CN113706382A - Image splicing method and device based on quadtree homogenization method and storage medium - Google Patents
Image splicing method and device based on quadtree homogenization method and storage medium Download PDFInfo
- Publication number
- CN113706382A CN113706382A CN202110997000.3A CN202110997000A CN113706382A CN 113706382 A CN113706382 A CN 113706382A CN 202110997000 A CN202110997000 A CN 202110997000A CN 113706382 A CN113706382 A CN 113706382A
- Authority
- CN
- China
- Prior art keywords
- image
- points
- characteristic
- feature points
- quadtree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 76
- 238000000265 homogenisation Methods 0.000 title claims abstract description 44
- 230000004927 fusion Effects 0.000 claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims description 25
- 238000000605 extraction Methods 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 238000007670 refining Methods 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 230000000007 visual effect Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000010187 selection method Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 229920001296 polysiloxane Polymers 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/005—Tree description, e.g. octree, quadtree
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a method for splicing images automatically based on quadtree homogenization, which solves the problems of low quality and visual dislocation of the spliced images in the prior art; the method comprises the following steps: acquiring a first image and a second image which need to be spliced; respectively extracting the characteristic points of the first image and the second image; performing quadtree homogenization on the feature points, and determining the homogenized feature points; registering the first image and the second image to determine a registered image; carrying out image fusion on the registration image to determine a complete image; and then realize accurate to many images splice to the image after the concatenation is a seamless high definition image.
Description
Technical Field
The application relates to the technical field of computer image splicing, in particular to an image splicing method and device based on a quadtree homogenization method and a storage medium.
Background
Image stitching is an increasingly popular area of research and has become a hotspot in photogrammetry, computer vision, image processing, and computer graphics research. The non-common form of image stitching is to construct a seamless, high-definition image with higher resolution and larger field of view than a single image by overlapping a series of spatially overlapping images.
In recent years, with research and development of image stitching technology, rendering based on images becomes a focus of combining two complementary fields, and currently, image stitching methods based on Feature correlation include an image stitching method based on SIFT Feature correlation (Scale artifact Feature Transform), an image stitching method based on FAST algorithm corner detection algorithm, an image stitching method based on speedup-robust Feature detection algorithm (speedup-up Feature) and the like.
For the increasingly demanding image stitching field, the quality of the stitching result obtained by image stitching by the methods is low.
Disclosure of Invention
The embodiment of the invention provides an image splicing method, an image splicing device and a storage medium based on a quadtree homogenization method, solves the problems of low quality and visual dislocation of spliced images in the prior art, and further realizes accurate splicing of multiple images, and the spliced images are seamless and high-definition images.
In a first aspect, an embodiment of the present invention provides an image stitching method based on a quadtree homogenization method, where the method includes:
acquiring a plurality of images, and determining a first image and a second image;
respectively extracting the characteristic points of the first image and the second image;
performing quadtree homogenization on the feature points, and determining homogenized feature points;
registering the first image and the second image to determine a registered image;
and carrying out image fusion on the registration image to determine a complete image.
Equally dividing the first image and the second image into four nodes in sequence;
when the splitting condition is satisfied, repeatedly executing the following steps: determining the number of the characteristic points in each node, and splitting the nodes with the number of the characteristic points larger than 1 into four nodes;
when the splitting condition is not met any more, selecting the characteristic point with better quality in each byte to determine as the homogenized characteristic point;
the splitting conditions are as follows: the total number of the nodes is less than or equal to a preset value, and the number of the characteristic points in each node is greater than 1.
With reference to the first aspect, in a possible implementation manner, the quadtree homogenization further includes:
and selecting the pixel points with better quality, wherein the selection method of the pixel points with better quality is to select the feature points with the highest angular point response value.
With reference to the first aspect, in a possible implementation manner, the separately extracting the feature points of the first image and the feature points of the second image includes:
constructing a Hessian matrix and calculating a characteristic value a;
constructing an image Gaussian pyramid, and determining the position and the size of each candidate position through a model;
and comparing the size of each pixel point processed by the Hessian matrix with 26 points in the three-dimensional field of the pixel point, determining whether the pixel point is the maximum value or the minimum value, and if so, keeping the pixel point as the feature point.
With reference to the first aspect, in a possible implementation manner, the separately extracting the feature points of the first image and the feature points of the second image includes:
and taking a pixel point and comparing the pixel point with surrounding pixel points, and determining the pixel point as the characteristic point if the pixel point is different from the surrounding points.
With reference to the first aspect, in a possible implementation manner, the registering the first image and the second image includes:
detecting the homogenized feature points in the first image and in the second image;
calculating the matching of the homogenized feature points, and calculating an initial value of a homography change matrix between the first image and the second image;
and iteratively refining the homography transformation matrix repeatedly, and performing guide matching on the first image and the second image until the number of the corresponding homogenized characteristic points is stable.
With reference to the first aspect, in a possible implementation manner, the image fusing the registration images includes: and carrying out weighted average on the gray values of the overlapped area of the first image and the second image.
In a second aspect, an embodiment of the present invention further provides an image stitching device based on a quadtree homogenization method, including:
the image acquisition module is used for acquiring a plurality of images and determining a first image and a second image;
the characteristic point extraction module is used for respectively extracting the characteristic points of the first image and the second image;
the characteristic point selection module is used for performing quadtree homogenization on the characteristic points and determining homogenized characteristic points;
the image registration module is used for registering the homogenized feature points to determine a registered image;
and the image fusion module is used for registering the first image and the second image to determine a registered image.
With reference to the second aspect, in a possible implementation manner, the feature point selecting module is specifically configured to:
equally dividing the first image and the second image into four nodes in sequence;
when the splitting condition is satisfied, repeatedly executing the following steps: determining the number of the characteristic points in each node, and splitting the nodes with the number of the characteristic points larger than 1 into four nodes;
when the splitting condition is not met any more, selecting the characteristic point with better quality in each byte to determine as the homogenized characteristic point;
the splitting conditions are as follows: the total number of the nodes is less than or equal to a preset value, and the number of the characteristic points in each node is greater than 1.
With reference to the second aspect, in a possible implementation manner, the feature point selecting module further includes: a pixel screening module, the pixel screening module is used for: and selecting the pixel points with better quality, wherein the selection method of the pixel points with better quality is to select the feature points with the highest angular point response value.
With reference to the second aspect, in a possible implementation manner, the feature point extraction module is specifically configured to:
constructing a Hessian matrix and calculating a characteristic value a;
constructing an image Gaussian pyramid, and determining the position and the size of each candidate position through a model;
and comparing the size of each pixel point processed by the Hessian matrix with 26 points in the three-dimensional field of the pixel point, determining whether the pixel point is the maximum value or the minimum value, and if so, keeping the pixel point as the feature point.
With reference to the second aspect, in a possible implementation manner, the feature point extraction module is specifically configured to:
and comparing a pixel point with surrounding pixel points, and determining the pixel point as the characteristic point if the pixel point is different from the surrounding points.
With reference to the second aspect, in one possible implementation manner, the image registration module is specifically configured to:
detecting the homogenized feature points in the first image and in the second image;
calculating the matching of the homogenized feature points, and calculating an initial value of a homography change matrix between the first image and the second image;
and iteratively refining the homography transformation matrix repeatedly, and performing guide matching on the first image and the second image until the number of the corresponding homogenized characteristic points is stable.
With reference to the second aspect, in a possible implementation manner, the image fusion module further includes a weighting module, and the weighting module is configured to:
and carrying out weighted average on the gray values of the overlapped area of the first image and the second image.
In a third aspect, the present application further provides an image stitching server based on a quadtree homogenization method, including a memory and a processor;
the memory is to store computer-executable instructions;
the processor is configured to execute the computer-executable instructions to implement the method of the first aspect as well as various possible implementations of the first aspect.
In a fourth aspect, the present application further provides a computer-readable storage medium, which stores executable instructions that, when executed by a computer, can implement the method according to the first aspect and various possible implementation manners of the first aspect.
One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:
the embodiment of the invention adopts an image splicing method based on quadtree homogenization, and adopts the steps of obtaining a plurality of images, determining a first image and a second image, wherein the second image and the second image are not only one image, the first image can be an image spliced by two images, but also can be a single image, and the images can be processed; the characteristic points of the first image and the characteristic points of the second image are respectively extracted, the extraction of the characteristic points can enable the images to be matched more accurately through the characteristic points when the images are spliced, and the images can be spliced more completely when the images are spliced; performing quadtree homogenization on the feature points, and determining the homogenized feature points, wherein in the step, the feature points with good quality can be obtained by performing quadtree homogenization on the feature points, so that accurate connection of images is realized; registering the homogenized feature points to determine a registered image, wherein the homogenized feature points are feature points with better quality, and fusing the registered image to determine a complete image; the problems that in the prior art, the quality of the spliced images is low and the images are staggered visually are solved, so that the multiple images can be spliced accurately, and the spliced images are seamless and high-definition images.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments of the present invention or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flowchart of an image stitching method based on a quadtree homogenization method according to an embodiment of the present invention;
FIG. 2 is a flowchart of feature point extraction in an image stitching method based on a quadtree homogenization method according to an embodiment of the present invention;
FIG. 3 is a flowchart of feature point homogenization in the image stitching method based on the quadtree homogenization method according to the embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating splitting in homogenization according to an embodiment of the present invention;
FIG. 5 is a flowchart of image registration in an image stitching method based on a quadtree homogenization method according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of image fusion in the image stitching method based on the quadtree homogenization method according to the embodiment of the present invention;
FIG. 7A is a first image in an image stitching method based on a quadtree homogenization method according to an embodiment of the present invention;
FIG. 7B is a second image of the image stitching method based on the quadtree homogenization method according to the embodiment of the present invention;
FIG. 7C is a diagram illustrating a well-stitched image in the image stitching method based on the quadtree homogenization method according to the embodiment of the present invention;
FIG. 8 is a schematic diagram of an image stitching apparatus based on a quadtree homogenization method according to an embodiment of the present invention;
fig. 9 is a schematic diagram of an image stitching server based on a quadtree homogenization method according to an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the present day life, image stitching is increasingly popular and has become a research hotspot, and compared with a single image before, the image stitching can form a seamless and high-definition image and has a larger visual field than the single image.
The embodiment of the invention provides an image splicing method based on a quadtree homogenization method, which comprises the following steps as shown in figure 1:
step S101: acquiring a plurality of images, and determining a first image and a second image.
Step S102: and respectively extracting the characteristic points of the first image and the second image.
Step S103: and performing quadtree homogenization on the feature points, and determining the homogenized feature points.
Step S104: and registering the first image and the second image to determine a registered image.
Step S105: and carrying out image fusion on the registration images to determine a complete image.
The method comprises the steps of obtaining a plurality of images, and determining a first image and a second image, wherein the second image and the second image are not only one image, the first image can be an image formed by splicing two images, but also can be a single image, and the plurality of images can be processed; respectively extracting the characteristic points of the first image and the second image, wherein the extraction of the characteristic points can enable the images to be more accurately matched through the characteristic points when the images are spliced, and can more completely splice the images when the images are spliced; performing quadtree homogenization on the feature points, and determining the homogenized feature points, wherein in the step, the feature points with good quality can be obtained by performing quadtree homogenization on the feature points, so that accurate connection of images is realized; registering the homogenized feature points to determine a registered image, wherein the homogenized feature points are feature points with better quality, and fusing the registered image to determine a complete image; the problems that in the prior art, the quality of the spliced images is low and the images are staggered visually are solved, so that the multiple images can be spliced accurately, and the spliced images are seamless and high-definition images.
In step S101, a first image and a second image are determined, where the first image may be a single image or a spliced image, and the second image is the same.
In step S102, the specifically extracting the feature points of the first image and the feature points of the second image respectively is to extract the feature points by using a SURF algorithm, and the specific steps are as shown in fig. 2:
step S201: a Hessian matrix is constructed and eigenvalues a are calculated.
Step S202: and constructing a Gaussian pyramid of the image, and determining the position and the size of each candidate position through a model.
Step S203: and comparing the size of each pixel point processed by the Hessian matrix with 26 points in the three-dimensional field of the pixel point, determining whether the pixel point is the maximum value or the minimum value, and if so, keeping the pixel point as a characteristic point.
The SURF algorithm is a robust algorithm for detecting and describing local feature points, and in step S201, a Hessian matrix is constructed for generating edge points for image stabilization, and the Hessian matrix is a square matrix formed by second derivatives of a multivariate function and describes local curvature of the function. The Hessian matrix for one image f (x, y) is as follows:
when the discrimination of the Hessian matrix obtains a local maximum value, the current pixel point is judged to be a point brighter or darker than other points in the surrounding neighborhood, and therefore the position of the key point is located.
In step S202, filters of the same size are used in different layers of the gaussian pyramid of the image, and the blur coefficients of the filters gradually increase.
After the above steps are performed, 4 × 4 rectangular region blocks around the feature point are extracted, the rectangular blocks are taken along the main direction of the feature point, Haar wavelet features of 25 pixel points in the horizontal direction and the vertical direction are counted in each rectangular region block, and the horizontal direction and the vertical direction are relative to the main direction.
The method for extracting the feature points of the first image and the feature points of the second image respectively further comprises the following steps: and comparing a pixel point with surrounding pixel points, and determining the pixel point as a characteristic point if the pixel point is different from the surrounding points. This method, called ORB algorithm, extracts feature points, which finds a specific region from the image, called keypoint, i.e. a small salient region in the image, which is usually able to determine some kind of edge. And extracting the characteristic points of the image by combining an ORB algorithm and an SURF algorithm.
In step S103, the quadtree is homogenized, and as shown in fig. 3, the method is divided into the following steps:
step S301: when the splitting condition is satisfied, repeatedly executing the following steps: and determining the number of the characteristic points in each node, and splitting the nodes with the number of the characteristic points larger than 1 into four nodes.
Step S302: and when the splitting condition is not met any more, selecting the characteristic point with better quality in each byte to determine the characteristic point as the homogenized characteristic point.
Step S303: the splitting conditions were: the total number of the nodes is less than or equal to a preset value, and the number of the characteristic points in each node is greater than 1.
In step S301, the four nodes may be divided into four first-level blocks according to the size of the image after the splitting, and the four first-level blocks respectively correspond to four coordinates of an upper left corner, an upper right corner, a lower left corner, and a lower right corner, as shown in fig. 4, some feature point coordinates may be shared by multiple blocks, and all feature points falling within a certain block range belong to an element of the node. When the feature points in each block are counted and the number of the feature points in the block is 0, the block is deleted. When the number of feature points in a block is 1, the block is not split. It should be noted here that after an upper block is split into four lower blocks, the node needs to be deleted from the original linked list, and three blocks are actually added in one split. During use, a block total is usually set in advance, and when the total is reached, splitting is not performed.
And when the total number of the splits is reached, selecting the pixel points with better quality, wherein the selection method of the pixel points with better quality is to select the characteristic points with the highest angular point response value. The corner point is that the movement of the window image in any direction causes the obvious change of the image gray scale. In the method of the application, the point with the highest response value of the corner point is selected as the characteristic point to be reserved.
The image registration is a technology for determining an overlapping area and an overlapping position between images to be stitched, and is the core of the whole image stitching. The algorithm flow of the solution is as follows.
In step S104, the first image and the second image are registered, as shown in fig. 5, including the following steps:
step S501: detecting the homogenized feature points in the first image and in the second image.
Step S502: the matching of the feature points after homogenization is calculated, and an initial value of a homography change matrix between the first image and the second image is calculated.
Step S503: and repeatedly carrying out iterative refining on the homography transformation matrix, and carrying out guide matching on the first image and the second image until the number of the corresponding homogenized characteristic points is stable.
After the registration step is carried out, the two images can be seamlessly spliced to the maximum extent, the image splicing precision is ensured, and the image splicing success rate is effectively improved.
In step S105, a complete image is determined, including: the gray values of the overlapping areas of the first image and the second image are weighted-averaged. The first image and the second image are shown in fig. 7A and 7B, respectively.
The gray value Piexl of the pixel point in the overlapping area of the images has the gray value Piexl of the corresponding pixel point in the first image and the second image_LAnd Piexl_RA weighted average is obtained, i.e. Piexl ═ Piexl_L+(1-k)*Pixel_RWhere k is an adjustable factor. Typically, 0 < k < 1, i.e., in the overlap region, as shown in FIG. 6, a greater correlation is established between the two images along points in the image overlap region, such that
Wherein d is1And d2The distances of the left and right borders of the overlap region are indicated, respectively, using the formula pixl k pixil_L+(1+k)*Pixel_RSuture thread treatment is performed.
After the above series of processing, we finally obtain the stitching result, as shown in fig. 7C, fig. 7C shows that two images are stitched into one image after the above steps.
The embodiment of the invention also provides an image splicing device based on the quadtree homogenization method, which comprises the following steps as shown in fig. 8: an image acquisition module 801, a feature point extraction module 802, a feature point selection module 803, an image registration module 804 and an image fusion module 805.
The image acquiring module 801 is configured to acquire a plurality of images and determine a first image and a second image.
The feature point extracting module 802 includes a SURF selecting module and an ORB selecting module. The feature point extraction module 802 is configured to extract feature points of the first image and feature points of the second image, respectively.
The SURF selection module is used for: constructing a Hessian matrix and calculating a characteristic value a; constructing an image Gaussian pyramid, and determining the position and the size of each candidate position through a model; and comparing the size of each pixel point processed by the Hessian matrix with 26 points in the three-dimensional field of the pixel point, determining whether the pixel point is the maximum value or the minimum value, and if so, keeping the pixel point as a characteristic point.
The ORB selection module is used for: and comparing a pixel point with surrounding pixel points, and determining the pixel point as a characteristic point if the pixel point is different from the surrounding points.
The first image and the second image are extracted through the SURF module and the ORG module, and the feature points can be completely and omnidirectionally extracted. The SURF module extracts the feature points to extract the feature points of the object which can smooth the first image and the second image, and the ORB module can find the feature points on the edges of the first image and the second image.
The feature point selection module 803 includes a homogenization module and a pixel point selection module. The feature point selection module 803 is configured to perform quadtree homogenization on the feature points, and determine homogenized feature points.
The homogenization module is for: equally dividing the first image and the second image into four nodes in sequence; when the splitting condition is satisfied, repeatedly executing the following steps: determining the number of the characteristic points in each node, and splitting the nodes with the number of the characteristic points larger than 1 into four nodes; when the splitting condition is not met any more, selecting the characteristic points with better quality in each byte to determine as the homogenized characteristic points; the splitting conditions were: the total number of the nodes is less than or equal to a preset value, and the number of the characteristic points in each node is greater than 1.
The pixel point screening module is used for: and selecting the pixel points with better quality, wherein the selection method of the pixel points with better quality is to select the characteristic points with the highest angular point response value.
In the homogenization module and the pixel point screening module of the feature point selection module 803, more feature points extracted by the two methods are screened, and the number of layers of the feature points distributed in the space is about the same, so that the more uniform the number of layers is, the more accurately the spatial set relationship can be expressed when the Jining image is registered.
The image registration module 804: and the method is used for registering the homogenized feature points to determine a registered image. The method is specifically used for: detecting the homogenized feature points in the first image and in the second image; calculating the matching of the homogenized feature points, and calculating the initial value of a homography change matrix between the first image and the second image; and repeatedly carrying out iterative refining on the homography transformation matrix, and carrying out guide matching on the first image and the second image until the number of the corresponding homogenized characteristic points is stable. The image registration module 804 enables more accurate stitching of the images when the first and second images are registered.
The image fusion module 805 also includes a weighting module. The image fusion module 805 is configured to perform image fusion on the registered images to determine a complete image.
The weighting module is to: the gray values of the overlapping areas of the first image and the second image are weighted-averaged.
The apparatuses or modules illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. The functionality of the modules may be implemented in the same one or more software and/or hardware implementations of the present application. Of course, a module that implements a certain function may be implemented by a plurality of sub-modules or a combination of sub-modules.
The methods, apparatus or modules described herein may be implemented in a computer readable program code means for a controller in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, Application Specific Integrated Circuits (ASICs), programmable logic controllers and embedded microcontrollers, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may therefore be considered as a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.
The application also provides an image stitching server based on the quadtree homogenization method, as shown in fig. 9, which includes a memory 901 and a processor 902; the memory 901 is used to store computer executable instructions; the processor 902 is configured to execute computer-executable instructions to implement the method for image stitching based on the quadtree homogenization method according to the embodiment of the present invention.
The storage medium includes, but is not limited to, a Random Access Memory (RAM), a Read-Only Memory (ROM), a Cache, a Hard disk (Hard disk), or a Memory card (HDD). The memory may be used to store computer program instructions.
The application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores executable instructions, and when the computer executes the executable instructions, the image stitching method based on the quadtree homogenization method provided by the embodiment of the invention can be realized.
Some of the modules in the apparatus described herein may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, classes, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by software plus necessary hardware. Based on such understanding, the technical solutions of the present application may be embodied in the form of software products or in the implementation process of data migration, which essentially or partially contributes to the prior art. The computer software product may be stored in a storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to perform the methods described in the various embodiments or portions of the embodiments of the present application.
The embodiments in the present specification are described in a progressive manner, and the same or similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. All or portions of the present application are operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, mobile communication terminals, multiprocessor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
Although the present application provides method steps as described in an embodiment or flowchart, additional or fewer steps may be included based on conventional or non-inventive efforts. The sequence of steps recited in this embodiment is only one of many steps performed and does not represent a unique order of execution. When an actual apparatus or client product executes, it can execute sequentially or in parallel (e.g., in the context of parallel processors or multi-threaded processing) according to the methods shown in this embodiment or the figures.
The above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications or substitutions do not depart from the spirit and scope of the present disclosure.
Claims (10)
1. An image splicing method based on a quadtree homogenization method is characterized by comprising the following steps:
acquiring a first image and a second image which need to be spliced;
respectively extracting the characteristic points of the first image and the second image;
performing quadtree homogenization on the feature points, and determining homogenized feature points;
registering the first image and the second image to determine a registered image;
and carrying out image fusion on the registration image to determine a complete image.
2. The method of claim 1, wherein said quadtree homogenizing said feature points comprises:
equally dividing the first image and the second image into four nodes in sequence;
when the splitting condition is satisfied, repeatedly executing the following steps: determining the number of the characteristic points in each node, and splitting the nodes with the number of the characteristic points larger than 1 into four nodes;
when the splitting condition is not met any more, selecting the characteristic point with better quality in each byte to determine as the homogenized characteristic point;
the splitting conditions are as follows: the total number of the nodes is less than or equal to a preset value, and the number of the characteristic points in each node is greater than 1.
3. The method of claim 2, wherein said selecting the better quality pixel points of said block comprises: and selecting the characteristic point with the highest corner response value.
4. The method according to claim 1, wherein the extracting the feature points of the first image and the feature points of the second image respectively comprises:
constructing a Hessian matrix and calculating a characteristic value a;
constructing an image Gaussian pyramid, and determining the position and the size of each candidate position through a model;
and comparing the size of each pixel point processed by the Hessian matrix with 26 points in the three-dimensional field of the pixel point, determining whether the pixel point is the maximum value or the minimum value, and if so, keeping the pixel point as the feature point.
5. The method according to claim 1, wherein the extracting the feature points of the first image and the feature points of the second image respectively comprises:
and selecting a pixel point and comparing the pixel point with the surrounding pixel points, and if the selected pixel point is different from the surrounding pixel points, determining the selected pixel point as the characteristic point.
6. The method of claim 1, wherein registering the first image and the second image comprises:
detecting the homogenized feature points in the first image and in the second image;
calculating the matching of the homogenized feature points, and calculating an initial value of a homography change matrix between the first image and the second image;
and iteratively refining the homography transformation matrix repeatedly, and performing guide matching on the first image and the second image until the number of the corresponding homogenized characteristic points is stable.
7. The method of claim 1, wherein the image fusing the registered images comprises:
and carrying out weighted average on the gray values of the overlapped area of the first image and the second image.
8. An image stitching device based on a quadtree uniformization method, comprising:
the image acquisition module is used for acquiring a first image and a second image which need to be spliced;
the characteristic point extraction module is used for respectively extracting the characteristic points of the first image and the second image;
the characteristic point selection module is used for performing quadtree homogenization on the characteristic points and determining homogenized characteristic points;
an image registration module, configured to register the first image and the second image, and determine a registered image;
and the image fusion module is used for carrying out image fusion on the registration image to determine a complete image.
9. An image splicing server based on a quadtree homogenization method is characterized by comprising a memory and a processor;
the memory is to store computer-executable instructions;
the processor is configured to execute the computer-executable instructions to implement the method of any of claims 1-7.
10. A computer-readable storage medium having stored thereon executable instructions that, when executed by a computer, are capable of implementing the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110997000.3A CN113706382A (en) | 2021-08-27 | 2021-08-27 | Image splicing method and device based on quadtree homogenization method and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110997000.3A CN113706382A (en) | 2021-08-27 | 2021-08-27 | Image splicing method and device based on quadtree homogenization method and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113706382A true CN113706382A (en) | 2021-11-26 |
Family
ID=78656135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110997000.3A Pending CN113706382A (en) | 2021-08-27 | 2021-08-27 | Image splicing method and device based on quadtree homogenization method and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113706382A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110390640A (en) * | 2019-07-29 | 2019-10-29 | 齐鲁工业大学 | Graph cut image split-joint method, system, equipment and medium based on template |
| CN110675437A (en) * | 2019-09-24 | 2020-01-10 | 重庆邮电大学 | Image matching method based on improved GMS-ORB characteristics and storage medium |
| CN111291768A (en) * | 2020-02-17 | 2020-06-16 | Oppo广东移动通信有限公司 | Image feature matching method and device, equipment and storage medium |
| CN111754402A (en) * | 2020-06-09 | 2020-10-09 | 南京师范大学 | Image splicing method based on improved SURF algorithm |
-
2021
- 2021-08-27 CN CN202110997000.3A patent/CN113706382A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110390640A (en) * | 2019-07-29 | 2019-10-29 | 齐鲁工业大学 | Graph cut image split-joint method, system, equipment and medium based on template |
| CN110675437A (en) * | 2019-09-24 | 2020-01-10 | 重庆邮电大学 | Image matching method based on improved GMS-ORB characteristics and storage medium |
| CN111291768A (en) * | 2020-02-17 | 2020-06-16 | Oppo广东移动通信有限公司 | Image feature matching method and device, equipment and storage medium |
| CN111754402A (en) * | 2020-06-09 | 2020-10-09 | 南京师范大学 | Image splicing method based on improved SURF algorithm |
Non-Patent Citations (2)
| Title |
|---|
| ELECTECH6计算机视觉LIFE: "VSLAM系列原创04讲|四叉树实现ORB特征点均匀化分布：原理+代码", pages 1 - 12, Retrieved from the Internet <URL:https://mp.weixin.qq.com/s/O2_wiwnWjiI2R4VVAzrbVQ> * |
| 最难不过二叉树: "OpenCV探索之路（二十四）图像拼接和图像融合技术", pages 1 - 17, Retrieved from the Internet <URL:https://www.cnblogs.com/skyfsm/p/7411961.html> * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6902611B2 (en) | Object detection methods, neural network training methods, equipment and electronics | |
| CN112348815B (en) | Image processing method, image processing apparatus, and non-transitory storage medium | |
| CN107301402B (en) | Method, device, medium and equipment for determining key frame of real scene | |
| KR20160048140A (en) | Method and apparatus for generating an all-in-focus image | |
| CN108986152B (en) | Foreign matter detection method and device based on difference image | |
| CN102713938A (en) | Scale space normalization technique for improved feature detection in uniform and non-uniform illumination changes | |
| CN109919971B (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| CN112967381B (en) | Three-dimensional reconstruction method, apparatus and medium | |
| JP2016212784A (en) | Image processing apparatus and image processing method | |
| CN105578070A (en) | Image processing method and terminal | |
| CN108229281B (en) | Neural network generation method, face detection device and electronic equipment | |
| CN115937546A (en) | Image matching method, three-dimensional image reconstruction method, image matching device, three-dimensional image reconstruction device, electronic apparatus, and medium | |
| CN108960247A (en) | Image significance detection method, device and electronic equipment | |
| CN113706382A (en) | Image splicing method and device based on quadtree homogenization method and storage medium | |
| JP2017103756A (en) | Image data processing apparatus and method | |
| CN115619636A (en) | Image stitching method, electronic device and storage medium | |
| CN114549825A (en) | Target detection method and device, electronic equipment and storage medium | |
| CN115424181A (en) | Target object detection method and device | |
| CN115731256A (en) | Vertex coordinate detection method, device, equipment and storage medium | |
| CN109644236B (en) | Angle detection method | |
| CN112149463A (en) | Image processing method and device | |
| CN110705575A (en) | Positioning method and device, equipment and storage medium | |
| JP2015170205A (en) | Feature amount generation device, feature amount generation method, and program | |
| Wei et al. | Objective image quality assessment based on saliency map | |
| CN116152382B (en) | Digital representation conversion method and device of structural floor plan and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |