64results about How to "Reduce Stitching Errors" patented technology

The invention discloses a rapid shooting equipment and method of panoramic image and panoramic image sequence in the mobile course, which comprises the following parts: plural image gathering devices, synchronous control system and panoramic image processing working station, wherein the plural image gathering devices are fixed on the same level with imaging point approximately superposed at center of circle of the same arc. The method comprises the following steps: proceeding 360 deg synchronous shooting for all-round of certain point or multiple points on the moving line in the mobile course continuously; disposing the shot image to obtain the panoramic image or continuous panoramic image sequence. The invention solves the real-time capture defect of traditional panoramic shooting pattern in the mobile course, which reduces the difficulty of image for split joint during after-treatment course and improves the efficiency greatly.

The invention discloses a panoramic image obtaining method and apparatus. The method comprises the steps of performing shooting by utilizing double fisheye lenses to obtain two fisheye images, wherein the double fisheyes lenses are two lenses arranged oppositely, and a sum of view field angles of the two lenses is greater than or equal to 360 degrees; mapping the two fisheye images to respective spherical models from a planar coordinate system; extracting feature points from the fisheye images subjected to the mapping, and performing matching on the extracted feature points to obtain feature point matching pairs; determining a transformation matrix between the two fisheye images subjected to the mapping by utilizing the feature point matching pairs; transforming the two fisheye images subjected to the mapping to be in the same spherical model by utilizing the transformation matrix; and splicing the two fisheye images in the same spherical model to obtain a spliced spherical panoramic image. The method and the apparatus are used for improving the splicing precision of the fisheye panoramic image.

The invention provides an unmanned aerial vehicle and a panoramic stitching method, device and system thereof. The method comprises the following steps of obtaining multiple frames of images for stitching; carrying out panoramic stitching on the multiple frames of images at an unmanned aerial vehicle side based on a preset rule to obtain a panoramic image; and sending the panoramic image to a mobile terminal. Since the multiple frames of images for stitching are obtained at the unmanned aerial vehicle side, and the panoramic stitching is carried out on the multiple frames of images based on the preset rule to obtain the panoramic image, image distortion caused by encoding and compression of the multiple frames of images for stitching is avoided, and stitching errors are reduced. Moreover, since the panoramic image is stitched at the unmanned aerial vehicle side, compared with the manner of stitching the panoramic image at the mobile terminal, the response time from obtaining the images to stitching the images is shortened, the stitching efficiency is improved, and the panoramic stitching instantaneity is improved.

The invention proposes a quick mosaic method for aviation images with high tilt rate and low overlapping rate. By utilizing the method, the robustness of registration of the images with high tilt rate and low overlapping rate can be improved, the mosaic error can be reduced, and the synthesis efficiency of a wide image can be enhanced. The method is implemented by the following technical scheme of firstly, calculating a position distribution relationship of the images by utilizing imaging assistance data of the aviation images, establishing an image position relationship matrix, and calculating a reference image; secondly, calculating overlapped regions between adjacent images on adjacent paths through an adjacent image path between a shortest path search image and the reference image, and extracting significant feature points of the overlapped regions; thirdly, normalizing the feature points or the regions into a vertex, constructing a multi-scale image by utilizing different adjacent distance point sets, and performing matching by applying image matching to calculate a spatial transformation matrix; and finally, calculating corresponding coordinates of synthesized image pixels in a source image by using the spatial transformation matrix, and calculating the pixel values of the synthesized image by using bilinear interpolation, thereby obtaining a final mosaic image.

The invention discloses a method for obtaining a panorama by using a code aperture camera. Distant-view panoramas that all parts are clear can be obtained by utilizing the method. The method comprises the following steps of: shooting a view by using the code aperture camera; calculating so as to obtain the sizes of point spread functions of points in shot images; clarifying the images by utilizing the corresponding relationship between the sizes and the depths of the point spread functions and utilizing a deconvolution algorithm; and splicing the clarified images so as to obtain the panorama. By utilizing the method, the splicing mistakes caused by shielding articles in the view are reduced when the images are spliced, and the problem of image blurring caused by the inconsistency of field depths in shooting when the panorama is spliced is avoided. According to the method, the code camera is adopted to acquire the view, so that the requirement on equipment is relatively low, and no extra depth acquiring equipment is used when obtaining an optimal panorama; and the method is simple and convenient in shooting process and easy to implement and popularize.

