CN114998551B - Grid reconstruction quality optimization method, system, computer and readable storage medium - Google Patents
Grid reconstruction quality optimization method, system, computer and readable storage medium Download PDFInfo
- Publication number
- CN114998551B CN114998551B CN202210925727.5A CN202210925727A CN114998551B CN 114998551 B CN114998551 B CN 114998551B CN 202210925727 A CN202210925727 A CN 202210925727A CN 114998551 B CN114998551 B CN 114998551B
- Authority
- CN
- China
- Prior art keywords
- camera
- image
- initial
- grid
- representing
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 85
- 238000003860 storage Methods 0.000 title claims abstract description 11
- 238000005457 optimization Methods 0.000 title claims description 27
- 238000005259 measurement Methods 0.000 claims abstract description 17
- 238000011478 gradient descent method Methods 0.000 claims abstract description 15
- 230000008859 change Effects 0.000 claims description 39
- 239000000126 substance Substances 0.000 claims description 25
- 238000004364 calculation method Methods 0.000 claims description 22
- 238000012216 screening Methods 0.000 claims description 19
- 238000013441 quality evaluation Methods 0.000 claims description 18
- 238000001303 quality assessment method Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 8
- 238000011524 similarity measure Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000005520 cutting process Methods 0.000 claims description 5
- 230000005484 gravity Effects 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 239000013307 optical fiber Substances 0.000 claims description 3
- 238000009826 distribution Methods 0.000 abstract description 8
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000004891 communication Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000009795 derivation Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000013461 design 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
- 238000003672 processing method Methods 0.000 description 1
- TURAMGVWNUTQKH-UHFFFAOYSA-N propa-1,2-dien-1-one Chemical compound C=C=C=O TURAMGVWNUTQKH-UHFFFAOYSA-N 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
- G06T17/205—Re-meshing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
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 Processing (AREA)
Abstract
The invention discloses a method, a system, a computer and a readable storage medium for optimizing grid reconstruction quality. The method comprises the following steps: acquiring a measurement image and point cloud of aerial photography of the unmanned aerial vehicle, and constructing an initial grid model; constructing an initial camera pair list, wherein the camera pair is a pair of cameras corresponding to two images; selecting an optimal and least number of preferred camera pairs from the initial camera pair list according to the indexes of the camera pairs paired on the grid surface of the initial grid model; calculating the variation amplitude of the grid vertex in the normal direction according to the similarity difference of the two cameras in the optimal camera pair on the images; at different image resolutions, the mesh vertices are updated by a gradient descent method. The beneficial effect of this application is: the method optimizes the model mesh through similarity difference of the model surface on the image, and solves the problems of low degree of detail and uneven distribution of detail areas of the mesh surface generated in the prior art.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a method, a system, a computer and a readable storage medium for optimizing grid reconstruction quality.
Background
With the high-speed development of the field of unmanned aerial vehicles, the large-scale scene aerial view visual construction can be realized by combining the method of shooting high-resolution images by the unmanned aerial vehicles and the multi-view stereo method, and based on the method, the large-scale model three-dimensional reconstruction is widely applied.
At present, the unmanned aerial vehicle shoots the shadowImage data is subjected to three-dimensional reconstruction, image processing is realized by adopting a grid reconstruction mode, the grid reconstruction refers to the operation of constructing a network by utilizing dense point clouds obtained by dense matching, however, in the process of grid reconstruction, under the condition of adopting an irregular camera pair shot by an unmanned aerial vehicle, in a grid model, a grid is continuous on an image but discontinuous on a real object, as shown in the attached figure 1 of the specification, due to the position structure particularity of a continuous part displayed in the image, the surface of the model is subjected to surface structure processingIs a part ofQuilt-only cameraIt is observed that the camera pair shot by the unmanned aerial vehicle at the present stage can cause partial loss of the finally generated grid surface, and the problems of low detail degree and uneven distribution of detail areas exist.
Disclosure of Invention
Based on this, an object of the present invention is to provide a method, a system, a computer and a readable storage medium for optimizing mesh reconstruction quality, which are based on optimizing a camera pair using an approximate scene model, and optimizing a model mesh through similarity difference between model surfaces on an image, so as to solve the problems of low detail degree and uneven distribution of detail areas on the generated mesh surface in the prior art.
In a first aspect, the present application provides a method for optimizing quality of mesh reconstruction, where the method includes the following steps:
acquiring a measurement image and point cloud of aerial photography of the unmanned aerial vehicle, and constructing an initial grid model; wherein the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedral graph cutting method;
constructing an initial camera pair list, wherein the camera pair is a camera pair corresponding to two images;
selecting an optimal and least number of preferred camera pairs from the initial list of camera pairs according to the index of pairing of the camera pairs on the mesh surface of the initial mesh model;
calculating the variation amplitude of the grid vertex in the normal direction according to the similarity difference of the two cameras in the preferred camera pair on the images;
at different image resolutions, the mesh vertices are updated by a gradient descent method.
The mesh reconstruction quality optimization method provided by the invention comprises the steps of obtaining an unmanned aerial vehicle measurement image and point cloud, and generating an initial mesh model according to the point cloud; pairing images of the initial grid model, optimizing pairing and generating an optimal and minimum camera pair set (namely an optimal camera pair); on the basis of the initial grid, according to the similarity measurement of the grid surface between the camera pair images, listing an equation of converting the similarity change into the point gradient change on the grid surface, calculating to obtain the gradient change of the point on the surface in the normal direction, discretizing the gradient change to the grid vertex, and obtaining the gradient change value of the grid vertex; and dividing the image into a plurality of scales according to the resolution, wherein each scale uses the last scale output grid as input to carry out grid subdivision, and under each scale, the input grid iteratively updates grid vertexes by adopting a gradient descent method, and a grid model with higher detail degree is output. By the method, the problems of low surface detail degree and uneven distribution of detail areas of the generated grid in the prior art are effectively solved.
Preferably, in the method for optimizing quality of mesh reconstruction described in the present application, the step of constructing an initial camera pair list includes:
calculating the midpoint of the point cloudIn the imageCorresponding cameraThe number of visible points;
for any camera pairIs subject to conditional restrictions such thatBefore the most number of points they collectively observeA camera asThe initial pairing of (1);
the conditions are limited to:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with a camera moduleAnd withA reprojection pixel distance error therebetween;presentation cameraIs centered toThe angle between the rays;
Preferably, the step of selecting an optimal and least number of preferred camera pairs from the initial camera pair list according to the indexes of pairing of the camera pairs on the mesh surface of the initial mesh model specifically includes:
inputting the initial grid model and an image sequence;
defining a disparity quality evaluation with a reference disparity of 50 DEG, the disparity quality evaluation being:
for model surfaceUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
defining a resolution quality assessment, the resolution quality assessment being:
wherein, the first and the second end of the pipe are connected with each other,representing a 25% difference in resolution; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera,as a camera centerTo pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
defining pixel overlap quality assessment:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe corresponding images are then displayed on the display screen,is the overall area of the image;
define camera pair evaluation based on symmetry on surface:
wherein the content of the first and second substances,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe difference in the angle of (c),if, ifOn the same side of the normal line,the value is-1, otherwise the value is 1;
defining an energy equation:(ii) a Wherein, the first and the second end of the pipe are connected with each other,represents the energy that measures the quality of the disparity,the energy representing the camera versus similar resolution quality,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera image with respect to the model surface,the weight of the energy term is represented by,;
the camera pair that minimizes the energy equation calculation is taken as the preferred camera pair.
Preferably, in the method for optimizing mesh reconstruction quality according to the present application, the step of calculating a variation range of a mesh vertex in a normal direction thereof according to a similarity difference between two cameras in the preferred camera pair on the image specifically includes:
inputting the initial mesh model, the preferred camera pair;
Wherein the content of the first and second substances,is the magnitude of the change of the vertex in the direction of its normal,is the scale factor between the cameras and is,
presentation cameraObservation ofThe average depth of the optical fiber,is the focal length of the camera and is,is the normal direction of the vertex;
calculating the variation amplitude of the vertex in the normal direction:
wherein, the first and the second end of the pipe are connected with each other,representing barycentric coordinates of points within the triangle plane adjacent to the vertex with respect to the vertex,representing the variation amplitude of points in the triangular surface;
points within the triangle plane are represented as the initial mesh model surfaceTo above is connectedContinuous spotCalculating continuous points on the initial mesh modelThe gradient of (c) is changed:
wherein, the first and the second end of the pipe are connected with each other,is a coordinate of a pixel, and is,in order to be the size of the image,the coefficient of the coordinate of the center of gravity is expressed,representing the gradient of the change of the similarity measure of the image areas,representing imagesThe gradient of (a) of (b) is,representing successive pointsDiscretized to image and image pointsA jacobian matrix of the transformation relations of (c),from camera center to image pointVector of direction.
Preferably, in the method for optimizing quality of mesh reconstruction described in the present application, the step of updating mesh vertices by a gradient descent method at different image resolutions specifically includes:
dividing the measurement image and the corresponding camera into three resolution levels of 0.25, 0.5 and 1.0 according to the scaling;
taking the initial grid as an input, and carrying out triangular subdivision according to the level with the scaling of 0.25; wherein, the subdivision condition is that the number of the projected area pixels on the image exceeds 8 for any triangular surface;
taking the grid after the current level subdivision as an input grid of the next level;
sequentially processing according to the zoom level from small to large, and obtaining the zoom level by a calculation formulaIteratively calculating new positions of all vertexes of the mesh model by using all camera pairs;
when the vertex is iteratively updated, if the variation amplitude of the current vertexIf the current vertex position is larger than or equal to the previous updating time, the current vertex position is not updated;
if the variation amplitude of the current vertexAnd if the current vertex position is smaller than the last updating time, updating the current vertex position.
In a second aspect, the present application provides a mesh reconstruction quality optimization system, including:
a model construction module: the method comprises the steps of obtaining a measurement image and point cloud of aerial photography of the unmanned aerial vehicle, and constructing an initial grid model; the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedral graph cutting method;
a camera pair list construction module: the method comprises the steps of constructing an initial camera pair list, wherein the camera pair is a camera pair corresponding to two images;
a screening module: means for selecting an optimal and least number of preferred camera pairs from the initial list of camera pairs according to an index of pairing of the camera pairs on a grid surface of the initial grid model;
the vertex change amplitude calculation module: the method is used for calculating the variation amplitude of the grid vertex in the normal direction according to the similarity difference of the two cameras in the preferred camera pair on the images;
an update module: and the method is used for updating the mesh vertexes through a gradient descent method under different image resolutions.
Preferably, the camera pair list building module specifically includes:
visible point calculation unit: for calculating the midpoint of a point cloudIn the imageCorresponding cameraThe number of visible points;
a condition limiting unit: for aiming at arbitrary camera pairIs subject to condition limitation such thatBefore the most number of points they collectively observeA camera asThe initial pairing of (1);
the conditions are limited to:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with a camera moduleAnda reprojection pixel distance error therebetween;presentation cameraIs centered toThe angle between the rays;
Preferably, the screening module specifically includes:
a first input unit: for inputting the initial mesh model, a sequence of images;
a parallax quality evaluation unit: for defining a disparity quality estimate for a reference disparity of 50 °, the disparity quality estimate being:
for model surfaceUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
a resolution quality evaluation unit: for defining a resolution quality assessment of:
wherein, the first and the second end of the pipe are connected with each other,representing a 25% difference in resolution; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera,as a camera centerTo pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
a pixel quality evaluation unit: for defining pixel overlap quality assessment:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe corresponding images are then displayed on the display screen,is the overall area of the image;
a symmetry evaluation unit: for defining camera pair evaluation based on symmetry on surface:
wherein the content of the first and second substances,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe difference in the angle of (c),if, ifOn the same side of the normal line,the value is-1, otherwise the value is 1;
the energy equation definition unit: for defining the energy equation:(ii) a Wherein the content of the first and second substances,represents the energy that measures the quality of the disparity,the energy representing the camera versus similar resolution quality,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera image with respect to the model surface,the weight of the energy term is represented by,;
preferably, the camera contrast value unit: for taking the camera pair that minimizes the energy equation calculation as the preferred camera pair.
In a third aspect, the present application proposes a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method for optimizing the quality of mesh reconstruction as described in the first aspect above when executing the computer program.
In a fourth aspect, the present application proposes a computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, implements the method for mesh reconstruction quality optimization as described in the first aspect above.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic diagram of a mesh model formed by mesh reconstruction using an irregular pair of cameras;
fig. 2 is a flowchart of a method for optimizing quality of mesh reconstruction according to an embodiment of the present invention;
fig. 3 is a flowchart of a method for constructing an initial camera pair list in the method for optimizing quality of mesh reconstruction according to an embodiment of the present invention;
fig. 4 is a flowchart of a method for screening a preferred camera pair from an initial camera pair list in a mesh reconstruction quality optimization method according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of symmetry of a camera pair in the method for optimizing quality of mesh reconstruction according to the first embodiment of the present invention;
fig. 6 is a flowchart of a method for calculating a variation amplitude of a vertex of a mesh in a normal direction of the vertex in a method for optimizing quality of mesh reconstruction according to an embodiment of the present invention;
FIG. 7 is a diagram illustrating a structure of mesh vertex update according to an embodiment of the present invention;
FIG. 8 is a schematic diagram of an image re-projection in an embodiment of the present invention;
fig. 9 is a flowchart of a method for updating mesh vertices by a gradient descent method in the mesh reconstruction quality optimization method according to the first embodiment of the present invention;
fig. 10 is a schematic structural diagram of a mesh reconstruction quality optimization system according to a second embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more clearly understood, the present application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the application, and that it is also possible for a person skilled in the art to apply the application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that such a development effort might be complex and tedious, but would nevertheless be a routine undertaking of design, fabrication, and manufacture for those of ordinary skill having the benefit of this disclosure, given the benefit of this disclosure, without departing from the scope of this disclosure.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless otherwise defined, technical or scientific terms referred to herein should have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The use of the terms "including," "comprising," "having," and any variations thereof herein, is meant to cover a non-exclusive inclusion; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "plurality" as referred to herein means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.
With the high-speed development of unmanned aerial vehicles, large-scale three-dimensional model reconstruction is widely applied by combining a method of shooting high-resolution images by the unmanned aerial vehicles and a multi-view stereo method. And the grid reconstruction refers to the network construction operation by using dense point cloud obtained by dense matching.
The invention provides a method for efficiently improving the quality of grid details, which is based on the use of an approximate scene model optimization camera pair and optimizes a model grid through similarity difference of a model surface on an image and is used for solving the problems of low degree of grid surface details and uneven distribution of detail areas generated in the prior art.
Referring to fig. 2, a method for optimizing mesh reconstruction quality according to a first embodiment of the present invention includes the following steps:
s11, obtaining a measurement image and point cloud of the unmanned aerial vehicle aerial photography, and constructing an initial grid model.
Wherein the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedral graph cutting method. The triangular mesh constructed by the Delaunay tetrahedron image segmentation method has the characteristics of a hollow circle (any circumscribed circle of 3 points does not contain the 4 th point, namely the circumscribed circle of any 3 points (triangles) is hollow) and the characteristics of the maximized minimum angle, so that the generation of long and narrow triangles in image segmentation is effectively avoided.
And S12, constructing an initial camera pair list.
The camera pair is a pair of two corresponding images. If a triangle of the model surface is observed by two cameras simultaneously, which are represented as images, the two images represent a camera pair of the triangle.
And S13, screening the optimal and least number of preferred camera pairs from the initial camera pair list according to the indexes matched by the camera pairs on the grid surface of the initial grid model.
In an embodiment of the present invention, the pair indicators specifically refer to a parallax quality, a resolution quality, a pixel overlapping parameter, and a symmetry indicator parameter of the camera pair. And calculating and screening the optimal and least camera pairs as the preferred camera pairs by calculating the parameter index values and performing weight distribution.
And S14, calculating the variation amplitude of the grid vertex in the normal direction according to the similarity difference of the two cameras in the optimal camera pair on the images.
In the embodiment of the invention, the calculation of the change amplitude of the grid vertex in the normal direction is to list an equation for converting the similarity change into the point gradient change on the grid surface according to the similarity measurement of the grid surface between the camera and the image, calculate the gradient change of the point on the surface in the normal direction by derivation calculation, discretize the gradient change on the grid vertex and obtain the gradient change value of the grid vertex.
And S15, under different image resolutions, updating the grid vertex by a gradient descent method.
In the embodiment of the invention, the image is divided into a plurality of scales by using the resolution, each scale uses the last scale output grid as input to carry out grid subdivision, and the input grid iteratively updates the grid vertex by adopting a gradient descent method under each scale so as to achieve the aim of outputting a grid model with higher detail degree.
In conclusion, according to the mesh reconstruction quality optimization method provided by the invention, the unmanned aerial vehicle measurement image and the point cloud are obtained, and the initial mesh model is generated according to the point cloud; pairing images of the initial grid model, optimizing pairing and generating an optimal and minimum camera pair set (namely an optimal camera pair); on the basis of the initial grid, according to the similarity measurement of the grid surface between the camera pair images, listing an equation of converting the similarity change into the point gradient change on the grid surface, calculating to obtain the gradient change of the point on the surface in the normal direction, discretizing the gradient change to the grid vertex, and obtaining the gradient change value of the grid vertex; and dividing the image into a plurality of scales according to the resolution, wherein each scale uses the last scale output grid as input to carry out grid subdivision, and under each scale, the input grid iteratively updates grid vertexes by adopting a gradient descent method, and a grid model with higher detail degree is output. By the method, the problems of low detail degree of the surface of the generated grid and uneven distribution of detail areas in the prior art are effectively solved.
Referring to fig. 3, a flowchart of a method for constructing an initial camera pair list in a method for optimizing quality of mesh reconstruction provided in an embodiment of the present invention is shown, where the method includes:
step S21, calculating the midpoint of the point cloudIn the imageCorresponding cameraThe number of visible points.
Step S22, for any camera pairIs subject to condition limitation such thatBefore the most number of points they collectively observeA camera asThe initial pairing of (a).
Specifically, the limiting conditions are:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with camera moduleAnda reprojection pixel distance error therebetween;presentation cameraIs centered toThe angle between the rays.
And S23, finally screening to obtain an initial camera pair set.
Preferably, please refer to fig. 4, which is a flowchart illustrating a method for screening a preferred camera pair from an initial camera pair list in a mesh reconstruction quality optimization method according to an embodiment of the present invention, wherein the method includes:
and S31, inputting an initial grid model and an image sequence.
The image sequence is a group of images, and may be a plurality of picture files, or a plurality of frame images in a video.
And step S32, defining parallax quality evaluation.
Based on the fact that camera pairs with smaller difference in reference parallax angles are more beneficial to model surface optimization, the reference parallax is taken as 50 degrees, parallax quality evaluation is defined, and the parallax quality evaluation is as follows:
wherein the content of the first and second substances,(ii) a Average parallaxFor model surfacesUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
and step S33, defining resolution quality evaluation.
The resolution quality assessment is:
wherein the content of the first and second substances,representing a resolution difference of 25%; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera,as a camera centerTo a pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
and step S34, defining pixel overlapping quality evaluation.
The pixel overlap quality is evaluated as:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe corresponding images are then displayed on the display screen,is the overall area of the image.
In embodiments of the present invention, the above metric can handle most situations, but in some cases can cause problems. As shown in fig. 5, the cameraAll have complete observation of the surfaceI.e. the pixel overlap ratio is similar; for camera polar lineAndare similar, camera pairAnd withAre considered equally good (or equally bad). There is therefore a need to further define the symmetry of the camera pair based on the surface.
Step S35, defining the camera pair based on the symmetry on the surface.
The camera pair is evaluated based on the on-surface symmetry as:
wherein the content of the first and second substances,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe difference in the angle of (c),if, ifOn the same side of the normal line,the value is-1, otherwise the value is 1;
and S36, defining an energy equation.
The energy equation is:
wherein the content of the first and second substances,represents the energy that measures the quality of the disparity,the energy representing the camera pair for similar resolution qualities,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera image to the model surface.The weight of the energy term is represented by,;
and step S37, taking the camera pair which minimizes the energy equation calculation result as a preferred camera pair.
Further, referring to fig. 6, in a mesh reconstruction quality optimization method provided in an embodiment of the present invention, a flow chart of a method for calculating a variation range of a mesh vertex in a normal direction of the mesh vertex is provided, where the method specifically includes:
step S51, input of an initial mesh model, preferably a camera pair.
And S52, calculating optimized vertexes in the grid vertexes in the initial grid model.
wherein the content of the first and second substances,is the magnitude of the change of the vertex in the direction of its normal,is the inter-camera scale factor that is,
presentation cameraObservation ofThe average depth of the optical fiber,is the focal length of the camera and is,is the normal direction of the vertex.
And step S53, calculating the variation range of the vertex in the normal direction.
In the embodiment of the invention, the vertexCan be decomposed into inner points of adjacent triangular surfacesThe calculation formula of the variation amplitude of the vertex in the normal direction is as follows:
wherein, the first and the second end of the pipe are connected with each other,representing barycentric coordinates of points within the triangle plane adjacent to the vertex with respect to the vertex,indicating the magnitude of the change of the points within the triangular surface.
Taking fig. 7 as an example, a schematic structural diagram of mesh vertex update is shown. Variation of mesh verticesFrom the amplitude of variation of points in adjacent triangular facesAnd (4) acting together. Points within the triangle plane are represented as the initial mesh model surfaceContinuous point of. Surface of modelUpper continuous pointThe gradient change of (d) can be expressed as:
wherein, the first and the second end of the pipe are connected with each other,is a coordinate of a pixel, and is,in order to be the size of the image,the coefficient of the coordinates of the center of gravity is expressed,representing the gradient of the change of the similarity measure of the image areas,representing an imageThe gradient of (a) of (b) is,representing successive pointsDiscretizing to image and image pointA jacobian matrix of the transformation relations of (c),from camera center to image pointVector of direction.
The specific derivation is as follows:
1. camera pairCorresponding to the image being. DotIn thatThe similarity measure on can be regarded asIs recorded as the objective function of. DotAmplitude of variation ofIs thatIn thatThe gradient of (d) is noted. When the triangular mesh is optimal, the mesh is,。
2. image processing methodIn thatTo a similarity measure ofThe measure of the similarity of (a) to (b),is thatWarp surfaceInduced reprojection toAs shown in fig. 8.
3. Pixel pointIn thatThe similarity between can be used with respect toIs expressed as. According to the derived chain rule, the similarity measure function is related to points on the surface of the modelThe gradient of (a) is:
4. image pointFor points on the surface of the modelCan be expressed in terms of the camera center to image point directionExpression (2)WhereinA jacobian matrix is represented that,a matrix of projections of the camera is represented,representing pixel pointsDepth in camera coordinates.
5、Representing imagesGradient at pixel points, processing the image by checking it using gradient convolution, usingThe convolution kernel is expressed as follows:
6、the degree of similarity change between images is represented, and the degree of similarity change is calculated by using an integral graph method, wherein the calculation method comprises the following steps:
、、respectively representing the calculation of variance, mean and integral graph of the image; r represents the similarity measure window size.
Preferably, referring to fig. 9, a flowchart of a method for updating a mesh vertex by a gradient descent method in a mesh reconstruction quality optimization method provided by an embodiment of the present invention is shown, where the method specifically includes:
and S81, dividing the measurement image and the corresponding camera into three resolution levels of 0.25, 0.5 and 1.0 according to the scaling.
Step S82, triangular subdivision is performed at a scale of 0.25 with the initial mesh as an input.
The subdivision condition is that the number of pixels of a projection area on an image exceeds 8 for any triangular plane.
And S83, taking the grid after the current level subdivision as an input grid of the next level.
And S84, sequentially processing according to the zoom level from small to large, and iteratively calculating new positions of all vertexes of the mesh model.
The calculation formula for calculating the new positions of all the vertexes of the mesh model is as follows:。
step S85, when the vertex is updated in the iteration, if the change range of the current vertexLarger than or the same as the last update, the current vertex position is not updated.
Step S86, if the variation range of the current vertexAnd if the current vertex position is smaller than the last updating time, updating the current vertex position.
Referring to fig. 10, a grid reconstruction quality optimization system according to a second embodiment of the present invention is specifically provided, which includes:
the model building module 101: the method is used for obtaining the measurement image and the point cloud of the unmanned aerial vehicle aerial photography and constructing an initial grid model.
Wherein the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedron graph cut method.
The camera pair list construction module 102: for constructing an initial list of camera pairs.
Wherein the camera pair is a camera pair corresponding to two images.
The screening module 103: means for selecting an optimal and least number of preferred camera pairs from the initial list of camera pairs according to an index of pairing of the camera pairs on a grid surface of the initial grid model.
Vertex change amplitude calculation module 104: and the method is used for calculating the variation amplitude of the grid vertex in the normal direction of the grid vertex according to the similarity difference of the two cameras in the preferred camera pair on the images.
The update module 105: and the method is used for updating the mesh vertexes through a gradient descent method under different image resolutions.
Preferably, in the system for optimizing quality of mesh reconstruction proposed by the present invention, the camera pair list building module specifically includes:
visible point calculation unit: for calculating the midpoint of a point cloudIn the imageCorresponding cameraThe number of visible points;
a condition limiting unit: for arbitrary camera pairIs subject to condition limitation such thatBefore the most number of points they have observed togetherA camera asThe initial pairing of (a);
the conditions are limited to:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with camera moduleAnd withA reprojection pixel distance error therebetween;presentation cameraCenter toThe angle between the rays;
Preferably, in the system for optimizing quality of mesh reconstruction proposed by the present invention, the screening module specifically includes:
a first input unit: for inputting the initial mesh model, a sequence of images;
a parallax quality evaluation unit: for defining a disparity quality estimate for a reference disparity of 50 °, the disparity quality estimate being:
wherein, the first and the second end of the pipe are connected with each other,(ii) a Average parallax:
for model surfaceUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
a resolution quality evaluation unit: for defining a resolution quality assessment, the resolution quality assessment being:
wherein the content of the first and second substances,representing a resolution difference of 25%; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera and,as a camera centerTo a pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
a pixel quality evaluation unit: for defining pixel overlap quality assessment:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe corresponding images are then displayed on the display screen,is the overall area of the image;
a symmetry evaluation unit: for defining camera pair evaluation based on symmetry on surface:
wherein, the first and the second end of the pipe are connected with each other,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe difference in the angle of (c),if at allOn the same side of the normal line,the value is-1, otherwise the value is 1;
the energy equation definition unit: for defining the energy equation:(ii) a Wherein, the first and the second end of the pipe are connected with each other,represents the energy that measures the quality of the disparity,the energy representing the camera versus similar resolution quality,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera pair to the model surface,the weight of the energy term is represented by,;
the preferred camera is as follows: for taking the camera pair that minimizes the energy equation calculation as the preferred camera pair.
By the grid reconstruction quality optimization system provided by the invention, an unmanned aerial vehicle measurement image and a point cloud are obtained by combining the grid reconstruction quality optimization method, and an initial grid model is generated according to the point cloud; pairing images of the initial grid model, optimizing pairing and generating an optimal and minimum camera pair set (namely an optimal camera pair); on the basis of the initial grid, according to the similarity measurement of the grid surface between the camera pair images, listing an equation of converting the similarity change into the point gradient change on the grid surface, calculating to obtain the gradient change of the point on the surface in the normal direction, discretizing the gradient change to the grid vertex, and obtaining the gradient change value of the grid vertex; and dividing the image into a plurality of scales according to the resolution, wherein each scale uses the last scale output grid as input to carry out grid subdivision, and under each scale, the input grid iteratively updates grid vertexes by adopting a gradient descent method, and a grid model with higher detail degree is output. By the method, the problems of low surface detail degree and uneven distribution of detail areas of the generated grid in the prior art are effectively solved.
It should be noted that the above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules may be located in different processors in any combination.
In addition, the mesh reconstruction quality optimization method of the embodiments of the present application described in conjunction with the drawings can be implemented by a computer device. The computer device may include a processor and a memory storing computer program instructions.
In particular, the processor may include a Central Processing Unit (CPU), or A Specific Integrated Circuit (ASIC), or may be configured to implement one or more Integrated circuits of the embodiments of the present Application.
The memory may include, among other things, mass storage for data or instructions. By way of example, and not limitation, memory may include a Hard Disk Drive (Hard Disk Drive, abbreviated to HDD), a floppy Disk Drive, a Solid State Drive (SSD), flash memory, an optical Disk, a magneto-optical Disk, tape, or a Universal Serial Bus (USB) Drive or a combination of two or more of these. The memory may include removable or non-removable (or fixed) media, where appropriate. The memory may be internal or external to the data processing apparatus, where appropriate. In a particular embodiment, the memory is a Non-Volatile (Non-Volatile) memory. In particular embodiments, the Memory includes Read-Only Memory (ROM) and Random Access Memory (RAM). The ROM may be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically Erasable PROM (EEPROM), electrically rewritable ROM (EAROM), or FLASH Memory (FLASH), or a combination of two or more of these, where appropriate. The RAM may be a Static Random-Access Memory (SRAM) or a Dynamic Random-Access Memory (DRAM), where the DRAM may be a Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), an Extended Data Out Dynamic Random Access Memory (EDODRAM), a Synchronous Dynamic Random Access Memory (SDRAM), and the like.
The memory may be used to store or cache various data files for processing and/or communication use, as well as possibly computer program instructions for execution by the processor.
The processor may be configured to read and execute the computer program instructions stored in the memory to implement any one of the mesh reconstruction quality optimization methods in the above embodiments.
The computer device may also include a communication interface and a bus. The processor, the memory and the communication interface are connected through a bus and complete mutual communication.
The communication interface is used for realizing communication among modules, devices, units and/or equipment in the embodiment of the application. The communication interface may also be implemented with other components such as: the data communication is carried out among external equipment, image/data acquisition equipment, a database, external storage, an image/data processing workstation and the like.
A bus comprises hardware, software, or both that couple components of a computer device to one another. Buses include, but are not limited to, at least one of the following: data Bus (Data Bus), address Bus (Address Bus), control Bus (Control Bus), expansion Bus (Expansion Bus), and Local Bus (Local Bus). By way of example and not limitation, a Bus may include an Accelerated Graphics Port (AGP) or other Graphics Bus, an Enhanced Industrial Standard Architecture (EISA) Bus, a Front Side Bus (FSB), a Hypertransport (HT) Interconnect, an ISA (ISA) Bus, an InfiniBand (LPC) Interconnect, a Low Pin Count (LPC) Bus, a memory Bus, a microchannel Architecture (MCA) Bus, a PCI-Express (PCI-interface) Bus, a PCI-Express (PCI-Express) Bus, a Serial Advanced Technology Attachment (vladvanced Technology, SATA) Bus, a Video Association (Video Association) Bus, or a combination of two or more of these or other suitable electronic buses. A bus may include one or more buses, where appropriate. Although specific buses are described and shown in the embodiments of the application, any suitable buses or interconnects are contemplated by the application.
The computer device may execute the mesh reconstruction quality optimization method in the embodiment of the present application based on the acquired data information, thereby implementing the mesh reconstruction quality optimization method described in conjunction with fig. 2.
In addition, in combination with the method for optimizing the reconstruction quality of the mesh in the foregoing embodiments, the embodiments of the present application may provide a computer-readable storage medium to implement. The computer readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement any of the above-described embodiments of a method for mesh reconstruction quality optimization.
In the description of the specification, reference to the description of "one embodiment," "some embodiments," "an example," "a specific example," or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (5)
1. A mesh reconstruction quality optimization method for optimizing images shot by an Unmanned Aerial Vehicle (UAV), the method comprising:
acquiring a measurement image and point cloud of aerial photography of the unmanned aerial vehicle, and constructing an initial grid model; wherein the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedral graph cutting method;
constructing an initial camera pair list, wherein the camera pair is a camera pair corresponding to two images;
selecting an optimal and least number of preferred camera pairs from the initial camera pair list according to the indexes of the camera pairs paired on the grid surface of the initial grid model;
calculating the variation amplitude of the grid vertex in the normal direction according to the similarity difference of the two cameras in the optimal camera pair on the images;
under different image resolutions, updating the grid vertex by a gradient descent method;
the step of building an initial list of camera pairs comprises:
calculating the midpoint of the point cloudIn the imageCorresponding cameraThe number of visible points;
for any camera pairIs subject to condition limitation such thatBefore the most number of points they collectively observeA camera asThe initial pairing of (1);
the conditions are limited to:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with a camera moduleAnda reprojection pixel distance error therebetween;presentation cameraCenter toThe angle between the rays;
The step of selecting an optimal and least number of preferred camera pairs from the initial camera pair list according to the indexes of pairing of the camera pairs on the mesh surface of the initial mesh model specifically comprises:
inputting the initial grid model and an image sequence;
taking the reference parallax as 50 ° to define a parallax quality evaluation, wherein the parallax quality evaluation is as follows:
wherein, the first and the second end of the pipe are connected with each other,(ii) a Average parallax:
for model surfaceUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
defining a resolution quality assessment, the resolution quality assessment being:
wherein, the first and the second end of the pipe are connected with each other,representing a 25% difference in resolution; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera and,as a camera centerTo a pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
defining pixel overlap quality assessment:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe corresponding images are then displayed on the display screen,is the overall area of the image;
defining an evaluation of camera pairs based on symmetry on a surface:
wherein, the first and the second end of the pipe are connected with each other,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe difference in the angle of (c),if, ifOn the same side of the normal line,the value is-1, otherwise the value is 1;
defining an energy equation:(ii) a Wherein, the first and the second end of the pipe are connected with each other,represents the energy that measures the quality of the disparity,the energy representing the camera pair for similar resolution qualities,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera image with respect to the model surface,the weight of the energy term is represented by,;
taking the camera pair which minimizes the energy equation calculation result as a preferred camera pair;
the step of calculating the variation range of the grid vertex in the normal direction thereof according to the similarity difference of the two cameras in the preferred camera pair on the image specifically comprises:
inputting the initial mesh model, the preferred camera pair;
Wherein the content of the first and second substances,is the magnitude of the change of the vertex in the direction of its normal,is the scale factor between the cameras and is,is the vertex of the mesh before optimization,
presentation cameraObservation ofThe average depth of (a) is determined,presentation cameraObservation ofThe average depth of (a) is determined,is the focal length of the camera and is,is the normal of the vertexDirection;
calculating the variation amplitude of the vertex in the normal direction:
wherein, the first and the second end of the pipe are connected with each other,representing barycentric coordinates of points within the triangle plane adjacent to the vertex with respect to the vertex,representing the variation amplitude of points in the triangular surface;
points within the triangle plane are represented as the initial mesh model surfaceContinuous point ofCalculating continuous points on the initial mesh modelThe gradient of (c) is changed:
wherein the content of the first and second substances,is a coordinate of a pixel, and is,is the size of the image to be displayed,the coefficient of the coordinate of the center of gravity is expressed,representing the gradient of the change of the similarity measure of the image areas,representing an imageThe gradient of (a) of (b) is,representing successive pointsDiscretized to image and image pointsA jacobian matrix of the transformation relations of (c),from camera center to image pointVector of direction.
2. The method for optimizing mesh reconstruction quality according to claim 1, wherein the step of updating mesh vertices by gradient descent method at different image resolutions specifically comprises:
dividing the measurement image and the corresponding camera into three resolution levels of 0.25, 0.5 and 1.0 according to the scaling;
taking the initial grid as input, and carrying out triangular subdivision according to the level with the scaling of 0.25; wherein, the subdivision condition is that the number of the projected area pixels on the image exceeds 8 for any triangular surface;
taking the grid after the current level subdivision as an input grid of the next level;
sequentially processing according to the zoom level from small to large, and obtaining the zoom level by a calculation formulaIteratively calculating new positions of all vertexes of the mesh model by using all the camera pairs;
when the vertex is updated in iteration, if the change amplitude of the current vertexIf the current vertex position is larger than or equal to the previous updating time, the current vertex position is not updated;
3. A system for optimizing quality of mesh reconstruction, the system comprising:
a model construction module: the method comprises the steps of obtaining a measurement image and point cloud of aerial photography of the unmanned aerial vehicle, and constructing an initial grid model; the initial mesh model is a mesh model with a triangular mesh constructed by a Delaunay tetrahedral graph cutting method;
a camera pair list construction module: the method comprises the steps of constructing an initial camera pair list, wherein the camera pair is a camera pair corresponding to two images;
a screening module: means for selecting an optimal and least number of preferred camera pairs from the initial list of camera pairs according to an index of pairing of the camera pairs on a grid surface of the initial grid model;
the vertex change amplitude calculation module: the method is used for calculating the variation amplitude of the grid vertex in the normal direction of the grid vertex according to the similarity difference of the two cameras in the preferred camera pair on the images;
an updating module: the method is used for updating the grid vertexes through a gradient descent method under different image resolutions;
the camera pair list building module specifically includes:
visible point calculation unit: for calculating the midpoint of a point cloudIn the imageCorresponding cameraThe number of visible points;
a condition limiting unit: for aiming at arbitrary camera pairIs subject to conditional restrictions such thatBefore the most number of points they have observed togetherA camera asThe initial pairing of (1);
the conditions are limited to:
wherein the content of the first and second substances,representing three-dimensional pointsCamera with a camera moduleAnd withA reprojection pixel distance error therebetween;presentation cameraIs centered toThe angle between the rays;
The screening module specifically comprises:
a first input unit: for inputting the initial mesh model, a sequence of images;
parallax quality evaluation unit: defining a disparity quality estimate for taking the reference disparity as 50 °, the disparity quality estimate being:
for model surfaceUpper point ofCentering of cameraAngle of parallax,Indicates the area of the image region, where the region is indicated asIt is a surfaceOn-cameraIn the area of the projection, the image is projected,representing pixels in an imageIs back projected to the surfaceA point on;
a resolution quality evaluation unit: for defining a resolution quality assessment, the resolution quality assessment being:
wherein the content of the first and second substances,representing a resolution difference of 25%; is provided withRepresenting points on the surface of a modelThe distance to the center of the camera is normalized by the focal length of the camera,which represents the focal length of the camera,as a camera centerTo pointThe vector of (a);indicating that the camera is relative to the two-camera resolutionNormalized difference in length:
a pixel quality evaluation unit: for defining pixel overlap quality assessment:
wherein, the first and the second end of the pipe are connected with each other,presentation cameraThe image is to be mapped to a corresponding image,is the overall area of the image;
a symmetry evaluation unit: for defining camera pair evaluation based on symmetry on surface:
wherein the content of the first and second substances,,representing points on the surface of a modelThe angular difference of the normal is averaged out,,is camera center to pointRelative to its normal lineThe angle difference of (a) to (b),if, ifOn the same side of the normal line,the value is-1, otherwise the value is 1;
the energy equation definition unit: for defining the energy equation:(ii) a Wherein, the first and the second end of the pipe are connected with each other,represents the energy that measures the quality of the disparity,the energy representing the camera versus similar resolution quality,the energy representing the quality of the overlap of the camera to the pixel,an energy function representing the quality of symmetry of the camera pair to the model surface,the weight of the energy term is represented by,;
preferably, the camera contrast value unit: the method is used for taking the camera pair which minimizes the energy equation calculation result as a preferred camera pair;
the vertex change amplitude calculation module is used for: inputting the initial mesh model, the preferred camera pair;
Wherein the content of the first and second substances,is the magnitude of the change of the vertex in the direction of its normal,is the inter-camera scale factor that is,is the vertex of the mesh before optimization,
presentation cameraObservation ofThe average depth of (a) is determined,presentation cameraObservation ofThe average depth of the optical fiber,is the focal length of the camera and is,is the normal direction of the vertex;
calculating the variation amplitude of the vertex in the normal direction:
wherein, the first and the second end of the pipe are connected with each other,representing barycentric coordinates of points within the triangle plane adjacent to the vertex with respect to the vertex,representing the variation amplitude of points in the triangular surface;
triangular surfacePoints within are represented as the initial mesh model surfaceContinuous point ofCalculating continuous points on the initial mesh modelThe gradient of (c) is changed:
wherein the content of the first and second substances,is the coordinates of the pixels and is the coordinates of the pixels,is the size of the image to be displayed,the coefficient of the coordinate of the center of gravity is expressed,representing the gradient of the change of the similarity measure of the image areas,representing an imageThe gradient of (a) is determined,representing successive pointsDiscretized to image and image pointsA jacobian matrix of the transformation relations of (c),from camera center to image pointVector of direction.
4. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the mesh reconstruction quality optimization method according to any one of claims 1 to 2 when executing the computer program.
5. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method of mesh reconstruction quality optimization according to any one of claims 1 to 2.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210925727.5A CN114998551B (en) | 2022-08-03 | 2022-08-03 | Grid reconstruction quality optimization method, system, computer and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210925727.5A CN114998551B (en) | 2022-08-03 | 2022-08-03 | Grid reconstruction quality optimization method, system, computer and readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114998551A CN114998551A (en) | 2022-09-02 |
CN114998551B true CN114998551B (en) | 2022-11-18 |
Family
ID=83022563
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210925727.5A Active CN114998551B (en) | 2022-08-03 | 2022-08-03 | Grid reconstruction quality optimization method, system, computer and readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114998551B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007135659A2 (en) * | 2006-05-23 | 2007-11-29 | Elbit Systems Electro-Optics Elop Ltd. | Clustering - based image registration |
CN107291093A (en) * | 2017-07-04 | 2017-10-24 | 西北工业大学 | Unmanned plane Autonomous landing regional selection method under view-based access control model SLAM complex environment |
CN110223383A (en) * | 2019-06-17 | 2019-09-10 | 重庆大学 | A kind of plant three-dimensional reconstruction method and system based on depth map repairing |
JP2019185070A (en) * | 2018-03-30 | 2019-10-24 | 東京ガスiネット株式会社 | Information processing system and program |
CN114708375A (en) * | 2022-06-06 | 2022-07-05 | 江西博微新技术有限公司 | Texture mapping method, system, computer and readable storage medium |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP4136611A4 (en) * | 2020-05-18 | 2023-06-07 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image reconstruction |
CN113393577B (en) * | 2021-05-28 | 2023-04-07 | 中铁二院工程集团有限责任公司 | Oblique photography terrain reconstruction method |
-
2022
- 2022-08-03 CN CN202210925727.5A patent/CN114998551B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007135659A2 (en) * | 2006-05-23 | 2007-11-29 | Elbit Systems Electro-Optics Elop Ltd. | Clustering - based image registration |
CN107291093A (en) * | 2017-07-04 | 2017-10-24 | 西北工业大学 | Unmanned plane Autonomous landing regional selection method under view-based access control model SLAM complex environment |
JP2019185070A (en) * | 2018-03-30 | 2019-10-24 | 東京ガスiネット株式会社 | Information processing system and program |
CN110223383A (en) * | 2019-06-17 | 2019-09-10 | 重庆大学 | A kind of plant three-dimensional reconstruction method and system based on depth map repairing |
CN114708375A (en) * | 2022-06-06 | 2022-07-05 | 江西博微新技术有限公司 | Texture mapping method, system, computer and readable storage medium |
Non-Patent Citations (3)
Title |
---|
利用飞行时间三维相机的非刚体形状三维重建;童晶等;《计算机辅助设计与图形学学报》;20110315(第03期);3-10 * |
宽基线视差图像的拼接算法;左森等;《计算机工程》;20070520(第10期);181-183 * |
影像三维重建的网格自适应快速优化;张春森等;《武汉大学学报(信息科学版)》;20200305(第03期);98-105 * |
Also Published As
Publication number | Publication date |
---|---|
CN114998551A (en) | 2022-09-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8331615B2 (en) | Match, expand, and filter technique for multi-view stereopsis | |
US8711143B2 (en) | System and method for interactive image-based modeling of curved surfaces using single-view and multi-view feature curves | |
JP5778237B2 (en) | Backfill points in point cloud | |
US9256980B2 (en) | Interpolating oriented disks in 3D space for constructing high fidelity geometric proxies from point clouds | |
CN107909640B (en) | Face relighting method and device based on deep learning | |
CN113393577B (en) | Oblique photography terrain reconstruction method | |
CN114119853B (en) | Image rendering method, device, equipment and medium | |
CN113643414B (en) | Three-dimensional image generation method and device, electronic equipment and storage medium | |
CN109816706A (en) | A kind of smoothness constraint and triangulation network equal proportion subdivision picture are to dense matching method | |
CN111627119A (en) | Texture mapping method, device, equipment and storage medium | |
CN111583398B (en) | Image display method, device, electronic equipment and computer readable storage medium | |
CN114202632A (en) | Grid linear structure recovery method and device, electronic equipment and storage medium | |
CN115631317A (en) | Tunnel lining ortho-image generation method and device, storage medium and terminal | |
CN114998551B (en) | Grid reconstruction quality optimization method, system, computer and readable storage medium | |
CN106600691B (en) | Fusion correction method and system of multi-channel two-dimensional video images in three-dimensional geographic space | |
CN115294277B (en) | Three-dimensional reconstruction method and device of object, electronic equipment and storage medium | |
CN115994993A (en) | Stylized face three-dimensional shape modeling method, system, equipment and storage medium | |
CN115421509B (en) | Unmanned aerial vehicle flight shooting planning method, unmanned aerial vehicle flight shooting planning device and storage medium | |
WO2019144281A1 (en) | Surface pattern determining method and device | |
CN115937395A (en) | Electrical equipment model rendering method and device, computer equipment and storage medium | |
JP2020187626A (en) | Image processing device, image processing method, and program | |
CN116295031B (en) | Sag measurement method, sag measurement device, computer equipment and storage medium | |
TWI782806B (en) | Point cloud rendering method | |
Wu et al. | An accurate novel circular hole inspection method for sheet metal parts using edge-guided robust multi-view stereo | |
CN116863085B (en) | Three-dimensional reconstruction system, three-dimensional reconstruction method, electronic equipment and storage medium |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |