CN116067891A - Meat marinating on-line detection system based on polarization image sensing and application thereof - Google Patents

Meat marinating on-line detection system based on polarization image sensing and application thereof Download PDF

Info

Publication number
CN116067891A
CN116067891A CN202310194546.4A CN202310194546A CN116067891A CN 116067891 A CN116067891 A CN 116067891A CN 202310194546 A CN202310194546 A CN 202310194546A CN 116067891 A CN116067891 A CN 116067891A
Authority
CN
China
Prior art keywords
meat
image
polarization
polarization degree
marinating
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
Application number
CN202310194546.4A
Other languages
Chinese (zh)
Inventor
汪莎
蒲永杰
朱杰
向丹
李琴
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan University
Original Assignee
Sichuan University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Sichuan University filed Critical Sichuan University
Priority to CN202310194546.4A priority Critical patent/CN116067891A/en
Publication of CN116067891A publication Critical patent/CN116067891A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/01—Arrangements or apparatus for facilitating the optical investigation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/21—Polarisation-affecting properties
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/27—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands using photo-electric detection ; circuits for computing concentration
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/01—Arrangements or apparatus for facilitating the optical investigation
    • G01N2021/0106—General arrangement of respective parts
    • G01N2021/0112—Apparatus in one mechanical, optical or electronic block
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N2021/1765—Method using an image detector and processing of image signal
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N2021/1765—Method using an image detector and processing of image signal
    • G01N2021/177—Detector of the video camera type
    • G01N2021/1776—Colour camera
    • 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
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
    • Y02A40/90—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in food processing or handling, e.g. food conservation

Landscapes

  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)

Abstract

The invention discloses a meat marinating on-line detection system based on polarized image sensing and application thereof, wherein the complete system comprises a sample preparation device, a meat marinating device, a polarized light generation device, an image sensing detection device and a data processing module; the preparation of the sample and the meat marinating device acquire a measured sample, are hung on a three-dimensional displacement platform and are marinated in an iron pan; the polarized light generating device generates linearly polarized light; the image sensing detection device is used for imaging a target object; the data processing module comprises a polarization degree calculation module, a texture feature extraction module, a color information extraction module, a simple meat identification module, a database module and a meat boiling time calculation module, and is mainly used for obtaining the relationship between DOP, ASM, H, CON, COR and A and the boiling time and the relationship after 6 parameters are subjected to multi-element fitting. The intelligent meat cooking system has the advantages of high reliability of obtained data, simple algorithm, realization of intelligent identification of simple meat and judgment of cooking degree at one time, and high measurement efficiency.

Description

Meat marinating on-line detection system based on polarization image sensing and application thereof
Technical Field
The invention relates to a meat marinating on-line detection system based on polarized image sensing and application thereof, belonging to the field of food safety and processing.
Background
The intelligent judgment of meat marinating is greatly helpful for people with busy work or insufficient life experience. Most importantly, the unified judgment standard can reduce the probability of misjudgment. The quality of food depends on a number of qualities, these parameters being divided into external and internal qualities. Internal quality includes moisture, sugar, ph, vitamins, internal defects, etc. External qualities include size, shape, color, odor, surface, hardness, softness, mouthfeel, and the like. In the traditional quality detection, the external quality is mainly detected manually, and the result has a manual error; the internal quality is mainly detected by physical and chemical detection, the result is accurate, but the cost is high, the detection object is required to be destroyed, and the process is tedious and time-consuming. The main method comprises the following steps: chromatography, spectrophotometry and titration. In recent years, many industries are looking for better methods to evaluate food quality. The nondestructive testing technology is a new technology which can help to improve the quality of food and has good development prospect. The quality nondestructive detection refers to a detection process for analyzing and acquiring food quality by utilizing physical characteristics of fruits such as sound, electricity, light, magnetism and the like and by integrating detection technologies such as spectral imaging, dielectric characteristics, nuclear magnetic resonance and the like on the premise of not damaging a detection object. The main specific methods at present are an electronic nose method, a spectrum method and a polarized image sensing method. The electronic nose method is most widely used, but is affected by the smell of other foods. The result of the spectrometry is accurate, but the light receiving needs an expensive probe, which is not beneficial to popularization. The polarization image sensing experimental device is simple, low in price and suitable for popularization in the food industry.
Disclosure of Invention
The invention aims to provide a meat marinating on-line detection system based on polarized image sensing and application thereof, wherein the meat doneness recognition system uses polarized light as a light source, and has the advantages of low cost, high reliability, simple algorithm and high recognition efficiency.
The invention discloses a meat marinating on-line detection system based on polarized image sensing, which comprises a sample selection and preparation device, a meat marinating device, a polarized light generating device, an image sensing detection device and a data processing module.
The sample selection and preparation process involves purchasing fresh meat in the fresh meat market and then rapidly transporting to the laboratory with a refrigerator. The surface fascia and fat were removed after the sample was placed in the console. The measured samples were then extracted with meat samplers at equal intervals and equal thickness.
The meat marinating device comprises an iron pot capable of controlling temperature stability, a set of three-dimensional displacement platform capable of directly fishing out meat blocks, and the like.
The polarized light generating device comprises a light source and a polaroid, and is used for generating polarized light to irradiate an area where meat to be measured is placed.
The image sensing detection device comprises a CMOS camera and a polaroid, and is arranged in front of the sample at a certain angle with polarized light and used for imaging the sample.
The data processing module comprises a polarization degree calculation module, a texture feature extraction module, a color feature information extraction module, a simple meat identification module, a database module and a meat boiling time calculation module.
The degree of polarization calculation module calculates a degree of polarization (DOP) based primarily on the stokes. The specific method comprises the following steps: the polarizing plate 1 for generating polarized light is set to 0 degrees, and the polarizing plate 2 placed in front of the CMOS camera is set to 0 degrees, 45 degrees, 90 degrees, and 135 degrees, respectively, to thereby obtain angles corresponding to the respective figures. Then, gray values are calculated for gray images of different angles shot by the CMOS camera, and the light intensity I corresponding to the image is obtained 0° 、I 45° 、I 90° And I 135° . And then calculating the polarization degree value of the light intensity corresponding to the 4 angles according to the stokes. If the picture shot by the CMOS camera is a color picture, three channels are separated by adopting RGB space, and then the polarization degree value is calculated.
The texture feature extraction module adopts a traditional extraction method, namely a gray level co-occurrence matrix (GLCM). Since the polarization degree map contains abundant texture information, it is necessary to obtain the polarization degree map. The specific method for calculating the polarization degree map is as follows: the polarizer 2 sets the polarization degree value of the corresponding pixel according to the stokes by the pixels of the pictures shot at different angles, and then the polarization degree map can be obtained. GLCM is defined as the joint probability density of two position pixels, and the matrix may reflect the position distribution characteristics and the luminance distribution characteristics between two pixels with close luminance. The method comprises the steps of equally taking 4 areas on a polarization degree image, extracting parameters such As Second Moment (ASM), entropy (H), contrast (CON), correlation value (COR) and the like from the 4 images, and then averaging 4 characteristic parameters of the 4 images to further reflect the doneness of meat.
The color characteristic information extraction module is used for reflecting the reflectivity and the scattering rate of meat by using the brightness average value (A) of an image according to the color change of the meat in the marinating process, so that the spectral reflectivity or the scattering rate is different, and the doneness of the meat is further reflected.
The simple meat identification module divides a picture obtained by the CMOS camera so as to obtain an image of a target sample, and then identifies the type of the target sample.
The database module stores ASM, H, CON, COR, A and DOP fitted curves of different target samples and meat boiling time. Then, final multi-fitting can be performed according to the transformation of the parameters, and a final fitting curve is obtained.
The meat boiling time calculation module calculates the boiling time of the target object according to 6 parameters of the food to be detected or the multi-fitting result. The specific method comprises the following steps: the target object boiling time can be predicted by calling a fitting curve or a multielement fitting curve of 6 parameters and time of a database, and then calculating the target object boiling time according to the 6 parameters.
Compared with the prior art, the invention has the advantages that: according to the invention, polarized light is used as a light source, a polarization degree value and a polarization degree image of a target object and time can be obtained under the condition of adjusting the angle of the polaroid, the polarization degree image contains abundant texture and reflection information, and further, gray level co-occurrence matrix and color characteristic information extraction analysis are carried out on the polarization degree image, and finally, the relation of 6 parameters along with time can be obtained. These analytical methods have a great effect on the accuracy of the final results. The invention effectively reduces the cost and the difficulty of data processing, and also simplifies the complexity of the polarized light generating device and the image sensing detection device, and has the advantages of high identification processing running speed, high reliability and simple algorithm. The system comprises a sample selection and preparation device, a meat marinating device, a polarized light generation device, an image sensing detection device and a data processing module, and can realize intelligent identification of the target object and call a judgment criterion at one time, so that the detection efficiency is greatly improved, and the judgment accuracy is also greatly improved.
Further, the polarization degree and polarization degree image calculation module described in the present invention obtains an image of the target object based on the adjustment of the angles of the polarizer (polarizer 1) and the analyzer (polarizer 2). If the gray level image is the gray level image, the brightness value of each angle image is directly obtained, and then the polarization degree and the polarization state image are obtained according to the stokes. If the image is a color image, firstly, three channels are separated according to RGB space, then, the light intensity I is obtained for the image of each channel, and finally, the polarization degree and the polarization degree image are calculated.
Furthermore, the angle difference between the target object and the analyzer in the invention is-1 to 1 degree to obtain I 0° The area with the angle of 44-46 degrees different from the analyzer can obtain I 45° The area with the angle of 89-91 degrees with the analyzer can obtain I 90° I can be obtained in a region with an angle of 134-136 DEG from the analyzer 135° 。
Further, when a polarized light beam is directed onto meat, the specularly reflected light preserves the polarization of the incident light. Whereas diffuse reflection changes the polarization state of incident light into partially polarized light. Partially polarized light is characterized by a polarized component in unpolarized light. The amount of polarization in the multiply scattered light is represented by Stokes vectors. The specific calculation process is as follows:
first the stokes component of the polarized light scattered by the food to be measured is calculated from the four light intensities,
S 0 ＝I 0 +I 90
S 1 ＝I 0 -I 90
S 2 ＝I 45 -I 135
DOP, as well as the polarization degree image, can then be found from the aforementioned stokes components.
Figure BDA0004106663410000041
Further, the light source usable in the present invention is various, and may be an LED white light source, a helium-neon laser, a semiconductor laser, a solid-state laser, a fiber laser, a vortex laser, or a halogen lamp.
Furthermore, the number of the polarized light generating devices and the number of the image sensing detecting devices can be N (different directions), the devices for generating the polarized light irradiate the target object during measurement, the image sensor device is used for placing the target object area at a certain angle for imaging, and the data processing module is used for processing the data of the pictures obtained by the image sensor to obtain the linear relation between 6 parameters and time.
Further, the present invention generally divides the photographed image into a color image and a gray image according to the CMOS camera. The pixels of a color map are made up of a plurality of color components, which are channels. Typically, an image can be decomposed into red (R), green (G), and blue (B) according to the RGB space. The pixels of the gray-scale image are made up of one color, typically luminance. The processing of the method is performed before the data processing according to the different image types. The specific process is as follows: if the color image is the color image, firstly, carrying out channel separation processing, then calculating the polarization degree and the polarization degree image, and then carrying out gray level co-occurrence matrix processing and color characteristic information extraction. If the gray level image is the gray level image, the brightness value is directly calculated, then the polarization degree and the polarization degree image are calculated, and then the gray level co-occurrence matrix is processed and the color characteristic information is extracted.
Further, the gray level co-occurrence matrix (GLCM) is one of the conventional texture extraction methods. The texture of the object is formed by the repeated occurrence of gray scale distribution, so that two pixels separated by a certain distance in the image necessarily have a gray scale relationship. The comprehensive information of the image gray scale in the direction, the interval and the variation amplitude can be obtained by statistically analyzing the correlation characteristics of the pixel point in the direction and the distance.
Further, one phenomenon often occurs in the present invention, which is often associated with a variety of factors. The optimal combination of multiple independent variables (denoted as X1, X2, … Xn) can predict or estimate the dependent variable (Y) more efficiently and practically than using only one independent variable for prediction or estimation. Therefore, it is necessary to establish a linear regression equation for a plurality of independent and dependent variables. The principle of the multiple linear regression equation is:
Y＝a 0 +a 1 X 1 +a 2 X 2 +a 3 X 3 +K+a m X m ,
a 0 is a constant term, a 1 、a 2 、a n Is a regression coefficient. If there are N observations in the experiment, Y i ,X 1i ，…X ni (i=1, 2, …, n). The relationship is shown as follows.
Y i ＝a 0 +a 1 X 1i +a 2 X 2i +a 3 X 3i +Λ+a m X mi ,
The matrix is represented as follows:
Figure BDA0004106663410000051
the formula can be written as:
Y＝aX
the purpose of the regression equation is to solve for the coefficients and constant terms of the independent variables. The more arguments, the more complex the computation.
In the invention, the automatic process automatically compares the data with the existing samples in the database, so as to judge whether the samples are cooked or not.
The invention further discloses application of the meat marinating online detection system based on the polarization image sensing to judgment of the doneness in the meat marinating process. The meat doneness recognition system is arranged in a food processing workshop, the polarized light generating device and the image detection device are arranged beside an iron pot for cooking meat, meat at equal time intervals in the iron pot is quickly taken out and placed on an operation platform, when the meat doneness recognition system works, light emitted by a light source is changed into polarized light after passing through a polarizer, the polarized light irradiates on meat blocks on the operation platform, and the image sensing device images a target object. The data processing module processes the image to obtain the boiling time of the target object, and the recognition of the maturity of the target object is realized.
The present invention will be described in further detail with reference to the following detailed description and the accompanying drawings.
Drawings
FIG. 1 is a schematic diagram of unpolarized light, partially polarized light, and linearly polarized light of the present invention.
FIG. 2 is a flow chart of an automated determination of doneness in the present invention;
FIG. 3 is a schematic diagram showing the overall structure of the meat doneness recognition system according to the present invention.
Fig. 4 is a schematic diagram of the overall structure of a meat doneness recognition system according to an embodiment of the invention.
Detailed Description
A meat marinating on-line identification system based on polarized image detection comprises a sample selection and preparation device, a meat marinating device, a polarized light generation device, an image sensing detection device and a data processing module.
The sample selection and preparation process involves purchasing fresh meat in the fresh meat market and then rapidly transporting to the laboratory with a refrigerator. The surface fascia and fat were removed after the sample was placed in the console. The measured samples were then extracted with meat samplers at equal intervals and equal thickness.
The meat marinating device comprises an electric iron pot capable of controlling temperature stability, a set of three-dimensional displacement platform capable of directly fishing out meat blocks, and the like.
The polarized light generating device comprises a light source and a polaroid, and is used for generating polarized light to irradiate an area where meat to be measured is placed. The light generated by a typical light source is a non-uniformly polarized light that varies with spatial distribution. Fig. 1 is a schematic diagram of polarized light. 1.1 is unpolarized light, 1.2 is partially polarized light, and 1.3 is linearly polarized light.
The image sensing detection device comprises a CMOS camera and a polaroid, and is arranged in front of the sample at a certain angle with polarized light and used for imaging the sample.
The data processing module comprises a polarization degree calculation module, a texture feature extraction module, a color feature information extraction module, a simple meat identification module, a database module and a meat boiling time calculation module.
The degree of polarization calculation module calculates a degree of polarization (DOP) based primarily on the stokes. The specific method comprises the following steps: the polarizing plate 1 for generating polarized light is set to 0 degrees, and the polarizing plate 2 placed in front of the CMOS camera is set to 0 degrees, 45 degrees, 90 degrees, and 135 degrees, respectively, to thereby obtain angles corresponding to the respective figures. Then, gray values are calculated for gray images of different angles shot by the CMOS camera, and the light intensity I corresponding to the image is obtained 0° 、I 45° 、I 90° And I 135° . And then calculating the polarization degree value of the light intensity corresponding to the 4 angles according to the stokes. If the picture shot by the CMOS camera is a color picture, three channels are separated by adopting RGB space, and then the polarization degree value is calculated.
The texture feature extraction module adopts a traditional extraction method, namely a gray level co-occurrence matrix (GLCM). Since the polarization degree map contains abundant texture information, it is necessary to obtain the polarization degree map. The specific method for calculating the polarization degree map is as follows: the polarizer 2 sets the polarization degree value of the corresponding pixel according to the stokes by the pixels of the pictures shot at different angles, and then the polarization degree map can be obtained. GLCM is defined as the joint probability density of two position pixels, and the matrix may reflect the position distribution characteristics and the luminance distribution characteristics between two pixels with close luminance. The method comprises the steps of equally taking 4 areas on a polarization degree image, extracting parameters such As Second Moment (ASM), entropy (H), contrast (CON), correlation value (COR) and the like from the 4 images, and then averaging 4 characteristic parameters of the 4 images to further reflect the doneness of meat.
The color characteristic information extraction module is used for reflecting the reflectivity and the scattering rate of meat by using the brightness average value (A) of an image according to the color change of the meat in the marinating process, so that the spectral reflectivity or the scattering rate is different, and the doneness of the meat is further reflected.
The simple meat identification module divides a picture obtained by the CMOS camera so as to obtain an image of a target sample, and then identifies the type of the target sample.
The database module stores ASM, H, CON, COR, A and DOP fitted curves of different target samples and meat boiling time. Then, final multi-fitting can be performed according to the transformation of the parameters, and a final fitting curve is obtained.
The meat boiling time calculation module calculates the boiling time of the target object according to 6 parameters of the food to be detected or the multi-fitting result. The specific method comprises the following steps: the target object boiling time can be predicted by calling a fitting curve or a multielement fitting curve of 6 parameters and time of a database, and then calculating the target object boiling time according to the 6 parameters.
Preferably, the polarization degree calculation module and the polarization degree diagram are calculated according to the stokes. The specific method is as follows: the angle of the polarizer (polarizing plate 1) was always maintained at 0 degrees in order to change unpolarized light into linearly polarized light. The angle of the analyzer (polarizing plate 2) may be set to 0 degrees, 45 degrees, 90 degrees, and 135 degrees. If the CMOS camera obtains the gray level map, the light intensity I corresponding to each angle can be directly obtained, then the polarization degree is obtained according to the stokes, and the pixel corresponding to each map is calculated according to the stokes to obtain the polarization degree map. If the CMOS camera obtains a color map, three channels can be separated by adopting RGB space, and then the polarization degree and the polarization degree map corresponding to each channel can be obtained according to the same processing method.
More preferably, the target is at an angle of-1 DEG to 1 DEG from the analyzer to obtain I 0° The area with the angle of 44-46 degrees different from the analyzer can obtain I 45° The area with the angle of 89-91 degrees with the analyzer can obtain I 90° I can be obtained in a region with an angle of 134-136 DEG from the analyzer 135° 。
Preferably, when a polarized light beam is directed onto the meat, the specularly reflected light preserves the polarization of the incident light. Whereas diffuse reflection changes the polarization state of incident light into partially polarized light. Partially polarized light is characterized by a polarized component in unpolarized light. The amount of polarization in the multiply scattered light is represented by Stokes vectors. The specific calculation process is as follows:
first the stokes component of the polarized light scattered by the food to be measured is calculated from the four light intensities,
S 0 ＝I 0 +I 90
S 1 ＝I 0 -I 90
S 2 ＝I 45 -I 135
DOP, as well as the polarization degree image, can then be found from the aforementioned stokes components.
Figure BDA0004106663410000081
Preferably, the on-line meat marinating detection system based on polarized image sensing is characterized in that the light source is of various types and can be an LED white light source, a helium-neon laser, a semiconductor laser, a solid laser, an optical fiber laser, a vortex laser and a halogen lamp.
Preferably, the number of the polarized light generating devices and the number of the image sensing detecting devices can be N (different directions), the devices for generating the polarized light are irradiated to the target object during measurement, the image sensor device is placed in a target object area at a certain angle for imaging, and the data processing module is used for processing the data of the pictures obtained by the image sensor to obtain the linear relation between 6 parameters and time.
Preferably, the image photographed by the CMOS camera is generally divided into a color map and a gray map. The pixels of a color map are made up of a plurality of color components, which are channels. Typically, an image can be decomposed into red (R), green (G), and blue (B) according to the RGB space. The pixels of the gray-scale image are made up of one color, typically luminance. The processing of the method is performed before the data processing according to the different image types. The specific process is as follows: if the color image is the color image, firstly, carrying out channel separation processing, then calculating the polarization degree and the polarization degree image, and then carrying out gray level co-occurrence matrix processing and color characteristic information extraction. If the gray level image is the gray level image, the brightness value is directly calculated, then the polarization degree and the polarization degree image are calculated, and then the gray level co-occurrence matrix is processed and the color characteristic information is extracted.
Preferably, a gray level co-occurrence matrix (GLCM) is one of the conventional texture extraction methods. The texture of the object is formed by the repeated occurrence of gray scale distribution, so that two pixels separated by a certain distance in the image necessarily have a gray scale relationship. The comprehensive information of the image gray scale in the direction, the interval and the variation amplitude can be obtained by statistically analyzing the correlation characteristics of the pixel point in the direction and the distance.
Preferably, a phenomenon is generally associated with a variety of factors. Optimal combination of multiple independent variables (denoted as X 1 ,X 2 ，…X n ) The dependent variable (Y) can be predicted or estimated, which is more efficient and practical than using only one independent variable for prediction or estimation. Therefore, it is necessary to establish a linear regression equation for a plurality of independent and dependent variables. The principle of the multiple linear regression equation is:
Y＝a 0 +a 1 X 1 +a 2 X 2 +a 3 X 3 +K+a m X m ,
a 0 is a constant term, a 1 、a 2 、a n Is a regression coefficient. If there are N observations in the experiment, Y i ,X 1i ，…X ni (i=1, 2, …, n). The relationship is shown as follows.
Y i ＝a 0 +a 1 X 1i +a 2 X 2i +a 3 X 3i +Λ+a m X mi ,
The matrix is represented as follows:
Figure BDA0004106663410000091
the formula can be written as:
Y＝aX
the purpose of the regression equation is to solve for the coefficients and constant terms of the independent variables. The more arguments, the more complex the computation.
In the invention, the automatic process automatically compares the data of the existing sample with the database to further judge whether the sample is cooked, and the automatic flow is shown in figure 2.
The meat marinating online identification system based on polarized image detection can be applied to food processing detection and is used for directly detecting marinating degree of meat in a production workshop, and a complete experimental device is shown in fig. 3.

Claims (11)

1. A meat marinating on-line detection system based on polarization image sensing is characterized in that: the system comprises a sample selection and preparation device, a meat marinating device, a polarized light generating device, an image sensing detection device and a data processing module;
the sample selection and preparation process comprises purchasing fresh meat in fresh meat market, then rapidly transferring to laboratory by displacing refrigerator, and then placing sample on operation table to remove fascia and fat on surface; then, extracting a measured sample with equal space and equal thickness by using a meat sampler;
the meat marinating device comprises an iron pot capable of controlling the temperature, a set of three-dimensional displacement platform capable of directly fishing out meat blocks and the like;
the polarized light generating device comprises a light source and a polaroid, and is used for generating polarized light and irradiating a region for placing meat to be detected;
the image sensing detection device comprises a CMOS camera and a polaroid, forms a certain angle with polarized light and is arranged in front of the sample and is used for imaging the sample;
the data processing module comprises a polarization degree calculation module, a texture feature extraction module, a color feature information extraction module, a simple meat identification module, a database module and a meat boiling time calculation module;
the polarization degree calculation module is used for calculating the polarization Degree (DOP) mainly according to the stokes; the specific method comprises the following steps: setting the polarizer 1 for generating polarized light at 0 degree, setting the polarizer 2 in front of CMOS camera at 0 degree, 45 degrees, 90 degrees and 135 degrees respectively to obtain images with corresponding angles, and calculating gray values of gray images of different angles to obtain imagesCorresponding light intensity I 0° 、I 45° 、I 90° And I 135° Calculating the polarization degree value of the light intensity corresponding to the 4 angles according to the stokes, and if the picture shot by the CMOS camera is a color picture, adopting RGB space to separate three channels and then calculating the polarization degree value respectively;
the texture feature extraction module adopts a traditional extraction method, namely a gray level co-occurrence matrix (GLCM), and the polarization degree diagram needs to be obtained because the polarization degree diagram contains rich texture information; the specific method for calculating the polarization degree map is as follows: the polaroid 2 sets the polarization degree value of the corresponding pixel calculated by the pixels of the pictures shot at different angles according to the stokes, and a polarization degree diagram can be obtained; GLCM is defined as the joint probability density of two position pixels, the matrix can reflect the position distribution feature and brightness distribution feature between two pixels with close brightness; taking 4 areas on the polarization degree image at equal intervals, extracting parameters such as A Second Moment (ASM), entropy (H), contrast (CON), a correlation value (COR) and the like from the 4 images, and averaging 4 characteristic parameters of the 4 images to further reflect the doneness of meat;
the color characteristic information extraction module is used for reflecting the reflectivity and the scattering rate of meat by using the brightness average value (A) of an image according to the color change of the meat in the marinating process, so that the doneness of the meat is reflected;
the simple meat identification module segments the image obtained by the CMOS camera so as to obtain an image of a target sample, and then identifies the type of the target sample;
the database module stores fitting curves of ASM, H, CON, COR, A and DOP of different target samples and meat boiling time, and then final multi-fitting can be carried out according to the transformation of the parameters, and a final fitting curve is obtained;
the meat boiling time calculation module calculates the boiling time of the target object according to 6 parameters of food to be detected or the multi-fitting result; the specific method comprises the following steps: the target object boiling time can be predicted by calling a fitting curve or a multielement fitting curve of 6 parameters and time of a database, and then calculating the target object boiling time according to the 6 parameters.
2. The polarization image sensing-based meat marinating online detection system according to claim 1, wherein: the polarization degree calculation module and the polarization degree graph are calculated according to the stokes; the specific method is as follows: the angle of the polarizer (the polaroid 1) is always kept at 0 degree, so that unpolarized light is changed into linearly polarized light, the angle of the polarization analyzer (the polaroid 2) can be set to 0 degree, 45 degrees, 90 degrees and 135 degrees, if a gray scale image is obtained by a CMOS camera, the light intensity I corresponding to each angle can be directly obtained, then the polarization degree is obtained according to the stokes, the polarization degree image can be obtained by calculating the pixel corresponding to each image according to the stokes, if a color image is obtained by the CMOS camera, three channels can be separated by adopting RGB space, and then the polarization degree and the polarization degree image corresponding to each channel can be obtained according to the same processing method.
3. The polarization image sensing-based meat marinating online detection system according to claim 2, wherein: the angle difference between the target object and the analyzer is-1 to 1 degree to obtain I 0° The area with the angle of 44-46 degrees different from the analyzer can obtain I 45° The area with the angle of 89-91 degrees with the analyzer can obtain I 90° I can be obtained in a region with an angle of 134-136 DEG from the analyzer 135° 。
4. The polarization image sensing-based meat marinating online detection system according to claim 1, wherein:
when light with polarization is irradiated on meat, the light reflected by the mirror surface maintains the polarization state of incident light, diffuse reflection changes the polarization state of the incident light, and becomes partial polarized light, the partial polarized light is characterized by polarized components in unpolarized light, and the polarization amount in multiple scattered light is expressed by Stokes vectors, and the specific calculation process is as follows:
first the stokes component of the polarized light scattered by the food to be measured is calculated from the four light intensities,
S 0 ＝I 0° +I 90°
S 1 ＝I 0v -I 90v
S 2 ＝I 45° -I 135°
DOP and polarization degree images can then be obtained from the following Stokes components
Figure FDA0004106663400000031
5. The on-line detection system for meat marinating based on polarized image sensing according to claim 1, wherein the light source is of various kinds, such as an LED white light source, a helium-neon laser, a semiconductor laser, a solid state laser, a fiber laser, a vortex laser and a halogen lamp.
6. The polarization image sensing-based meat marinating online detection system according to claim 1, wherein: the polarized light generating device and the image sensing detecting device can be N (different directions), the device for generating the polarized light irradiates the target object during measurement, the image sensor device is used for placing the target object area at a certain angle for imaging, and the data processing module is used for processing the data of the pictures obtained by the image sensor to obtain the linear relation between 6 parameters and time.
7. The polarization image sensing-based meat marinating online detection system according to claim 1, wherein the image photographed by the CMOS camera is generally divided into a color image and a gray image; the pixels of the color map are made up of a plurality of color components, which are channels; typically an image can be decomposed into red (R), green (G), blue (B) according to RGB space; the pixels of the gray scale map are made up of one color, typically luminance; processing of the method is performed according to different image types before data processing; the specific process is as follows:
if the color image is the color image, firstly carrying out channel separation, then calculating the polarization degree and the polarization degree image, and then carrying out gray level co-occurrence matrix processing and color characteristic information extraction;
if the gray level image is the gray level image, the brightness value is directly calculated, then the polarization degree and the polarization degree image are calculated, and then the gray level co-occurrence matrix is processed and the color characteristic information is extracted.
8. The polarization image sensing-based meat marinating online detection system of claim 1, wherein a gray level co-occurrence matrix (GLCM) is one of the conventional texture extraction methods; the texture of the target is formed by repeatedly appearing gray distribution, so that two pixels which are separated by a certain distance in the image have a gray relation, and the comprehensive information of the gray of the image in the directions, the intervals and the variation amplitude can be obtained by statistically analyzing the correlation characteristics of the pixel point directions and the distances.
9. The polarization image sensing-based meat marinating online detection system of claim 1, wherein a phenomenon is generally related to a plurality of factors; optimal combination of multiple independent variables (denoted as X 1 X2, … Xn) can predict or estimate the dependent variable (Y), which is more efficient and practical than predicting or estimating using only one independent variable; therefore, it is necessary to establish a linear regression equation for a plurality of independent and dependent variables; the principle of the multiple linear regression equation is:
Y＝a 0 +a 1 X 1 +a 2 X 2 +a 3 X 3 ++a m X m ,
a 0 is a constant term, a 1 、a 2 、a n Is a regression coefficient, if there are N observations in the experiment, Y i ,X 1i ，…X ni (i=1, 2, …, n), then the relationship is as follows:
Y i ＝a 0 +a 1 X 1i +a 2 X 2i +a 3 X 3i ++a m X mi ,
the matrix is represented as follows:
Figure FDA0004106663400000041
the formula can be written as:
Y＝aX
the purpose of the regression equation is to solve for the coefficients and constant terms of the independent variables, the more complex the calculation.
10. The polarization image sensing-based on-line detection system for marinating meat according to claim 1, wherein the detection system is characterized in that the detection system automatically compares the data of the existing samples in the database to determine whether the samples are boiled.
11. Use of a meat marinating on-line detection system based on polarized image sensing as claimed in any one of claims 1-10 in food processing.
CN202310194546.4A 2023-03-03 2023-03-03 Meat marinating on-line detection system based on polarization image sensing and application thereof Pending CN116067891A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310194546.4A CN116067891A (en) 2023-03-03 2023-03-03 Meat marinating on-line detection system based on polarization image sensing and application thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310194546.4A CN116067891A (en) 2023-03-03 2023-03-03 Meat marinating on-line detection system based on polarization image sensing and application thereof

Publications (1)

Publication Number Publication Date
CN116067891A true CN116067891A (en) 2023-05-05

Family

ID=86180265

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310194546.4A Pending CN116067891A (en) 2023-03-03 2023-03-03 Meat marinating on-line detection system based on polarization image sensing and application thereof

Country Status (1)

Country Link
CN (1) CN116067891A (en)

Similar Documents

Publication Publication Date Title
Pedreschi et al. Computer vision classification of potato chips by color
Kucheryavskiy et al. Determination of fat and total protein content in milk using conventional digital imaging
TWI668436B (en) Systems and methods for enhancing inspection sensitivity of an inspection tool
KR20100015628A (en) Lumber inspection method, device and program
CN104049068B (en) The non-destructive determination device of fresh poultry meat freshness and assay method
WO2019151393A1 (en) Food inspection system, food inspection program, food inspection method and food production method
CN112308832A (en) Bearing quality detection method based on machine vision
US11193756B2 (en) Object shape measurement apparatus and method
Xiao et al. An improved method of detecting pork freshness based on computer vision in on-line system
US5208870A (en) Image inspection methods and apparatus
JP2019168388A (en) Image inspection method and image inspection device
CN102565044A (en) Method for detecting pesticide residue by adopting rapid-determination card on basis of image processing
CN101207834A (en) Method for checking digital camera articulation
WO2015198401A1 (en) Method for setting inspection condition for fastener element, and method for inspecting fastener element
Feng et al. Detection of blood spots in eggs by hyperspectral transmittance imaging
WO2021195817A1 (en) Method for extracting spectral information of object to be detected
CN116067891A (en) Meat marinating on-line detection system based on polarization image sensing and application thereof
CN114565563B (en) Color steel plate surface abnormity detection method based on artificial intelligence
JP7411155B2 (en) Color unevenness inspection device and color unevenness inspection method
JP2005315748A (en) Data compression method, defect inspection method, and defect inspection device
TWI447382B (en) Wood inspection methods and apparatus and program
De Kerf et al. Quantitative detection of corrosion minerals in carbon steel using shortwave infrared hyperspectral imaging
CN117288692B (en) Method for detecting tannin content in brewing grains
CN210533976U (en) Kiwi fruit maturity fast discrimination device
CN108444798A (en) A kind of beef adulteration detection method based on Biospeckles and the moment of inertia spectrum analysis

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