CN114660112B - Method for detecting temperature barrier property of recycled polyester material - Google Patents

Method for detecting temperature barrier property of recycled polyester material Download PDF

Info

Publication number
CN114660112B
CN114660112B CN202210571598.4A CN202210571598A CN114660112B CN 114660112 B CN114660112 B CN 114660112B CN 202210571598 A CN202210571598 A CN 202210571598A CN 114660112 B CN114660112 B CN 114660112B
Authority
CN
China
Prior art keywords
image
standard
polyester material
barrier property
detecting
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
Application number
CN202210571598.4A
Other languages
Chinese (zh)
Other versions
CN114660112A (en
Inventor
董冲
李沅鸿
宋厚春
王洋
陈家磊
王会军
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.)
Henan Yuanhong Polymer New Materials Co ltd
Original Assignee
Henan Yuanhong Polymer New Materials Co ltd
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 Henan Yuanhong Polymer New Materials Co ltd filed Critical Henan Yuanhong Polymer New Materials Co ltd
Priority to CN202210571598.4A priority Critical patent/CN114660112B/en
Publication of CN114660112A publication Critical patent/CN114660112A/en
Application granted granted Critical
Publication of CN114660112B publication Critical patent/CN114660112B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N25/00—Investigating or analyzing materials by the use of thermal means
    • 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
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/8422—Investigating thin films, e.g. matrix isolation method
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • Y02W—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT
    • Y02W30/00—Technologies for solid waste management
    • Y02W30/50—Reuse, recycling or recovery technologies
    • Y02W30/80—Packaging reuse or recycling, e.g. of multilayer packaging

Abstract

A method for detecting the temperature barrier property of a recycled polyester material comprises the following steps: (1) forming a film from a polyester material to obtain a polyester film; (2) and collecting melting images (3) of the standard body under the two conditions that the polyester film does not cover the standard container and covers the standard container, processing and identifying the plurality of images by using an algorithm, marking the number of ice and the number of water in the images pixel by pixel, and judging the heat preservation performance according to preset conditions. The method disclosed by the invention is simple and quick in arrangement, can be directly applied to a production line, and is accurate in detection.

Description

Method for detecting temperature barrier property of recycled polyester material
Technical Field
The invention relates to the field of performance detection of a recycled polyester material, in particular to the detection of temperature barrier performance.
Background
In order to save resources, the current recycled polyester material is a widely applicable material. And sometimes designed with various functionalities, such as heat retention, depending on the needs of the application. Polyester materials are widely used for food outer packaging such as freezing, steaming, pickling, cakes, candies, beverages and the like due to excellent properties of the polyester materials. In recent years, with the rise of the internet and the take-out industry, polyester materials are further used for packaging outer belts of various foods. In this application scenario, the thermal insulation performance of polyester becomes an important factor. The rapid increase of the demand of disposable food outer packages also puts higher requirements on the production efficiency and the detection efficiency of the products. The traditional detection of the heat preservation performance of the material is usually completed by professional institutions by adopting professional instruments, the material needs to be fixed in a special and complex way, and a single test usually takes tens of minutes or even hours. And a large amount of professional equipment is needed for detection, the equipment can only be used in a laboratory, cannot be compatibly configured on a production line, is not suitable for detection of mass products, and cannot be used for quick self-checking of manufacturers. At present, a using mode is also proposed for detection, but the arrangement of the detection process and the design of an algorithm cannot meet the requirements on accuracy and rapidity.
Therefore, under the background of rapid increase of material demand, the requirements for timeliness, convenience in use and accuracy of detection equipment become urgent requirements of relevant manufacturers.
Disclosure of Invention
In order to solve one or more technical problems, the invention provides a method for detecting the heat insulation performance of a polyester film outer package, which is used for measuring the heat insulation performance of the polyester film outer package. The universal and easy-to-operate equipment is used as a measuring tool for quickly detecting the heat insulation performance of the polyester film outer package, and the capacity of efficiently screening out unqualified products is achieved, so that the method has a wide prospect under the commercial application of a new era.
A method for detecting the temperature barrier property of a regenerated polyester material,
step 1: preparation before detection and image acquisition steps
(1) Forming a film from a polyester material to obtain a polyester film;
(2) starting a heat source; the first standard container is conveyed to the detection area by the conveyor belt, and the camera acquires the image of the first standard body
Figure DEST_PATH_IMAGE002
After a time T, the camera acquires the image of the first standard body
Figure DEST_PATH_IMAGE004
；
(3) Starting a heat source; the second standard container is conveyed to the detection area by the conveyor belt, the polyester film to be detected is conveyed to the detection area by the conveying roller, the projection of the conveyor belt and the conveying direction of the conveying roller on the horizontal plane is vertical, the polyester film covers the standard container, and the camera acquires the image of the second standard body
Figure DEST_PATH_IMAGE006
After a time T, the camera acquires the image of the second standard body
Figure DEST_PATH_IMAGE008
；
Wherein the standard body is arranged in a standard container; the upper part of the standard container is provided with an opening, and the lower part of the standard container is transparent; a heat source is arranged above the detection area and used for radiating heat to the standard body; a camera is arranged below the detection area and used for shooting an image of the standard body from the bottom of the standard container; the standard body is an ice block with standard size and standard temperature;
step 2: for any image
Figure DEST_PATH_IMAGE010
Using multiple kernel functions
Figure DEST_PATH_IMAGE012
Calculating to obtain response spectrum
Figure DEST_PATH_IMAGE014
Figure DEST_PATH_IMAGE016
Wherein
Figure DEST_PATH_IMAGE018
，
Figure DEST_PATH_IMAGE020
The size of the corresponding kernel function is represented,
Figure DEST_PATH_IMAGE022
，
Figure DEST_PATH_IMAGE024
，
Figure DEST_PATH_IMAGE026
，
Figure DEST_PATH_IMAGE028
the size difference of adjacent kernel functions is shown, and N is the number of the kernel functions;
and step 3: computing images using step 2
Figure 571524DEST_PATH_IMAGE002
Image, and image
Figure 931574DEST_PATH_IMAGE004
Image, and image
Figure 609680DEST_PATH_IMAGE006
Image, and image
Figure 840941DEST_PATH_IMAGE008
Inputting the response spectrum of each image into the trained neural network model to perform pixel-by-pixel ice water identification and labeling;
and 4, step 4: calculating images according to the result of the step 3
Figure 288103DEST_PATH_IMAGE002
And image
Figure 805672DEST_PATH_IMAGE004
Number of pixels marked as ice in
Figure DEST_PATH_IMAGE030
、
Figure DEST_PATH_IMAGE032
And the number of pixels marked as water
Figure DEST_PATH_IMAGE034
、
Figure DEST_PATH_IMAGE036
(ii) a And calculating the image
Figure 784123DEST_PATH_IMAGE006
And image
Figure 553496DEST_PATH_IMAGE008
Number of pixels marked as ice in
Figure DEST_PATH_IMAGE038
、
Figure DEST_PATH_IMAGE040
And the number of pixels marked as water
Figure DEST_PATH_IMAGE042
、
Figure DEST_PATH_IMAGE044
；
If it is
Figure DEST_PATH_IMAGE046
And judging that the heat-insulating property of the film meets the requirement.
The standard container is constructed of clear glass.
The standard body is an ice block with the bottom side of 1cm and the thickness of 0.1cm, and is prepared by continuously freezing for 24 hours at the temperature of minus 18 ℃.
The ice cubes have a color.
The conveyor belts are two in number, and respectively bear two sides of the standard container, and the center of the bottom surface of the standard container is exposed. .
A plurality of detection areas are provided at a plurality of positions.
Each detection zone corresponds to one standard container and a pair of conveyor belts.
Each detection zone corresponds to a heat source and a camera.
And 3, in the training process in the step 3, collecting data samples of a plurality of images, and labeling the samples.
In step 2
Figure DEST_PATH_IMAGE048
。
The invention has the advantages that:
1. the invention provides a method for forming a polyester film, which is characterized in that the film and a detection standard unit are respectively conveyed to a detection area to realize rapid detection, simultaneously the composition of a detection standard container and a standard body is optimized, and an image acquisition mode is matched for use, so that the method is more suitable for arrangement on a production line and can realize rapid and accurate detection.
2. By the aid of the kernel function with variable size, accurate and comprehensive extraction of the response spectrum is achieved, and a foundation is laid for accurate and rapid detection of the next step.
3. The precision of the neural network detection algorithm is improved in a pixel-by-pixel identification and marking mode. And a final discrimination condition is obtained through a large number of experiments. By means of the mutual matching of the labeling mode and the judging condition, the accurate detection of the performance is realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a schematic view of the structure of the detecting device of the present invention.
Detailed Description
Step 1: preparation step before detection
Detection device structure
The detection device comprises an optical camera R4 which is arranged below the standard container and is used for shooting the image of the standard body from the bottom;
a standard container R2 of transparent glass construction, having one and only one open side on which the mylar R7 to be tested can be placed; it is intended to receive a standard body, a standard body R3, usually in the form of ice frozen in purified or distilled water, which can be prepared in advance in large quantities for testing purposes. The standard bodies are arranged in the standard container and jointly form a standard detection unit.
A preset heat source R1 arranged on the detection area and used for emitting heat to the polyester film to be detected with constant power and partially penetrating the polyester film to completely/partially melt the standard body in the standard container;
the polyester film to be tested is conveyed to the inspection area by the conveying roller R5. The standard container is transported to the inspection area by a conveyor R6. The conveyor belt is perpendicular to the projection of the conveying direction of the conveying roller on the horizontal plane, so that the conveying roller conveys the polyester film to be detected to the detection area, and the conveyor belt conveys the standard detection unit to the detection area, so that the polyester film covers the standard container of the standard detection unit, namely, the polyester film isolates the standard body in the standard container from external heat from the upper part. In order to avoid blocking the camera shooting, the conveyor belts are two belts which respectively bear two sides of the standard container and expose the center of the bottom surface of the standard container.
Preferably, a plurality of detection areas may be provided at a plurality of positions, each detection area corresponding to one detection unit and a pair of conveyor belts, i.e., a plurality of conveyor belts perpendicular to the projection of the conveying direction of the conveying roller on the horizontal plane. Of course, the heat source and the camera should be correspondingly provided in plural numbers. Therefore, the temperature barrier performance can be detected at a plurality of positions simultaneously, and the detection efficiency is improved.
Preparation of the assay
Before detection, the polyester material to be detected is prepared into a polyester film, so that the detection of the temperature barrier property is facilitated. The polyester film is then fed into the inspection apparatus. As one preference, the preparation may be by stretching, film blowing, or the like.
Before detection, preparing a reference detection object with a special shape and a special size through a general refrigeration device such as a freezer or a refrigerator, wherein the reference detection object is usually a cuboid with the bottom side length of 1cm and the thickness of 0.1cm, is convenient to place in a standard container and is kept stable in a frozen state; purified water or distilled water is selected for preparation, so that the influence of impurities on an optical image is reduced; as a preferable value of a long-term experiment, the sample is frozen at the temperature of minus 18 ℃ for 24 hours, and a reference detection object meeting the requirement can be formed; preferably, the pure water or distilled water is colored so that the ice cubes have a certain color after being frozen, thereby facilitating image recognition. In order to keep the consistency of the test, all the scenes using the reference detection object have the consistent preparation process of the reference detection object, including freezing temperature, continuous freezing time and refrigeration equipment.
Placing the standard container on a hollow conveyor belt, so that the transparent bottom surface can be shot by an optical camera; the optical cameras are arranged on the opposite sides of the opening of the standard container, and the shooting direction is vertical to the bottom surface.
The heat source is arranged on one side of the opening of the standard container, the heating direction of the heat source is opposite to the opening, and a proper distance is kept, so that the heat of the heat source uniformly and efficiently acts on the opening.
When the detection is carried out, firstly, taking out an ice block which refers to a detected object, putting the ice block into a standard container, selecting whether a film is used for sealing an opening or not, recording the time as the detection starting time, and shooting an image by adopting an optical camera; after a certain time, for example, 1 minute, the examination is ended, the time is recorded as the examination end time, and an image is taken.
The method specifically comprises the following steps: (1) starting a heat source; the first standard container is conveyed to the detection area by the conveyor belt, and the camera acquires the image of the first standard body
Figure 527268DEST_PATH_IMAGE002
After a time T, the camera acquires the image of the first standard body
Figure 91105DEST_PATH_IMAGE004
(ii) a (2) Starting a heat source; the second standard container is conveyed to the detection area by the conveyor belt, the polyester film to be detected is conveyed to the detection area by the conveying roller, the projection of the conveyor belt and the conveying direction of the conveying roller on the horizontal plane is vertical, the polyester film covers the standard container, and the camera acquires the image of the second standard body
Figure 743803DEST_PATH_IMAGE006
After a time T, the camera acquires the image of the second standard body
Figure 441500DEST_PATH_IMAGE008
. It is to be understood that, in the above (1), the acquisition is not required to be performed every time, and the acquired image may be used as a standard image.
Step 2: and processing the collected image to obtain a response spectrum.
When the detection device is arranged and implemented according to the step 1, an image is shot at the detection starting time, an image is shot at the detection ending time, and an image processing method of a response spectrum is provided.
Given a captured image, the image is represented by a matrix
Figure DEST_PATH_IMAGE050
Expressing that each element in the matrix, i.e. a corresponding pixel of the image, is used
Figure 800938DEST_PATH_IMAGE010
It is shown that,
Figure DEST_PATH_IMAGE052
indicating the position of the element in the matrix. Element(s)
Figure 942200DEST_PATH_IMAGE010
The value of (d) represents the luminance value of the pixel at that location of the image.
Get the
Figure 816615DEST_PATH_IMAGE010
A neighborhood of
Figure DEST_PATH_IMAGE054
Representing an image
Figure 786845DEST_PATH_IMAGE050
Middle pixel
Figure 63106DEST_PATH_IMAGE010
A subset of the matrices that are centered on,
Figure DEST_PATH_IMAGE056
one of the elements in the subset is represented,
Figure DEST_PATH_IMAGE058
，
Figure DEST_PATH_IMAGE060
，
Figure 841181DEST_PATH_IMAGE018
，
Figure 140576DEST_PATH_IMAGE020
representing a neighborhood
Figure 117759DEST_PATH_IMAGE054
The size of (c).
Accordingly, a kernel function is defined
Figure DEST_PATH_IMAGE062
Is and neighborhood
Figure 779684DEST_PATH_IMAGE054
The same size matrix.
Definition of
Figure 184121DEST_PATH_IMAGE010
For kernel function
Figure 846178DEST_PATH_IMAGE012
The response of (a) is expressed as:
Figure DEST_PATH_IMAGE064
further, define
Figure 158210DEST_PATH_IMAGE010
Response spectrum of
Figure DEST_PATH_IMAGE066
Is composed of
Figure 612325DEST_PATH_IMAGE010
A set of responses to a set of kernel functions of different sizes,
Figure 187663DEST_PATH_IMAGE018
，
Figure 337016DEST_PATH_IMAGE020
represents the size of the corresponding kernel function, and:
Figure DEST_PATH_IMAGE068
Figure DEST_PATH_IMAGE070
it shows that the sizes of the adjacent kernel functions in the response spectrum become larger in sequence, and the difference between the sizes of the two adjacent kernel functions is as
Figure 452739DEST_PATH_IMAGE028
. The number of kernel functions in the response spectrum is N. Different sized kernel functions are used to capture different sized features in an image. By setting the flexible and variable kernel function, the image feature extraction is more comprehensive and accurate.
The response spectrum is used for processing the shot image, parameters of the response spectrum are obtained through training, and then the kernel function is applied to the shot image to obtain the response spectrum. As a large number of experimentally preferred values, N =8,
Figure DEST_PATH_IMAGE072
.
and step 3: utilizing response spectrum to carry out pixel-by-pixel ice water identification and train model at the same time
During training, data samples of a plurality of images are collected, corresponding labels are carried out on the samples, and a group of response spectrum parameters are determined according to the samples and the labels.
The acquired data sample comprises a plurality of images which are further divided into two types, namely an image in a completely frozen state when a reference detection object (namely ice cubes) is taken out, and an image in a completely melted state after the reference detection object is taken out. Due to the difference of absorption and reflection of light by ice and water, the pixel distribution of the two types of images has different characteristics, and the characteristics can be obtained and the pixel types can be distinguished by establishing the response spectrum of the images.
Defining data sample images annotated as two matrices：
Figure DEST_PATH_IMAGE074
And
Figure DEST_PATH_IMAGE076
same as the image size, and:
Figure DEST_PATH_IMAGE078
Figure DEST_PATH_IMAGE080
the meaning of the image is the same as in step 2,
Figure 433465DEST_PATH_IMAGE074
and
Figure 117387DEST_PATH_IMAGE076
the annotation can be done manually.
For the response spectrum obtained in step 2, in order to further reduce the influence of noise, the kernel functions adjacent to the response spectrum are differentiated to improve the robustness of the model, according to step 2,
Figure DEST_PATH_IMAGE082
order:
Figure DEST_PATH_IMAGE084
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE086
Figure DEST_PATH_IMAGE088
by the equation (5), the influence of the local noise signal in the image is reduced, which contributes to improvement of the robustness of the model.
Further, two sets of independent parameters are defined
Figure DEST_PATH_IMAGE090
、
Figure DEST_PATH_IMAGE092
：
Figure DEST_PATH_IMAGE094
Figure DEST_PATH_IMAGE096
In (1),
Figure DEST_PATH_IMAGE098
、
Figure DEST_PATH_IMAGE100
represents a deviation parameter, and
Figure DEST_PATH_IMAGE102
is a non-linear function:
Figure DEST_PATH_IMAGE104
the non-linear function is used to establish a non-linear relationship between the response spectrum and the label,
Figure DEST_PATH_IMAGE106
for a parameter for adjusting the sensitivity of the classifier, preference is given according to experiments
Figure DEST_PATH_IMAGE108
。
Each kernel function in the response spectrum describes a linear response of the image to a certain feature, and equations (4) and (5) jointly establish a relational model of the response spectrum and the label, wherein the output of the model is two matrixes with the same size as the image space, and one element of each matrix is in a corresponding relation with the input image pixel in space, so that the class label (ice, water or other) of any pixel in the image is judged.
According to the data sample, the Back propagation method is adopted to train the formula (4), and parameters can be obtained
Figure 32997DEST_PATH_IMAGE012
、
Figure DEST_PATH_IMAGE110
、
Figure DEST_PATH_IMAGE112
、
Figure 499882DEST_PATH_IMAGE098
、
Figure 397431DEST_PATH_IMAGE100
，
Figure 314571DEST_PATH_IMAGE026
。
And 4, step 4: discrimination of temperature barrier performance by using pixel identification result
The image-based heat preservation performance detection method obtains two images at the beginning and the end time without using a film to seal an opening according to the implementation method in the step 1
Figure 687784DEST_PATH_IMAGE002
And
Figure 348572DEST_PATH_IMAGE004
(ii) a Respectively calculating corresponding response spectrums according to the method in the step 2, and calculating according to the method of the formula (4) in the step 3
Figure 100627DEST_PATH_IMAGE002
And
Figure 454248DEST_PATH_IMAGE004
number of pixels marked as ice in
Figure 62560DEST_PATH_IMAGE030
、
Figure 464722DEST_PATH_IMAGE032
And the number of pixels marked as water
Figure 133601DEST_PATH_IMAGE034
、
Figure 720440DEST_PATH_IMAGE036
. Two images were acquired at the start and end times under conditions where the film seal was open
Figure 943611DEST_PATH_IMAGE006
And
Figure 680623DEST_PATH_IMAGE008
the corresponding number of pixels labeled ice is also obtained
Figure 204008DEST_PATH_IMAGE038
、
Figure 633852DEST_PATH_IMAGE040
And the number of pixels marked as water
Figure 422948DEST_PATH_IMAGE042
、
Figure 776700DEST_PATH_IMAGE044
。
Count value to be obtained without sealing the opening with a film
Figure 230291DEST_PATH_IMAGE030
、
Figure 627774DEST_PATH_IMAGE032
And with
Figure 356695DEST_PATH_IMAGE034
、
Figure 435510DEST_PATH_IMAGE036
For reference, if the count value after the seal is opened satisfies the following condition:
Figure DEST_PATH_IMAGE114
the heat insulation performance of the film is judged to meet the requirement.
The determination condition is preferably a condition that enables accurate determination by the determination condition on the premise of using the neural network, and is most efficient.
Method of the invention verification
The performance detection result obtained by the method is compared with the detection requirement standard of related manufacturers. The detection requirement criteria are defined as follows:
is provided with
Figure DEST_PATH_IMAGE116
Is the mass of ice in the frozen state,
Figure DEST_PATH_IMAGE118
、
Figure DEST_PATH_IMAGE120
the mass of the ice mass is partially melted before and after the film is used; the heat insulation performance of the film meets the standard definition as follows:
Figure DEST_PATH_IMAGE122
according to the comparison of the results in the following table, the detection result of the method disclosed by the invention is small in standard error, and the automatic detection of the heat insulation performance of the film is realized.
Figure DEST_PATH_IMAGE124

Claims (10)

1. A method for detecting the temperature barrier property of a regenerated polyester material is characterized by comprising the following steps:
step 1: preparation before detection and image acquisition steps
(1) Forming a film from a polyester material to obtain a polyester film;
(2) starting a heat source; the first standard container is conveyed to the detection area by a conveyor belt, and the camera acquires an image E of the first standard body 1 After a time T, the camera captures an image E of the first standard volume 2 ；
(3) Starting a heat source; the second standard container is conveyed to the detection area by a conveyor belt, the polyester film is conveyed to the detection area by a conveying roller, the conveyor belt is perpendicular to the projection of the conveying direction of the conveying roller on the horizontal plane, the polyester film covers the standard container, and the camera acquires an image E of the second standard body 1 ' the camera acquires an image E of the second standard body after a time T 2 ′；
Wherein the standard body is arranged in a standard container; the upper part of the standard container is provided with an opening, and the lower part of the standard container is transparent; a heat source is arranged above the detection area and used for radiating heat to the standard body; a camera is arranged below the detection area and used for shooting an image of the standard body from the bottom of the standard container; the standard body is an ice block with standard size and standard temperature;
step 2: for any image
Figure FDA0003764459820000011
Using multiple kernel functions
Figure FDA0003764459820000012
Calculating to obtain response spectrum
Figure FDA0003764459820000013
Figure FDA0003764459820000014
Wherein A is i ，B i Representing the size of the corresponding kernel function, A i+1 ＝A i +Γ，B i+1 ＝B i + delta, i is more than or equal to 1 and less than or equal to N, gamma is the size difference of adjacent kernel functions, and N is the number of the kernel functions;
and step 3: computing image E using step 2 1 Image E 2 Image E 1 ', image E 2 The response spectrum of' inputs the response spectrum of each image into the trained neural network model to carry out pixel-by-pixel ice water identification and labeling;
wherein, the neural network model is as follows:
labels defining the data sample image are two matrices:
Figure FDA0003764459820000015
and with
Figure FDA0003764459820000016
Same as the image size, and:
Figure FDA0003764459820000021
Figure FDA0003764459820000022
according to the step 2, the process is carried out,
Figure FDA0003764459820000023
order:
Figure FDA0003764459820000024
wherein the content of the first and second substances,
Figure FDA0003764459820000025
Figure FDA0003764459820000026
further, two sets of independent parameters are defined
Figure FDA0003764459820000027
Figure FDA0003764459820000028
(5) In, phi 1 、ψ 2 Represents a deviation parameter, and Ω is a non-linear function:
Figure FDA0003764459820000029
and 4, step 4: calculating an image E according to the result of the step 3 1 And image E 2 Number of pixels ice marked therein 1 、ice 2 And the number of pixels marked as water 1 、water 2 (ii) a And calculating an image E 1 ' sum image E 2 Number of pixels ice marked in ` 1 ′、ice 2 ' with the number of pixels marked as water Water 1 ′、water 2 ′；
If it is
Figure FDA00037644598200000210
Figure FDA00037644598200000211
The heat-insulating performance of the polyester film is judged to meet the requirement.
2. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: the standard container is constructed of clear glass.
3. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: the standard body is an ice block with the bottom edge length of 1cm and the thickness of 0.1cm, and is prepared by continuously freezing at the temperature of 18 ℃ below zero for 24 hours.
4. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 3, wherein: the ice cubes have a color.
5. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: the conveyor belts are two in number, and respectively bear two sides of the standard container, and the center of the bottom surface of the standard container is exposed.
6. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: a plurality of detection areas are provided at a plurality of positions.
7. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 6, wherein: each detection zone corresponds to one standard container and a pair of conveyor belts.
8. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 6, wherein: each detection zone corresponds to a heat source and a camera.
9. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: and 3, in the training process in the step 3, collecting data samples of a plurality of images, and labeling the samples.
10. The method for detecting the temperature barrier property of the recycled polyester material as claimed in claim 1, wherein: in step 2, Γ is 4 and Δ is 4.
CN202210571598.4A 2022-05-25 2022-05-25 Method for detecting temperature barrier property of recycled polyester material Active CN114660112B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210571598.4A CN114660112B (en) 2022-05-25 2022-05-25 Method for detecting temperature barrier property of recycled polyester material

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210571598.4A CN114660112B (en) 2022-05-25 2022-05-25 Method for detecting temperature barrier property of recycled polyester material

Publications (2)

Publication Number Publication Date
CN114660112A CN114660112A (en) 2022-06-24
CN114660112B true CN114660112B (en) 2022-09-02

Family

ID=82036810

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210571598.4A Active CN114660112B (en) 2022-05-25 2022-05-25 Method for detecting temperature barrier property of recycled polyester material

Country Status (1)

Country Link
CN (1) CN114660112B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109299715A (en) * 2017-07-24 2019-02-01 图灵通诺（北京）科技有限公司 The settlement method and device of image recognition technology based on convolutional neural networks
US10402699B1 (en) * 2015-12-16 2019-09-03 Hrl Laboratories, Llc Automated classification of images using deep learning—back end

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5622430A (en) * 1993-11-05 1997-04-22 Degussa Aktiengesellschaft Method of testing the heat insulation action of bodies especially of heat insulation bodies
JP2001228105A (en) * 2000-02-15 2001-08-24 Toshiba Corp Characteristic evaluating device for thermally insulating coating and charcteristic evaluating method tehrefor
CN101576519B (en) * 2009-06-16 2011-08-31 上海市建筑科学研究院(集团)有限公司 Outdoor dynamic testing method for heat-insulation property of heat-insulation coating at outer wall of building
JP2012198142A (en) * 2011-03-22 2012-10-18 Tokyo Electric Power Co Inc:The Area measurement device and area measurement method
CN102661972B (en) * 2012-05-16 2015-02-11 北京黎明航发动力科技有限公司沈阳黎明燃机分公司 measurement method of heat insulating property of gas turbine heat shield
CN110501085A (en) * 2018-12-11 2019-11-26 中国航空工业集团公司北京长城计量测试技术研究所 A kind of thermal barrier coating heat-proof quality measurement method based on film thermometric
CN111933868B (en) * 2020-10-13 2020-12-22 河南银金达新材料股份有限公司 Polyethylene film and preparation method thereof
CN114074469A (en) * 2021-07-27 2022-02-22 河南源宏高分子新材料有限公司 Polyester film with high barrier property and preparation and application thereof

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10402699B1 (en) * 2015-12-16 2019-09-03 Hrl Laboratories, Llc Automated classification of images using deep learning—back end
CN109299715A (en) * 2017-07-24 2019-02-01 图灵通诺（北京）科技有限公司 The settlement method and device of image recognition technology based on convolutional neural networks

Also Published As

Publication number Publication date
CN114660112A (en) 2022-06-24

Similar Documents

Publication Publication Date Title
TWI499669B (en) Method for detecting a microorganism, apparatus for detecting a microorganism, and program
CN112881467A (en) Large-size composite material damage imaging and quantitative identification method
US20200049632A1 (en) Method and system for determining package integrity
CN106153639B (en) Circuit board detecting method based on artificial intelligence and its detection device
US11568532B2 (en) Methods and systems for thermal imaging of moving objects
CN109801287B (en) Label breakage detection method based on template matching and image quality evaluation
CN102246186A (en) Method and system for item identification
CN114660112B (en) Method for detecting temperature barrier property of recycled polyester material
CN102547079A (en) Night scene light source detection device and method
KR20110018081A (en) Side surface inspection apparatus of cylinder type rechargeable battery
CN115165897A (en) Defect detection method, system and device
CN101021488A (en) Method and apparatus for detecting rice damage
CN106560910A (en) Tube core defect detection method and device
CN107515198A (en) A kind of online test method of starch burn degree and gelatinization point
CN114656666B (en) PET material preparation process capable of realizing temperature barrier
CN116883893A (en) Tunnel face underground water intelligent identification method and system based on infrared thermal imaging
US11836908B2 (en) Method and system for determining package integrity
CN116091506A (en) Machine vision defect quality inspection method based on YOLOV5
CN211553736U (en) Hyperspectral detection device of fingerprint on plane carrier
CN111307805B (en) Threshing quality detection device and detection method based on visual feature fusion
CN105427335A (en) Method for detecting and locating skip plating defects of continuous strip-shaped porous metal material
CN110197227B (en) Multi-model fusion intelligent instrument reading identification method
CN108896557B (en) Device and method for detecting edge cracking defect of positive edge chamfer of lens
JPH06229957A (en) Material discrimination method
CN210358127U (en) Military ceramic capacitor appearance detection equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant