponding pixels of a moving object are obtained; then motion estimation is performed on the corresponding pixels of all the moving objects so that the corresponding relations of all the pixel points between the image to be registered and the reference image are found; and finally weighted averaging is performed on all the corresponding pixels between the two images and a de-noised image is outputted after completion. Registering precision can be effectively guaranteed so that a great de-noising effect can be achieved, and the method particularly has obvious de-noising effect on low-light video images with large amount of noise.

The invention provides a panorama generating method and device. The method comprises the steps that at least two original two-dimensional images are acquired by a smart mobile device which rotatinglyphotographs a real scene through a fisheye lens, wherein the boundaries of adjacent original two-dimensional images overlap or intersect; the rotation matrix of the smart mobile device in a world coordinate system when each original two-dimensional image in at least two original two-dimensional images is photographed; according to the rotation matrix of the smart mobile device in the world coordinate system when each original two-dimensional image is photographed, the relative position between at least two original two-dimensional images is determined; and at least two original two-dimensionalimages are stitched into a panorama according to the relative position between at least two original two-dimensional images. According to the method provided by the invention, the panorama generatingefficiency can be improved.

The invention discloses a multi-supervision image super-resolution reconstruction method based on a generative adversarial network. The method comprises the following steps: registering slice images with different resolutions of the same slice; making a training data set by using the registered slice images; on the training data set, training a generative adversarial model by using a multi-supervision multi-stage generation model; and reconstructing the low-resolution image into a high-resolution image by using the trained generative adversarial model. According to the method, the image shot under the low-power lens is reconstructed into the high-resolution image, the imaging time can be saved, the hardware space for storing the image can also be saved, and blurring and artifacts of commonmethods on pathological data are overcome.

The invention discloses a global measurement data registration method based on multi-vector constraint, belongs to the field of large-size measurement, and relates to a global measurement data registration method based on multi-vector constraint. According to the method, a plurality of non-coplanar target balls are reasonably arranged in a certain space to serve as common points, a common point data set is obtained through multi-station measurement of a laser tracker, and the station with the minimum overall measurement error serves as a global coordinate system. Normal vectors are then solvedbased on the common points and unitized. And finally, a conversion parameter of the local coordinate system is solved based on the constraint of the normal vector of the redundant unit, and the localdata is registered to the global coordinate system. According to the method, the problem that the data registration precision in a large-size space is abnormally sensitive to common point measurementerrors and the number of common points is fully considered; registration errors of global measurement data can be effectively reduced, the method can be quickly applied to field practice to guaranteethe overall measurement accuracy of geometrical characteristics of large aviation parts, and the method has wide application prospects in the field of large-size measurement.

The invention relates to the manufacturing technical field of a flexible circuit, and particularly to a manufacturing method and apparatus for a liquid metal multilayer circuit. According to the manufacturing method and apparatus for the liquid metal multilayer circuit, the circuit can be printed and packaged in different layers; after the circuits in different layers are registered and aligned through a registering mechanism, the packaging materials among the circuits in different layers are fused through a fusion mechanism; the circuits in different layers are punched and connected through a punching mechanism according to requirements; the problem of frequent damage to circuits caused by directly turning over or stacking the liquid metal circuit is avoided; the multilayer liquid metal circuit can be manufactured accurately and effectively, so that the applicable range of the liquid metal circuit is greatly expanded; the packaging thickness, the alignment precision, and the punching accuracy can be effectively ensured; the problems of circuit mismatch, wrong connections, or punched circuits due to carelessness, and the like are avoided; and in addition, the manufacturing method is simple to use, and the manufacturing apparatus has the advantages of simple assembling, low cost, no pollution, and the like.

The invention provides a machining calibration method and device for a three-dimensional blank workpiece. The machining calibration method comprises the steps: determining a preset number of first sampling areas on the surface of the three-dimensional blank workpiece; obtaining a first cloud data set corresponding to the first sampling area; obtaining a three-dimensional model, wherein the three-dimensional model comprises a machine tool model provided with a workpiece model; obtaining a first coordinate system of the workpiece model through matching of the first cloud data set and a model data set of a first target area in the workpiece model; obtaining a second coordinate system of the machine tool model; and generating a machining path according to the first coordinate system, the second coordinate system and the workpiece model, and then machining the three-dimensional blank workpiece installed on the machine tool. According to the method, the model data set of the corresponding local area of the theoretical three-dimensional model is registered through the point cloud data set corresponding to the local area of the entity workpiece, so that the data processing amount is reduced and the processing efficiency is improved on the basis of ensuring the registration precision.

The embodiment of the invention provides an intraoperative registration method and device, a storage medium and a server. The method in the embodiment of the invention comprises the steps of acquiringa focus image; performing coarse registration on the focus image through a non-rigid registration algorithm and a probe to generate an initial matrix; through an iterative nearest point algorithm, generating a registration matrix according to a pre-collected source point set, a pre-generated target point set and an initial matrix, registration equipment does not need to be purchased, and the registration cost can be saved; the registration precision is ensured, so that the precision of a navigation operation is ensured, and the operation safety is further improved; the consumed time is short,and accurate registration can be efficiently realized.

The embodiment of the invention discloses an image splicing method and device, equipment and a storage medium. The method comprises the steps: obtaining K original images which are sequentially arranged according to a first dimension, and K is an integer greater than or equal to 1; reducing each original image to obtain K reduced images which are sequentially arranged according to the first dimension; determining a target image overlapping rate corresponding to two adjacent reduced images; cutting two corresponding adjacent reduced images according to the overlapping rate of each target imageto obtain an overlapping region between K-1 pairs of two adjacent reduced images; and splicing the K original images according to the overlapping regions of the K-1 pairs of adjacent reduced images.

The invention provides an automatic transferring device. The device comprises a holding mechanism for holding a bed board, a bearing mechanism which is connected with the holding mechanism and used for bearing the holding mechanism, a movement mechanism which is connected with the bearing mechanism and used for driving the bearing mechanism to move, and a control mechanism for controlling the movement mechanism to move along a preset path. The automatic transferring device is a high-precision automatic transferring device through which a patient can automatically and safely move among different image devices with high precision and be accurately carried among the image devices, namely, the patient can automatically move among a plurality of medical devices for examination and switching with the assistance of the automatic transferring device, instead of the assistance of a doctor; with the adoption of the device, the damage of the medical devices to the doctor can be avoided, the pain of the patient can be relieved, the intelligent level of hospital devices can be improved, and the purpose of efficient and accurate diagnosis effect can be achieved.

The invention discloses a point cloud data registration and updating method and device, equipment and a storage medium, relates to the technical field of high-precision maps, and can be used for automatic driving. According to the specific implementation scheme, the method comprises the steps of collecting multiple frames of point cloud data of a target area and initial pose information of each frame of point cloud data; performing registration operation on adjacent point cloud data in the multiple frames of point cloud data to obtain registration parameters of the adjacent point cloud data; adopting the registration parameters of the adjacent point cloud data in the multiple frames of point cloud data to evaluate the credibility of the initial pose information of each frame of point clouddata in the adjacent point cloud data; and registering each frame of point cloud data in a manner of matching the credibility of the initial pose information of each frame of point cloud data. According to the embodiment, the registration efficiency can be improved, the registration precision is improved on the whole, and the registration quality is comprehensively improved.

At present, point cloud registration is widely applied to the traditional manufacturing industry and the emerging high-tech industry. In the fields of unmanned driving, the building industry and the like, point cloud registration lays a foundation for point cloud processing related work such as subsequent point cloud fusion and point cloud surface reconstruction, and the precision of point cloud registration also directly influences the precision of multiple point cloud high-level applications. Research on a point cloud registration algorithm mainly focuses on improvement of the point cloud registration precision and reduction of the operation time of the point cloud registration algorithm. The laser contour sensor is widely applied in industry in recent years, the measurement precision ofthe laser contour sensor can reach the micron order, hundreds of thousands of point cloud data can be obtained within a very small scanning range, and if a traditional point cloud registration algorithm is used, the registration time is long, and aiming at point cloud data obtained by the laser contour sensor, the point cloud registration precision is greatly improved. The invention provides a laser point cloud rapid registration algorithm combined with contour features. Line laser point cloud registration can be completed within several milliseconds.

The invention provides a registration method for hyperspectral image. The method comprises steps: S1, band images photographed by a spectral camera are arranged according to a band sequence, and the known photographing sequence priori knowledge is used for dividing the band images into multiple groups; S2, the groups are ranked, and the groups are well ranked in a sequential end-to-end mode according to the photographing sequence; and S3, the group at the very beginning of the arrangement serves as a first group, according to a reference transformation mode, a first scene image is selected as a reference in the group, intra-group registration is completed, the last scene in the first group serves as a reference next to register a first scene image in a second group, the first scene image of the second group after registration serves as a reference to complete registration of the second group, and thus, inter-group and intra-group registration work can be completed sequentially. The image distortion problem after multiple times of sampling can be effectively solved, and an excellent registration precision result can be obtained.

The invention provides a high-resolution remote sensing image geometric registration method and a high-resolution remote sensing image geometric registration device, and the method comprises the following steps: obtaining high-resolution image data containing projection information, including a reference image and an image to be registered; Performing geographic information coarse matching to obtain a preliminary correction result A1; Cutting out an area A2 having the same geographic coordinate as A1 from the reference image; Uniformly partitioning the A1 and the A2 according to the same size,calculating a standard deviation mean value of each pair of partitioned images with the same geographic coordinate area, and selecting the pair of partitioned images with the standard deviation meanvalue in the middle position; Fourier transform is carried out, and affine parameters between the block image pairs are calculated in a frequency domain according to phase information; And performingaffine correction on the A1 according to the affine parameters to obtain a registration result. According to the method and the device provided by the invention, the service required time can be reduced, the efficiency is improved, the registration precision is ensured, and the cost is reduced. The invention belongs to the field of surveying and mapping remote sensing.

The invention discloses an image control point target system and a method for unmanned aerial vehicle measurement and airborne laser radar. The image control point target system comprises a target panel, laser radar transmitting targets, a center bolt hole and a center mark target. The laser radar transmitting targets are fixedly arranged at four corners of the target panel; the central bolt holeis fixedly formed in the central position of the target panel; and the central mark targets and the central bolt holes are fixedly connected and are uniformly and fixedly arranged on the target panel.The image control point target system is simple in structure, easy to manufacture and low in requirement for the operation environment, and basically meets the requirement for terrain operation. Theimage control point target method is easy to operate, the workload is reduced, the operation efficiency is improved, and the image control point target method can be used as an image control point forunmanned aerial vehicle photogrammetry and airborne laser radar photogrammetry at the same time and is dual-purpose.

A non-rigid registering method of a mouse three-dimensional CT image comprises the following steps: A, a direction and a position of a source image of a mouse three-dimensional CT image are calibrated; B, a threshold segmentation algorithm is used to segment the mouse image to obtain a mouse bone of the source image and a mouse bone of a target image, a center of mass of a bone slice connected domain is extracted to serve as a characteristic point, so that a source image characteristic point and a target image characteristic point are obtained; C, the source image characteristic point and the target image characteristic point in the step B are matched, the matched characteristic point is used to calculate a thin plate batten transformation matrix, and a hierarchical transformation mode is used to carry out the thin plate batten transformation to the mouse source image; and D, the mouse source image after the thin plate batten transformation and the target image carry out a fine registering. According to the invention, when the characteristic points are extracted, the number of characteristic points is reduced; the hierarchical transformation mode is used to overcome the defect of long registering time; an automatic registering in the image registering is basically realized; and registering the three-dimensional mouse image is easier.

The invention provides a single-line point cloud and multi-line point cloud registration method and device, and the method comprises the steps: enabling a single-line point cloud and a multi-line point cloud to be converted to the same coordinate axis, and carrying out the uniform re-sampling of point clouds in a point cloud space through voxels with a preset size; dividing space grids, counting the number of point clouds in each grid, filtering the grids of which the number of the point clouds is lower than a preset threshold value, counting z value distribution of the point clouds in the remaining grids, marking the grids in the threshold value range as the ground, and marking the remaining grids as non-ground points; for point clouds marked as non-ground point clouds, taking all point clouds of an adjacent region of each point cloud, solving a corresponding covariance matrix for a matrix formed by the point clouds, carrying out feature decomposition on the covariance matrix, and filtering out irregular object point clouds based on a proportional relation between feature values; registering the regular object point clouds by adopting an NDT algorithm, and iterating translation amounts of the point clouds in three directions. Optimal registration of the single-line point cloud and the multi-line point cloud is realized, the registration efficiency is high, and the registrationresult is accurate and reliable.

The invention provides a three-dimensional reconstruction data processing method and device, electronic equipment and a storage medium. The method comprises the steps of acquiring first point cloud information, a first picture set, second point cloud information and a second picture set , wherein the first point cloud information is obtained during online reconstruction of a three-dimensional scene, the first picture set is a set of pictures used during online reconstruction of the three-dimensional scene, the second point cloud information is obtained during offline reconstruction of the three-dimensional scene, and the second picture set is a set of pictures used during offline reconstruction of the three-dimensional scene; determining K first target pictures in the first picture set andK second target pictures in the second picture set; determining first pose information corresponding to the first target pictures and second pose information corresponding to the second target pictures; and determining target transformation parameters of the first point cloud information and the second point cloud information according to the K pieces of first pose information and the K pieces ofsecond pose information.

The invention provides a coal mine underground full working face scene space digital construction system and method. The system comprises fixed-point sensors which are disposed on hydraulic support top plates at the head and tail of a working face, and are used for obtaining fixed-point space data; a routing inspection sensor which is arranged on the routing inspection walking mechanism and walks along a track of the routing inspection walking mechanism to obtain routing inspection space data; and a controller which is used for receiving the fixed-point spatial data and the inspection spatial data and obtaining spatial data of the full-working-face scene space according to the fixed-point spatial data and the inspection spatial data. According to the system, a fixed-point scanning mode and an inspection scanning mode are combined to obtain scene space data of the whole working face, and in the head area and the tail area of the working face, an inspection sensor automatically rotates a scanning rotating shaft which is consistent with a fixed-point sensor at the position; registration precision when the two sensors respectively acquire the spatial data is ensured, and splicing acquisition of the spatial data of the whole working face scene is realized.

The embodiment of the invention provides a printer control method and device. The method comprises the steps that a control panel obtains overprinting error information which is used for indicating the corresponding relation between the operation speed of a mobile platform and an overprinting error, wherein the overprinting error is an error between the actual position of a color code sensor and the target position of the color code sensor; the control panel determines the operating speed of the mobile platform; and the control panel adjusts the color code sensor according to the overprintingerror information and the operation speed of the mobile platform, so that the actual position of the color code sensor is aligned with the target position of the color code sensor, wherein after the overprinting error corresponding to the current speed of the mobile platform is obtained according to the overprinting error information, the color code sensor is adjusted to eliminate the overprintingerror. Therefore, according to the embodiment, the overprinting error can be eliminated by combining the real-time operating speed of the mobile platform, so that the overprinting error in each operating state can be eliminated, and the overprinting precision of the printing machine is effectively ensured.

The invention discloses a multilayer film registration control system and method for flexible RFID tag coiling recombination, and the method comprises the steps: (a), forming the system through a feeding unit, a recombination unit, and a material receiving unit; (b), inputting technological reference values: tag step, working tension and tension boundary; (c), enabling a plurality of sensors to cooperatively detect the registration deviation of a multilayer film in a movement feeding direction according to the triggering position relation, building a tension and registration deviation relation model, and converting the registration deviation into a tension adjustment variable; (d), synchronizing the feeding speed of the multilayer film under the target tension in a mode of tension-position-speed closed loop control, correcting the registration deviation, and carrying out the recombination; (e), building a feeding and material receiving coiling diameter algorithm, adjusting the torque output of a driver according to the changes of a coiling diameter, and guaranteeing the constant-tension operation of material feeding and receiving. The method can achieve the high-speed and high-precision registration requirements of the multilayer film under the condition of a controllable tension, and is especially suitable for the preparation of a flexible RFID tag recombination product.