The invention belongs to the digit image processing technical field, and discloses a fingerprint image splicing method based on regression line characteristics and a TPS deformation model; the methodcomprises the following steps: respectively extracting detail points of two to-be-spliced small area fingerprint images, carrying out registering for the detail points pair after pair, thus obtainingthe detain point pair with the highest similarity fraction; calculating the rotary translation parameter for the primary registering, thus obtaining the primarily spliced fingerprint image; solving arefinement image of the two to-be-spliced small area fingerprint images; using a N neighborhood block method to search a matching regression line of an overlapping area of the refinement image, and searching the matched regression line pair; introducing the TPS deformation model, correcting elastic deformation of the fingerprint image in the regression line pair characteristic layer, and carryingout secondary registering so as to obtain the final result of the secondarily spliced fingerprint image. The method uses the detail points for the primary registering, and uses the N neighborhood block method to search matching regression lines of the overlapping area, thus reducing splicing errors caused by wrongly matched detail points.

The invention belongs to the field of digital image processing technology and discloses a fingerprint image splicing method based on detailed points and a distance image. According to the method, a detailed point pair with the highest matching score in two fingerprint images is obtained according to a fingerprint image matching algorithm based on the detailed points to serve as a reference point pair of rough matching, and rotational translation parameters corresponding to rough matching are solved; and whether rotational translation parameters of fine matching are optimal rotational translation parameters is judged according to a ridge matching error of an overlapping region between a detailed image of the first image and the distance image of the second image nearby the rotational translation parameters of rough matching. Through the method, the probability of erroneous matching is lowered, and splicing errors generated due to erroneous matching of the detailed point pairs are reduced; the problem that the number of the matching detailed point pairs is small due to the fact that the overlapping region is small is avoided; and moreover, the ridge matching error of the overlapping region between the fingerprint images is used to solve the optimal rotational translation parameters, so that fingerprint splicing errors are further reduced.

The invention discloses an auxiliary device of point cloud registration of a three-dimensional scanner and a registration method of the auxiliary device. The auxiliary device comprises a cubic body and a magnetic base, wherein three planes, which are vertical to one another on the cubic body, comprise a first plane, a second plane and a third plane respectively; the magnetic base is fixedly connected with a fourth plane of the cubic body through a connection rod; the fourth plane is opposite to the second plane; and the magnetic base and a product are adsorbed together. With the adoption of the auxiliary device, a scanned part does not need to be scanned for the second time; however, the three planes of the cubic body with a relatively small size are scanned for the second time, and the efficiency can be improved by about 50%; a lug boss on the cubic body is used as a reference point for point cloud registration and the reference point can be easily found; the surface of the lug boss has a regular shape and point cloud data of the reference point are stable, so that errors of the point cloud registration are greatly reduced; and the auxiliary device disclosed by the invention is simple in structure, low in cost and convenient and rapid to operate.

The invention discloses a locomotive body deflection measurement method. The method comprises the steps that height values of a first preset point and a second preset point of the transverse center of a locomotive chassis to a preset horizontal reference plane are respectively measured, and a first height value and a second height value are obtained via the measured height values; the height values of a third preset point, a fourth preset point, a fifth preset point and a sixth preset point of the longitudinal two ends of the locomotive chassis to the preset horizontal reference plane are respectively measured, and a third height value, a fourth height value, a fifth height value and a sixth height value are obtained via the measured height values respectively; and locomotive body deflection is calculated based on the height values. Splicing error caused by locomotive body side beams can be reduced and measurement error of side beam height can be reduced in the measurement process of locomotive body deflection through measurement of an average height value based on the basic principles of geometry so that measurement precision of locomotive body deflection can be enhanced.

The invention discloses a computer-assisted mosaicing method for color patterns and belongs to the technical field of computer image processing. The method comprises the steps of scanning, coding, pairing, simulation mosaicing and real-object mosaicing, wherein n scanned mosaicing modules and edges of mosaicing modules are numbered, points are sampled at equal intervals at edges, color codes of sampled points are extracted, data are established, color codes of edges having the same point sampling number are matched to judge whether edges are paired, and real mosaicing is conducted. According to the method, color information of edge characteristic points and edge lengths serve as mosaicing evidences, and initial screening is conducted with edge information length matching serving as the first condition for mosaicing, so that the time for mosaicing computation is saved.

The invention discloses an aerial remote sensing image splicing method assisted by pose information, and relates to the field of aerial photography remote sensing and image splicing. Firstly, aerial remote sensing image data and navigation pose data are obtained and preprocessed; then, the image displacement in the air strip is solved according to the pose information, and an adjacent image overlapping region is obtained; image features are extracted in the image overlapping area for matching, and an image homography matrix is solved for image registration; the image registration quality is analyzed, if registration is judged to be invalid, a homography matrix is directly solved according to the camera pose converted from the navigation data, and image registration is carried out; and finally, fusion is carried out by adopting a suture band feathering fusion method, and panorama geographic positioning is carried out according to the pose information. The pose information of the airborne navigation system is used for assisting in aerial remote sensing image splicing, the remote sensing image splicing speed is increased, splicing of images with rare features can be achieved, the remote sensing panorama with geographic information is finally obtained, and the method has high practical value.

The invention provides a scene space model construction method and device and a storage medium, and relates to the technical field of computers, and the method comprises the steps: obtaining first point cloud information collected by a mobile point cloud collection device, and obtaining depth image information collected by a fixed-point depth camera device; determining a rotation matrix, corresponding to the global coordinate system, of a camera coordinate system of the fixed-point depth camera equipment, generating a first panoramic image based on the depth image information, and mapping the first panoramic image to a three-dimensional unit spherical surface; performing rotation processing on the three-dimensional unit spherical surface based on the rotation matrix to generate a second panoramic image; generating a scene space model according to the first point cloud information and the second panoramic image; according to the method and device, the electronic equipment and the storage medium, the distortion phenomenon of the scene space model can be effectively reduced, the shooting efficiency can be improved, and the modeling efficiency and accuracy of the scene space model can be improved.

A roller-type fingerprint identification module includes an image acquiring unit, a main roller, a plurality of sensing points, and a sensing unit. By rolling the main roller with a finger in a fixed direction to pass the sensing points on the main roller through the sensing unit, the image acquiring unit is driven by the sensing unit to continuously acquire fingerprint image on each part. Compared with conventional image acquiring unit with 8-dimentional arrangement, the invention can adopt image acquiring unit with 1-dimentional or 2-dimentional arrangement to greatly reduce production cost. A finger can move along a linear direction and difficult image combination caused by passing fingers through the image acquiring unit in different speeds is prevented.

The invention, which belongs to the foundation moon observation field, provides a ground-based moon observation apparatus by using a slit type imaging spectrometer. With the apparatus, problems that the structure of the existing ground-based moon observation device is complicated and the cost is high and a continuous moon spectral image with a high spectral resolution can not be obtained can be solved. The ground-based moon observation apparatus is composed of a two-dimensional tracking and positioning turntable, a slit type imaging spectrometer, a moon imaging instrument, and a computer control and acquisition system. The slit type imaging spectrometer and the moon imaging instrument are installed at the two-dimensional tracking and positioning turntable and are installed by sharing a same optical axis. The slit type imaging spectrometer and the moon imaging instrument are electrically connected with the computer control and acquisition system; and moon moving track simulation software is installed in the computer control and acquisition system. According to the invention, the structure is simple and the cost is low; and a continuous moon spectral image with a high spectral resolution can be obtained. Moreover, spectral radiance information detection is completed simultaneously and the observation efficiency and observation data qualities are effectively improved.

The invention discloses a large-area continuous roller-driven exposure device and a method. The large-area continuous roller-driven exposure device comprises a roller conveying system and a light path system, wherein the roller conveying system comprises a platform, a roller conveying mechanism, a glue liquid tank and a cleaning tank, the roller conveying mechanism is used for conveying a PET (polyethylene terephthalate) flexible substrate by rollers, and the roller conveying mechanism, the glue liquid tank and the cleaning tank are respectively arranged on the platform; the light path system comprises a computer, DMD (digital mirror device) digital masks, an illumination subsystem, a focusing subsystem and a projection and imaging subsystem. The large-area continuous roller-driven exposure device has the advantages that on the basis of tension limiting of the flexible substrate, the displacement of the substrate and the inputting of DMD mask patterns for synchronous conversion can be precisely controlled at low software and hardware cost; the continuous glue applying, photoetching and cleaning can be realized, so as to complete the large-area pattern non-mask photoetching.

The invention discloses a method for rapidly extracting long-fragment DNA from fresh fungus mycelia. The method comprises the following steps: thallus preparation: extracting fresh cultured fungus mycelia, carrying out centrifugation to remove moisture, weighing the mycelia, and rapidly grinding the mycelia into powder in liquid nitrogen; cell lysis: adding a DNA extracting buffer solution, adding60 mu L of 20mg / mL proteinase K, uniformly mixing, and carrying out water bath at 65 DEG for 30 minutes; DNA extraction: adding 0.33mL of KAc, carrying out water bath for 20 minutes, carrying out extraction by virtue of a mixture of phenol, chloroform and isoamyl alcohol, carrying out centrifugation, extracting supernate, adding absolute ethyl alcohol for precipitation, standing at a low temperature for 30 minutes, carrying out centrifugation, and collecting precipitates; purification: adding RNaseA, processing at 37 DEG C for 30 minutes, respectively extracting by virtue of a mixture of phenol, chloroform and isoamylol in a ratio of 25 to 24 to 1 and a mixture of chloroform and isoamylol in a ratio of 24 to 1, and carrying out low-temperature centrifugation for 5 minutes; and extractingsupernate, adding absolute ethyl alcohol for precipitation, selecting precipitates, rinsing, carrying out air drying, dissolving the precipitates into a TE buffer solution, and carrying out preservation for later use. The DNA amount of filamentous fungi extracted by virtue of the method is increased twice, the concentration of the filamentous fungi reaches 277ng / mu L, the lengths of fragments aremore than or equal to 20000bps, and the filamentous fungi are not degraded and are high in purity.

The invention discloses a splicing structure of a microscopic scattering dark field image on surface of an optical element. The method comprises the following steps: scanning the surface of the optical element along a specified patch through a line-scan digital camera; extracting SIFT (Scale Invariant Feature Transform) feature matching point pairs from left and right adjacent interested areas of the microscopic scattering dark field images; screening the SIFT feature matching point pairs by the parallel clustering algorithm to obtain the optimal feature matching point pair set; calculating transformation matrix of the overlapped parts of two pictures according to the optimal feature matching point pair set; splicing the images through the transformation matrix. With the adoption of the method, the splicing of the microscopic scattering dark field image on the surface of the optical element can be stably and quickly performed, little time is cost, the robustness is high, and the splicing error is small.

The invention discloses a construction method for prefabricating pipe galleries. The method comprises the following steps: (1) carrying out construction for prefabricating parts by prefabricating an outer side wall, a top plate and a bottom plate, and arranging connection parts at the upper end and the lower end of the outer side wall, the left end and the right end of the top plate and the left end and the right end of the bottom plate respectively; (2) assembling the prefabricated parts by firstly hoisting the bottom plate, then hoisting the outer side wall, then mutually fixing the connection part at the lower end of the outer side wall with the corresponding connection part on the bottom plate, then mounting a die frame device, finally hoisting the top plate and mutually fixing the connection part at the upper end of the outer side wall with the corresponding connection part on the top plate; and (3) pouring in site and forming by pouring concrete on the inner surface of the outer side wall and the outer surface of the top plate. Through the construction method for prefabricating the pipe galleries, the pipe galleries are constructed by a manner with combination of prefabrication and in-site pouring; the method is simple to operate, rapid in speed and high in safety coefficient; the overall speed is far higher than that of the overall prefabrication and hoisting method; the good structural water prevention effect is achieved; the problem of water prevention is effectively solved.

The invention discloses a three-dimensional configuration splicing method for a large component based on laser vision sensing, relates to the field of image measurement and data processing, and aims to solve the problems of large workload, low efficiency, poor anti-interference capability and difficulty in realizing on-line measurement of large complex profile straightness measurement. The three-dimensional configuration splicing method adopts an autonomously designed straightness splicing algorithm, breaks through the length limitation of single measurement, and is used for subsequent three-dimensional modeling and the like. The laser vision measurement method is derived from the idea of a pitch method, and the straightness of a large complex molded surface is measured by utilizing the advantages of accuracy, stability, real-time performance and the like of short-distance measurement of a laser vision sensor. The size of the molded surface of the large complex molded surface is large,so that the molded surface can be divided into a plurality of partially-overlapped sections with the length equal to that of laser stripes, segmented measurement is carried out in each segmented section, then data processing is carried out to carry out splicing reconstruction on a straightness curve, and the straightness of the large complex molded surface is calculated through the straightness curve obtained after splicing reconstruction. The method is applied to the field of image measurement.

The invention discloses a pavement image acquisition method based on a laser projection supplementary lighting manner, which comprises the following step: fixing a photographic part and a laser on a fixed frame above the rear part of an automobile, with the irradiation direction facing the rear part of the automobile. The method is characterized in that the angle of an emitted light curtain of the laser and the photographic angle of the photographic part are regulated to make the emitted light curtain of the laser and the emission surface of scanning lines of the photographic part non-coplanar and staggered at an included angle; in a ground projection, the scanning lines are in a light curtain of projection; the photographic part is controlled to acquire an image. According to the method, the effect of the obtained image is effectively improved.

The invention provides a rapid panoramic stitching technology suitable for a space complex environment, and the method at least comprises the following steps: S101, carrying out the calibration test of a camera, and extracting an angular point on a calibration image photographed by an effective view field; S102, obtaining an internal parameter of each camera in a single-target calibration manner;S103, performing double-target calibration to obtain external parameters between the adjacent cameras; S104, performing spherical back projection to form a 360-degree*180-degree panoramic plane image,selecting one path of camera as a basic reference coordinate system, completing image coordinate conversion of the other three paths, and realizing indexing of image pixel points under the same coordinate system; s105, adopting a gradual-in gradual-out algorithm to eliminate splicing seams, and carrying out smoothing processing on a large parallax pixel area; and S106, making the pixel point indexes of the four images corresponding to the panoramic image and the fusion coefficient of the stitching area into a LUT table used by the FPGA.

The invention provides a method for intelligently classifying and splicing a plurality of disordered images. The method comprises the following steps of: 1, carrying out feature matching on the plurality of disordered images by adopting an improved BBF-based K-D tree feature matching method; 2, defining splicing relevancy to describe the correlation between the images, and defining splicing credibility to describe the correlation strength between the images; 3, performing sorting priority width traversal on the related images to realize high-speed intelligent classification of the disordered images; and 4, reducing image splicing errors by optimizing a transmission type splicing method, and combining an image splicing line and a fusion algorithm to realize image splicing. Practice shows that faster feature matching is realized, more accurate image correlation description quantity is defined, higher-speed intelligent classification and splicing of the images are realized, and the methodis successful practice of intelligent classification and splicing of the disordered images, is practical, efficient, accurate and rapid, can be easily expanded, and solves the problems of intelligentclassification and accurate splicing of multiple disordered images.

The invention belongs to the field of kiln refractory materials, and particularly discloses a trapezoid-groove refractory brick. When the refractory brick is horizontally arranged, a first protruding block is arranged on the upper surface, a first groove matched with the first protruding block in shape and size is formed in the lower surface, and the first protruding block and the first groove are located on the same vertical line; and a second protruding block is arranged on the right surface, a second groove matched with the second protruding block in shape and size is formed in the left surface, and the second protruding block and the second groove are located on the same horizontal line. When the refractory bricks are spliced, errors are small.

The invention discloses a refrigerator shelf commodity statistics method and system based on an image processing technology, and relates to the field of artificial intelligence, and the method comprises the steps: enabling a terminal controller to splice a first image and a second image into a third image, and enabling a cloud server to segment the third image into a plurality of fourth images; segmenting each fourth image, and performing calculating to obtain the total amount of the first commodities placed in the first row in each storage interlayer in the refrigerator; summing the first commodity total amounts corresponding to all the storage compartments in the refrigerator to obtain a second commodity total amount placed in the first row in the refrigerator and storing the second commodity total amount in a database. The invariant splicing image is rotated according to the circle center; feature point matching errors can be reduced; the image effect obtained through splicing is better, the spliced image is layered into the fourth image, the fourth image is refined and segmented, the commodities placed in the first row in each storage interlayer can be accurately segmented, thestatistical efficiency can be improved, the commodity statistical accuracy can be improved, and the refrigerator commodity management quality can be improved.

The invention discloses a positioning and mapping method and apparatus, and an electronic device. The method comprises the steps of obtaining a spliced vehicle surrounding environment image; semantic segmentation is carried out on the vehicle surrounding environment image, the vehicle surrounding environment image is projected to a camera coordinate system of a top view angle corresponding to the position of the vehicle, and point cloud data of each pixel in the vehicle surrounding environment image in the camera coordinate system of the top view angle is determined; acquiring inertial sensor data; determining point cloud data of each pixel in a world coordinate system in the vehicle surrounding environment image according to the inertial sensor data and the point cloud data of each pixel in the camera coordinate system of the overlooking view angle; performing block matching on the point cloud data under the camera coordinate system and the point cloud data under the world coordinate system according to a semantic segmentation result to obtain multiple groups of matched data; determining deformation of the multiple groups of matched data; and performing non-rigid matching on the matching data of which the deformation is greater than a preset deformation threshold value.

The invention belongs to the technical field of integrated wallboards and discloses a hollow lock catch integrated wallboard and a manufacturing method thereof. A product subjected to cooling moldingis subjected to traction and cut through a traction machine; a required board product is obtained; a board is coated with various design colors and textures of PVC film paper, melamine paper and transfer printing film paper in an attached mode through overlapping of different attaching veneers, and the board product with patterns and colors is formed; and grooves are formed in the two sides of theboard with the patterns and colors through a grooving machine, and the seamlessly spliced lock catch integrated wallboard is obtained. According to the hollow lock catch integrated wallboard and themanufacturing method thereof, the situation of loose splicing of a master buckle and a sub-buckle after extrusion molding is overcome through independent grooving, and the seamless splicing effect isachieved; and the disadvantages that a spliced seam of a traditional product is obvious, and the overall effect is poor are overcome, and the pursuit of people for the overall effect is met.

The invention belongs to the technical field of underground comprehensive pipe galleries, and discloses a double-sided superposed pipe gallery side wall with embedded columns and a manufacturing method thereof. An outer prefabricated layer, an inner prefabricated layer and the embedded columns are connected with a steel bar framework into a whole through pre-cast concrete; a side wall cavity is reserved between the inner prefabricated layer and the outer prefabricated layer; and the inner prefabricated layer and the outer prefabricated layer are connected into a whole through the concrete embedded columns in the side wall cavity. According to the double-sided superposed pipe gallery side wall with the embedded columns and the manufacturing method thereof, only simple turnover equipment isneeded for pouring and manufacturing the double-sided superposed pipe gallery side wall during pouring and manufacturing; the manufacturing environment is good, the concrete curing quality is good, and the control effect of the thickness and the distance of a steel bar protection layer is good; high splicing precision is not needed during production and manufacturing, and the yield of prefabricated parts and the convenience during production are improved; an anti-cracking steel bar net can be accurately placed; the production process of a manufacturer does not need to be greatly innovated; andthe parts are not prone to deformation in the demolding, transporting and hoisting processes, the stress of the parts is good, obvious cracks do not occur on the surface of the concrete, and the self-waterproof performance is good.

The invention discloses a preparation method of a high-quality film lithium niobate micro-nano grating. The preparation method comprises the following steps of cleaning the surface of an LNOI material, preparing a high-precision electron beam mark, coating electron beam positive photoresist ZEP 520A, and exposing an LNOI grating pattern and an electron beam mark protection area, evaporating and stripping Ni metal to serve as a grating waveguide etching mask, adopting F-based RIE dry etching and NH4OH: H2O2: H2O wet processing, and achieving the specified etching depth through multiple times ofcirculation, removing the residual Ni mask by adopting 30% HNO3, and growing a silicon oxide upper cladding. The method provided by the invention is used for preparing the LNOI grating, effectively solves the problem of adhesion of products after lithium niobate etching, and improves the smoothness and verticality of the side wall of the LNOI grating.

The invention relates to the technical field of panoramic image splicing, in particular to a panoramic image splicing method, which comprises the following steps: S1, carrying out basic operation of digital image processing, establishing a matching template of an image, and carrying out Fourier transform and wavelet transform on the image; s2, using a certain matching strategy to find out corresponding positions of templates or feature points in the images to be spliced in the reference image, and then determining the transformation relation between the two images; s3, calculating each parameter value in the mathematical model according to the corresponding relationship between the template or the image features, thereby establishing a mathematical transformation model of the two images; s4, converting the images to be spliced into a coordinate system of the reference image according to the established mathematical conversion model, and completing unified coordinate conversion; and S5,fusing the overlapped areas of the images to be spliced to obtain a spliced and reconstructed smooth seamless panoramic image. According to the invention, misplaced splicing fusion of the images canbe realized, so that the splicing error of the images is greatly reduced.

The invention relates to a method for image mosaic without control points, which comprises the following steps: (1) acquiring an image sequence; (2) extracting a characteristic point set of each image in the image sequence; (3) searching homonymy point pairs of characteristic points between two adjacent images; (4) calculating a Homograph transformation relation between the adjacent images by using an RANSAC error-tolerant algorithm; and (5) using a continual multiplication formula and fusion technology to obtain mosaic results. Aiming at each characteristic point, the method can automatically collect a color channel with highest characteristic significance degree, so as to not only improve amount of the characteristic points greatly, but also improve significance and accuracy of the characteristic points. The method can be used as a correction method for human-computer interactive air photos to correct micro errors accumulated during mosaic. The method needs no control points, and has the advantages of high accuracy, low application cost and high efficiency.