CN113096232A - Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions - Google Patents
Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions Download PDFInfo
- Publication number
- CN113096232A CN113096232A CN202110341923.3A CN202110341923A CN113096232A CN 113096232 A CN113096232 A CN 113096232A CN 202110341923 A CN202110341923 A CN 202110341923A CN 113096232 A CN113096232 A CN 113096232A
- Authority
- CN
- China
- Prior art keywords
- reflection
- camera
- coordinates
- height
- light source
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/50—Lighting effects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- 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
- G06T2207/10012—Stereo images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention provides a feasible and effective mirror surface object surface three-dimensional reconstruction scheme based on light reflection, which is used for calibrating a multi-view camera system and unifying a world coordinate system. Then, coordinates of the light source point, the reflection point and the optical center are calculated in sequence. And then calculating gradient point cloud data of the surface according to the coordinates obtained by calculation in the steps and by combining the reflection principle of light. And finally, integrating the gradient point cloud to obtain reconstructed height point cloud data, and drawing a surface map after performing smooth filtering processing on the point cloud to obtain a three-dimensional reconstruction map of the surface of the object.
Description
Technical Field
The invention belongs to the technical field of three-dimensional reconstruction, and particularly relates to a method for three-dimensional reconstruction of a mirror surface object surface based on light reflection.
Background
In modern scientific research, three-dimensional reconstruction technology plays an important role in many research fields including but not limited to computer, communication engineering, optical engineering and the like, and a great deal of technicians devote themselves to the three-dimensional reconstruction technology in order to achieve great breakthrough. The method has the advantages that efficiency and accuracy are guaranteed, and meanwhile, the method for acquiring the three-dimensional shape of the object is significant in various fields such as medicine, buildings, military affairs, buildings and photography. Therefore, it is needless to say that the three-dimensional reconstruction technique will take an important position in the future. The three-dimensional surface measurement technology is divided into a contact type and a non-contact type, the traditional contact type has great limitations, one is that a reconstruction target can be damaged, and the other is that the technical difficulty is great when flowing liquid such as a water surface and the like is used as the reconstruction target. The non-contact measurement technology can better solve the problems, and typical non-contact measurement technologies such as a photoelectric technology and an information processing technology have both precision and speed, and are the main sites of research and development. The projection method, the laser triangulation method and the reflection method are common technologies for optical three-dimensional surface measurement, and are suitable for diffuse reflection targets, and the reflection rule is widely applied to optical devices with high precision requirements.
The working principle of the camera is to project an object in a three-dimensional world onto a plane to complete the conversion from three dimensions to two dimensions. On the contrary, the three-dimensional reconstruction aims to realize the conversion from two dimensions to three dimensions and complete the reconstruction of an object, and the three-dimensional reconstruction can be regarded as the inverse process of camera imaging. Generally, the information contained in the two-dimensional image is seriously insufficient because the information can only reflect a certain specific angle, the three-dimensional model is quite different, the information can be observed from any angle, and the information is comprehensive, rich and accurate. The three-dimensional reconstruction technology has great significance no matter in the entertainment field for pursuing the extreme visual experience or in the military medical field for pursuing the accuracy and the comprehensiveness of information.
The developed countries begin earlier research in the direction of three-dimensional reconstruction, and obtain more achievements. The domestic research investment in this area is started late, but is rapidly developed. Although the three-dimensional reconstruction technology has achieved great performance in many technical directions, the reconstruction of flowing liquid such as water surface still has a difficult problem, and in addition, the precision of the three-dimensional reconstruction and the complexity of the reconstruction method are the common contradictions in the reconstruction work.
Disclosure of Invention
The invention provides a novel method for three-dimensional reconstruction of a mirror surface object surface based on light reflection, aiming at the problems that the traditional object surface three-dimensional reconstruction method can cause contact to an object and the reconstruction precision is low.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
(1) the calibration of the multi-view system adopts a two-by-two calibration method, namely, the camera 1 and the camera 2, the camera 1 and the camera 3, and the camera 1 and the camera 4 are respectively calibrated in a binocular mode. The three sets of calibrated world coordinate systems are all overlapped with the camera coordinate system of the camera 1, so that the world coordinate systems are smoothly unified, and the relationship between the cameras 2, 3, 4 and the camera 1 can be obtained through the rotation and translation matrixes obtained by the sets of binocular calibration.
(2) The formula for solving the object surface gradient point cloud is deduced according to the reflection principle of light, and the gradient of the reflection point on the object surface can be solved as long as the world coordinates of a group of light source points, the reflection point and the optical center are obtained.
(3) The height of the reflecting point is preset, the world coordinate of the reflecting point is calculated according to the pixel coordinate of the object surface reflection image collected by the multi-view system and the camera calibration principle, and the error of the preset height is reduced by adopting an iterative approximation algorithm. And solving the world coordinates of the light source points and the world coordinates of the optical centers by utilizing a camera calibration principle and the pixel coordinates of the light source template.
(4) And solving the object surface gradient according to the world coordinates of the reflection points, the light source points and the light center, calculating a height point cloud according to the gradient point cloud, and reconstructing a three-dimensional image of the object surface after smooth filtering.
Further, a formula for solving the object surface gradient point cloud is derived according to the reflection principle of light as follows: according to the law of light reflection, there are
R-I＝kN
Wherein R is the reflected light vector, I is the incident light vector, N is the normal vector, and the world coordinates of the light source point S, the reflected point P, and the camera optical center C are S (X)s Ys Zs)P(Xp Yp Zp)C(Xc Yc Zc)
The coordinates of vector I and vector R are then expressed as:
if the normal direction is (x)n,yn,zn) I.e. kN ═ (x)n,yn,zn) Then, from R and I, the direction vector of the normal, i.e., kN, can be found.
The slopes of the reflection point P on the x and y axes are:
fx(x,y)＝-xn/zn
fy(x,y)＝-yn/zn
namely:
finally, the gradient of the surface of the mirror surface object to be measured is obtained
Namely the slope of the x direction and the slope of the y direction on the irregular surface of the mirror surface object to be measured.
Further, the iterative approximation algorithm of the reflection points is that the height of the reflection points is preset to be h1, the world coordinate of the reflection points is solved, the world coordinate of the light center is combined with the light source points to solve the gradient, then the integration is carried out to obtain the height point cloud, the height of the actually obtained height point cloud is set to be h2, h2 is preset to be the height of the reflection points, the world coordinate of the reflection points is solved again, the gradient and the height are solved again, the ith calculation height is set to be hi, the (i + 1) th calculation height is hi +1, and the loop is terminated until (hi +1-hi)/hi is less than 0.01.
Furthermore, when the coordinates of the light source points are obtained, the coordinates of the mirror images of the light source points are obtained by utilizing the high-reflection plane mirror, and then the coordinates of the light source points are obtained by utilizing the symmetry characteristics of the high-reflection plane mirror and the plane mirror.
Further, the algorithm for reconstructing the height point cloud from the gradient point cloud is a fourier integration algorithm.
Drawings
FIG. 1 is a schematic representation of the reflection of light;
FIG. 2 is a diagram of a light source template;
FIG. 3 is a schematic diagram of light source point determination;
fig. 4 is a system work flow diagram.
Detailed Description
The present invention will be further described with reference to the accompanying drawings.
The invention provides a new method for quickly and efficiently reconstructing the surface of a mirror target in three dimensions, which mainly comprises the following steps: the calibration of the multi-objective acquisition system, the calculation of light source point reflection points and light center world coordinates, the calculation of gradient point cloud according to the light reflection principle, the integral calculation of the gradient point cloud to the height point cloud and the reconstruction of an object surface three-dimensional image.
(1) Calibrating a multi-view system: the three-dimensional calibration tool of Matlab guides in pictures of three groups of cameras, namely a camera 12, a camera 13 and a camera 14 in sequence, so that calibration of the three groups of binocular cameras is smoothly completed, internal parameters of the four cameras and respective rotation matrixes and translation matrixes of the cameras 2, 3 and 4 compared with the camera 1 are obtained, and then the world coordinate system of the whole measurement system can be unified only by reasonably utilizing the calibration parameters.
(2) As shown in the figure I, the method for solving the gradient of the object surface based on the principle of light reflection is that, if I is a unit vector taking incident light as the direction, R is a unit vector taking reflected light as the direction, and N is a unit normal vector of a reflection point, then I is a unit vector taking incident light as the direction, R is a unit vector taking reflected light as the direction, N is a unit normal vector of the reflection point, and
sinβ＝sinα
|R|·|N|·sinβ＝|I|·|N|·sin(π-α)
according to the equation of length of vector product | c | ═ a × b | ═ a | · | b | sin < a, b >, the equation is transformed to obtain:
|R×N|＝|I×N|
from the combination of I N and R N being in the same direction and both being out of the plane of the paper, the equation can be derived:
R×N＝I×N
equation pair) equation two sides are taken and are arranged to obtain:
|(R-I)×N|＝|R-I|·|N|·sinλ＝0
in equation), since I and R are not collinear, | R-I | ≠ 0, where λ is the angle between R-I and N, it is easy to know that the equation holds only when λ ═ 0 or λ ═ pi, i.e. R-I and N are collinear, then:
R-I＝kN
where k is a positive number.
To this end, a fast and efficient specular object surface reconstruction system based on light reflection has been implemented and validated. The invention provides a non-contact accurate surface reconstruction system based on the light reflection principle, and solves the problems that the conventional three-dimensional reconstruction may damage an accurate object and the reconstruction accuracy is limited.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.
Claims (6)
1. A fast and efficient new method for three-dimensional reconstruction of a mirror surface object surface is characterized by having high reconstruction accuracy without contacting an object and comprises the following steps:
step 1: calibrating the multi-view camera, and unifying a world coordinate system;
step 2: the method comprises the steps of carrying out image processing on a reflection image of a light source template on a reflection target, which is acquired by a multi-view acquisition system, and respectively solving a light source point coordinate, a reflection point coordinate and an optical center coordinate according to an acquired pixel coordinate and a calibration principle of a multi-view camera;
and step 3: according to the reflection principle of light, combining the coordinates of each group of light source points, the coordinates of the reflection points and the coordinates of the optical center to solve and reconstruct the gradient point cloud on the surface of the target;
and 4, step 4: and performing integral calculation according to the gradient point cloud, solving the height point cloud of the object surface, performing smooth filtering treatment, and finally drawing a three-dimensional reconstruction map of the object surface according to the height point cloud.
2. The method of claim 1, wherein the method comprises: by utilizing the geometrical optics principle and carrying out three-dimensional reconstruction on the irregular surface of the mirror surface target through the image processing process, the whole process does not need to contact the object, and the method has very important significance on the safety of the high-precision object.
3. The method of claim 1, wherein the method comprises: for the calibration of the multi-view reconstruction system, the invention adopts a two-by-two calibration method, namely, the camera 1 and the camera 2, the camera 1 and the camera 3, and the camera 1 and the camera 4 are respectively calibrated in a binocular mode; the three sets of calibrated world coordinate systems are all overlapped with the camera coordinate system of the camera 1, so that the world coordinate systems are unified, and the relationship between the cameras 2, 3, 4 and the camera 1 can be obtained through the rotation and translation matrixes obtained by the sets of binocular calibration.
4. The method of claim 1, wherein the method comprises: when the coordinates of the light source points are solved, a high-reflection plane mirror is placed on the object carrying platform, the camera collects the mirror image of the display screen relative to the object carrying surface, the coordinates of the mirror image of the light source points are solved by combining the pixel coordinates of the collected image with the multi-view calibration principle, and the coordinates of the light source points are successfully obtained according to the symmetrical characteristics of the mirror image of the light source points and the light source points about the high-reflection plane mirror.
5. The method of claim 1, wherein the method comprises: when the coordinates of the reflection points are solved, firstly, the height value is preset, then the error is reduced through an iterative approximation method, the iterative approximation algorithm is that the height of the preset reflection points is h1, the world coordinates of the reflection points are solved, the world coordinates of the light center and the light source points are combined to solve the gradient, then the height point cloud is solved through integration, the height of the position in the actually solved height point cloud is h2, h2 is preset as the height of the reflection points, the world coordinates of the reflection points are solved again, the gradient and the height are solved again, the ith calculation height is set as hi, the ith +1 th calculation height is set as hi +1, and the loop is terminated until (hi +1-hi)/hi is less than 0.01.
6. The method of claim 1, wherein the method comprises: according to the light reflection principle, after gradient point cloud is solved by utilizing coordinates of light source points, reflection points and light centers, height point cloud is obtained through integration, and a reconstructed surface image is obtained after smooth filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110341923.3A CN113096232A (en) | 2021-03-30 | 2021-03-30 | Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110341923.3A CN113096232A (en) | 2021-03-30 | 2021-03-30 | Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113096232A true CN113096232A (en) | 2021-07-09 |
Family
ID=76671263
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110341923.3A Pending CN113096232A (en) | 2021-03-30 | 2021-03-30 | Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113096232A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113739718A (en) * | 2021-08-26 | 2021-12-03 | 江南大学 | Object three-dimensional modeling method based on multi-circle fitting |
| CN114255313A (en) * | 2022-02-28 | 2022-03-29 | 深圳星坊科技有限公司 | Three-dimensional reconstruction method and device for mirror surface object, computer equipment and storage medium |
-
2021
- 2021-03-30 CN CN202110341923.3A patent/CN113096232A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113739718A (en) * | 2021-08-26 | 2021-12-03 | 江南大学 | Object three-dimensional modeling method based on multi-circle fitting |
| CN113739718B (en) * | 2021-08-26 | 2022-06-10 | 江南大学 | Object three-dimensional modeling method based on multi-circle fitting |
| CN114255313A (en) * | 2022-02-28 | 2022-03-29 | 深圳星坊科技有限公司 | Three-dimensional reconstruction method and device for mirror surface object, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109000582B (en) | Scanning method and system of tracking type three-dimensional scanning device, storage medium and equipment | |
| CN103499302B (en) | The camshaft diameter dimension On-line Measuring Method of structure based light Vision imaging system | |
| CN109579695B (en) | Part measuring method based on heterogeneous stereoscopic vision | |
| CN113096232A (en) | Novel method for quickly and efficiently reconstructing surface of mirror surface object in three dimensions | |
| CN114111637B (en) | Virtual dual-purpose stripe structured light three-dimensional reconstruction method | |
| CN106952309A (en) | The device and method of Fast Calibration TOF depth camera many kinds of parameters | |
| CN113108721B (en) | High-reflectivity object three-dimensional measurement method based on multi-beam self-adaptive complementary matching | |
| CN112945141B (en) | Structured light rapid imaging method and system based on micro-lens array | |
| CN106500625A (en) | A kind of telecentricity stereo vision measuring apparatus and its method for being applied to the measurement of object dimensional pattern micron accuracies | |
| CN113724337B (en) | Camera dynamic external parameter calibration method and device without depending on tripod head angle | |
| CN111127540A (en) | Automatic distance measurement method and system for three-dimensional virtual space | |
| CN112258586A (en) | Calibration method for stereoscopic vision model parameters of single plane mirror | |
| Yamaguti et al. | A method of distance measurement by using monocular camera | |
| CN116295113A (en) | Polarization three-dimensional imaging method integrating fringe projection | |
| CN113763540A (en) | Three-dimensional reconstruction method and equipment based on speckle fringe hybrid modulation | |
| Meng et al. | Defocused calibration for large field-of-view binocular cameras | |
| CN109712139A (en) | The dimension measurement method of monocular vision based on linear movement mould group | |
| CN116592792A (en) | Measurement method and system for assisting relative phase stereo matching by using speckle | |
| CN110728745A (en) | Underwater binocular stereoscopic vision three-dimensional reconstruction method based on multilayer refraction image model | |
| CN112697065B (en) | Three-dimensional shape reconstruction method based on camera array | |
| Li et al. | A novel method for 3D reconstruction in tight and restricted space based on improved line-structured light | |
| Li et al. | A virtual binocular line-structured light measurement method based on a plane mirror | |
| Fu et al. | A novel phase measuring deflectometry method using a curved screen | |
| Zhu et al. | Rotation axis calibration of laser line rotating-scan system for 3D reconstruction | |
| CN113989105B (en) | Single-camera spherical mirror reflection imaging projection device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |