阅读说明：本技术 一种利用偏振信息对不同材质物体识别的方法 (Method for identifying objects of different materials by using polarization information ) 是由 谭治英 赵宝来 徐孝彬 费仲文 骆敏舟 沙刘力 庞凤麟 赵鸣晖 王子龙 张瀛 高 于 2020-04-20 设计创作，主要内容包括：本发明公开了一种利用偏振信息对不同材质物体识别的方法,包括如下步骤：工控机控制光源点亮,并触发偏振相机采集图像与处理；偏振相机采集物体在0°,45°,90°和135°四个方向的偏振图像,获得四个方向下的偏振强度I<Sub>0</Sub>,I<Sub>45</Sub>,I<Sub>90</Sub>和I<Sub>135</Sub>,然后计算斯托克斯参数S<Sub>0</Sub>、S<Sub>1</Sub>、S<Sub>2</Sub>；利用斯托克斯参数计算偏振强度I、方向角φ和偏振度ρ；基于偏振度ρ与折射率n、天顶角θ的关系推导出折射率n的数学模型,其中天顶角θ近似方位角φ,以方位角φ代入折射率n的数学模型,获得近似折射率<Image he="52" wi="35" file="DDA0002457871540000011.GIF" imgContent="drawing" imgFormat="GIF" orientation="portrait" inline="no"></Image>数学模型；利用聚类算法对近似折射率<Image he="52" wi="45" file="DDA0002457871540000012.GIF" imgContent="drawing" imgFormat="GIF" orientation="portrait" inline="no"></Image>数据进行聚类,实现对不同材质物体的识别。该方法计算精度高、成本低、易推广。(The invention discloses a method for identifying objects of different materials by utilizing polarization information, which comprises the following steps: the industrial personal computer controls the light source to be lightened and triggers the polarization camera to acquire and process images; the polarization camera collects polarization images of an object in four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees to obtain polarization intensities I in the four directions 0 ，I 45 ，I 90 And I 135 Then calculating the Stokes parameter S 0 、S 1 、S 2 (ii) a Calculating the polarization intensity I, the direction angle phi and the polarization degree rho by utilizing the Stokes parameters; deducing a mathematical model of the refractive index n based on the relation between the polarization degree rho and the refractive index n and the zenith angle theta, wherein the zenith angle theta is approximate to the azimuth angle phi, and the azimuth angle phi is substituted into the mathematical model of the refractive index n to obtain the approximate refractive index A mathematical model; using clustering algorithm to approximate refractive index And clustering the data to realize the identification of objects with different materials. The method has the advantages of high calculation precision, low cost and easy popularization.)

1. A method for recognizing objects made of different materials by utilizing polarization information comprises a polarization camera, an industrial lens, a light source and an industrial personal computer, and is characterized in that: the method comprises the following steps:

the industrial personal computer controls the light source to be lightened and triggers the polarization camera to acquire and process images;

the polarization camera collects polarization images of an object in four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees to obtain polarization intensities I in the four directions₀，I₄₅，I₉₀And I₁₃₅By means of I₀，I₄₅，I₉₀And I₁₃₅Calculating the Stokes parameter S₀、S₁、S₂；

Calculating the polarization intensity I, the direction angle phi and the polarization degree rho by utilizing the Stokes parameters;

deducing a mathematical model of the refractive index n based on the relation between the polarization degree rho and the refractive index n and the zenith angle theta, wherein the zenith angle theta is approximate to the azimuth angle phi, and substituting the azimuth angle phi into the mathematical model of the refractive index n instead of the zenith angle theta to obtain the approximate refractive indexA mathematical model;

using clustering algorithm to approximate refractive indexAnd clustering the data to realize the identification of objects with different materials.

2. The method of claim 1, wherein the polarization information is used to identify the different material objects, and the method further comprises: the Stokes parameter S₀、S₁、S₂The calculation process of (2) is as follows:

S₁＝I₀-I₉₀，

S₂＝I₄₅-I₁₃₅，

wherein S is₀Represents the total intensity of light, S₁Represents the difference between linearly polarized light components of 0 DEG and 90 DEG, S₂Representing the difference between the 45 deg. and 135 deg. linearly polarized light components.

3. The method of claim 1, wherein the polarization information is used to identify the different material objects, and the method further comprises: the calculation processes of the polarized light intensity I, the direction angle phi and the polarization degree rho are as follows:

I＝S₀，

4. the method of claim 1, wherein the polarization information is used to identify the different material objects, and the method further comprises: the relation among the polarization degree rho, the refractive index n and the zenith angle theta is as follows:

5. the method of claim 4, wherein the polarization information is used to identify the different material objects, and the method further comprises: the mathematical model of the refractive index n:

wherein the content of the first and second substances,

A＝ρ²cos⁴θ

B＝2ρ²cos²θ(2sin⁴θ-sin²θ)-4sin⁴θcos²θ

C＝ρ²(2sin⁴θ-sin²θ)²+4sin⁶θcos²θ。

6. the method of claim 1, wherein the polarization information is used to identify the different material objects, and the method further comprises: said approximate refractive indexThe mathematical model is as follows:

wherein the content of the first and second substances,

A′＝ρ²cos⁴φ

B′＝2ρ²cos²φ(2sin⁴φ-sin²φ)-4sin⁴φcos²φ

C′＝ρ²(2sin⁴φ-sin²φ)²+4sin⁶φcos²φ。

Technical Field

The invention relates to a machine vision material identification method, in particular to a method for identifying objects of different materials by utilizing polarization information.

Background

With the application and development of machine vision technology, robots or automation equipment have stronger and stronger ability to perform identification by means of vision. However, the conventional visual recognition method relies on a color camera or a black and white camera to acquire an RGB color image or a grayscale image, and recognition of an object is limited to information such as color, texture, intensity and direction of reflected light on the surface of the object, and lacks information expression of material of the object. Although the spectral camera can be used for collecting multiband information of an object and identifying materials, the requirement on the illumination environment is strict and the price is high, so that the spectral camera is not widely applied to robots or automatic equipment.

Disclosure of Invention

The purpose is as follows: in order to overcome the defects in the prior art, the invention provides a method for identifying objects made of different materials by using polarization information, and solves the problems of difficulty in identifying the materials and high cost on the existing robot or automation equipment.

The technical scheme is as follows: in order to solve the technical problems, the technical scheme adopted by the invention is as follows:

a method for recognizing objects made of different materials by utilizing polarization information is disclosed, the used device comprises a polarization camera, an industrial lens, a light source and an industrial personal computer, and the method comprises the following steps:

the industrial personal computer controls the light source to be lightened and triggers the polarization camera to acquire and process images;

the polarization camera collects polarization images of an object in four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees to obtain polarization intensities I in the four directions₀，I₄₅，I₉₀And I₁₃₅By means of I₀，I₄₅，I₉₀And I₁₃₅Calculating the Stokes parameter S₀、S₁、S₂；

Calculating the polarization intensity I, the direction angle phi and the polarization degree rho by utilizing the Stokes parameters;

deducing a mathematical model of the refractive index n based on the relation between the polarization degree rho and the refractive index n and the zenith angle theta, wherein the zenith angle theta is approximate to the azimuth angle phi, and substituting the azimuth angle phi into the mathematical model of the refractive index n instead of the zenith angle theta to obtain the approximate refractive indexA mathematical model;

using clustering algorithm to approximate refractive indexAnd clustering the data to realize the identification of objects with different materials.

Further, the Stokes parameter S₀、S₁、S₂The calculation process of (2) is as follows:

S₁＝I₀-I₉₀，

S₂＝I₄₅-I₁₃₅，

wherein S is₀Represents the total intensity of light, S₁Represents the difference between linearly polarized light components of 0 DEG and 90 DEG, S₂Representing the difference between the 45 deg. and 135 deg. linearly polarized light components.

Further, the calculation process of the polarized light intensity I, the direction angle Φ and the polarization degree ρ is as follows:

I＝S₀，

further, the relation between the polarization degree rho and the refractive index n and the zenith angle theta is as follows:

further, the mathematical model of the refractive index n is:

wherein the content of the first and second substances,

A＝ρ²cos⁴θ

B＝2ρ²cos²θ(2sin⁴θ-sin²θ)-4sin⁴θcos²θ

C＝ρ²(2sin⁴θ-sin²θ)²+4sin⁶θcos²θ。

further, the approximate refractive indexThe mathematical model is as follows:

wherein the content of the first and second substances,

A′＝ρ²cos⁴φ

B′＝2ρ²cos²φ(2sin⁴φ-sin²φ)-4sin⁴φcos²φ

C′＝ρ²(2sin⁴φ-sin²φ)²+4sin⁶φcos²φ。

has the advantages that: compared with the existing machine vision identification method, the method for identifying the objects made of different materials by utilizing the polarization information disclosed by the invention has the following advantages:

the invention provides an object surface approximate refractive index calculated by utilizing a polarization image, and clustering the approximate refractive index data by utilizing a clustering algorithm to provide information for visual identification of objects, so that the objects of different materials can be identified; the approximate refractive index depends on strict mathematical theory derivation, and has the advantages of high calculation precision and high speed; compared with the existing spectral object material identification method, the method for identifying the object by using the polarization information has the characteristics of simplicity, low cost, easiness in popularization and the like, and can effectively meet the requirements of robots or automation equipment.

Drawings

FIG. 1 is a schematic diagram of the calculation process of the approximate refractive index of the identification method of the present invention;

FIG. 2 is a gray scale image normalized for polarization intensity for four comparative examples of the present invention;

FIG. 3 is a gray scale image normalized in azimuth for four comparative embodiments of the present invention;

FIG. 4 is a gray scale image normalized for polarization for four comparative examples of the present invention;

FIG. 5 is a bar graph of the approximate indices of refraction for four comparative examples of the present invention;

FIG. 6 is a graph of approximate refractive index clustering results for four comparative examples of the present invention.

Detailed Description

The present invention will be further described with reference to the accompanying drawings.

The invention provides a method for identifying material by using polarization information of an object surface, the calculation flow of the method is shown in figure 1, and the device used by the method comprises a polarization camera, a lens, a light source and an industrial personal computer. Four groups of objects made of different materials are selected for a comparison test, wherein the background material is metal, (a) the group is a metal block and a wood block, the metal block and the background material are the same, (b) the group is the wood block and a ceramic cup, (c) the group is a red plastic block and a green plastic shell, and (d) the group is the red plastic block and red paper.

After the device is started, the industrial personal computer controls the light source to be lightened, the polarization camera is triggered to collect polarization images of each group of objects in four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees, and light intensity I in the four directions is obtained according to gray level calculation of the polarization images₀，I₄₅，I₉₀And I₁₃₅According to the polarization intensity I in four directions₀，I₄₅，I₉₀And I₁₃₅Calculating the Stokes parameter S₀、S₁、S₂(ii) a The polarization intensity I, the direction angle phi and the polarization degree rho are calculated by utilizing the Stokes parameters.

The polarized image is normalized to the polarized intensity I as a gray scale image, i.e., as shown in fig. 2.

And (3) normalizing the radian (-pi, pi) gray scale at each pixel of the polarization image to 0-255 to obtain the image shown in FIG. 3.

The degree of polarization is the ratio of the light intensity of the polarized part in the light beam to the whole light intensity, and the value range is [0,1], and the ratio is scaled up to the range [0,255], that is, the image shown in fig. 4 is obtained by standardized gray scale processing.

Deducing a mathematical model of the refractive index n based on the relation between the polarization degree rho and the refractive index n and the zenith angle theta, wherein the zenith angle theta is approximate to the azimuth angle phi, and substituting the azimuth angle phi into the mathematical model of the refractive index n instead of the zenith angle theta to obtain the approximate refractive indexA mathematical model;

wherein the content of the first and second substances,

A′＝ρ²cos⁴φ

B′＝2ρ²cos²φ(2sin⁴φ-sin²φ)-4sin⁴φcos²φ

C′＝ρ²(2sin⁴φ-sin²φ)²+4sin⁶φcos²φ。

the results of the approximate refractive indices of the four comparative examples are shown in fig. 5, in which the ordinate represents the approximate refractive index. As can be seen from fig. 5, the similar refractive indexes of objects having the same material are similar even if the colors are different, and the similar refractive indexes of objects having different materials are greatly different even if the colors are the same, so that it is possible to identify objects having different materials based on the similar refractive indexes.

Using clustering algorithm to approximate refractive indexAnd clustering the data to realize the identification of objects with different materials. As shown in fig. 6, the result of clustering the approximate refractive index by using the clusterdata clustering function of matlab is shown.

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.