CN113989286A - Deep learning method, device, equipment and storage medium for belt tearing strength - Google Patents
Deep learning method, device, equipment and storage medium for belt tearing strength Download PDFInfo
- Publication number
- CN113989286A CN113989286A CN202111628686.5A CN202111628686A CN113989286A CN 113989286 A CN113989286 A CN 113989286A CN 202111628686 A CN202111628686 A CN 202111628686A CN 113989286 A CN113989286 A CN 113989286A
- Authority
- CN
- China
- Prior art keywords
- tearing
- image
- current
- preset
- belt
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013135 deep learning Methods 0.000 title claims abstract description 32
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 230000005540 biological transmission Effects 0.000 claims abstract description 51
- 230000011218 segmentation Effects 0.000 claims abstract description 44
- 239000011159 matrix material Substances 0.000 claims abstract description 28
- 238000013136 deep learning model Methods 0.000 claims abstract description 21
- 238000001914 filtration Methods 0.000 claims description 29
- 238000000605 extraction Methods 0.000 claims description 14
- 238000005259 measurement Methods 0.000 claims description 11
- 238000000691 measurement method Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000000875 corresponding effect Effects 0.000 description 10
- 230000008569 process Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 239000000463 material Substances 0.000 description 4
- 239000002390 adhesive tape Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004566 building material Substances 0.000 description 1
- 239000003638 chemical reducing agent Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005272 metallurgy Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- -1 power plants Substances 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention relates to the technical field of deep learning, and discloses a method, a device, equipment and a storage medium for deep learning of belt tearing strength, wherein the method comprises the following steps: extracting a current tearing image of the transmission belt according to a preset image ROI strategy; segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area; obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image; measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt; compared with the prior art that the tearing degree of the belt is determined manually, the accuracy of measuring the tearing of the belt can be effectively improved, and the transmission danger is reduced.
Description
Technical Field
The invention relates to the technical field of deep learning, in particular to a method, a device, equipment and a storage medium for measuring tearing strength of a belt in deep learning.
Background
The belt conveyer is used as important transportation equipment for transporting materials by enterprises in various fields, and occupies an important position in production, the fields comprise ports, metallurgy, mines, chemical industry, petroleum, power plants, building materials and the like, the main materials of the conveying belt are rubber, fiber core adhesive tapes, common canvas core adhesive tapes or steel rope core adhesive tapes and other core ropes made of different materials, the tearing condition is easy to occur in the continuous operation process of the conveying belt, but the belt is expensive and approximately occupies 50 percent of the cost of the whole belt conveyer, so that the belt conveyer cannot be replaced when one thread of the belt is torn, and also cannot be replaced after the belt is completely torn because the belt conveyer is still in a working state when the belt conveyer is completely torn, the materials are spilled, the speed reducer, the motor and other equipment are damaged, and even the frame structure is damaged when the condition is serious, the personal safety of field personnel is threatened, the tearing degree is used as the measurement standard of the replacement of the transmission belt, the existing belt tearing degree is determined based on manual experience, certain errors occur in manual judgment more or less, and the accuracy of finally determining the belt tearing degree is low.
The above is only for the purpose of assisting understanding of the technical aspects of the present invention, and does not represent an admission that the above is prior art.
Disclosure of Invention
The invention mainly aims to provide a method, a device, equipment and a storage medium for deeply learning belt tearing strength, and aims to solve the technical problem that transmission accidents are frequent due to low belt tearing measurement accuracy in the prior art.
To achieve the above object, the present invention provides a deep learning belt tearing strength method, comprising the steps of:
acquiring a current tearing image of a transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image;
segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area;
obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image;
and measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt.
Optionally, the obtaining a current torn image of the transmission belt, extracting the current torn image according to a preset image ROI policy, and obtaining a torn length image includes:
acquiring a current tearing image and a target belt characteristic set of a transmission belt;
performing image extraction on the current tearing image through a preset image ROI strategy according to the target belt feature set;
performing single-channel extraction on the extracted current tearing image to obtain a target channel tearing image;
and filtering the target channel tearing image through preset average filtering equipment to obtain a tearing length image.
Optionally, the filtering the target channel tearing image through a preset average filtering device to obtain a tearing length image includes:
obtaining a corresponding window specification and a center coordinate according to the target channel tearing image;
carrying out average calculation on image pixel points in the window specification to obtain the average value of the current pixels;
and filtering the target channel tearing image through preset average filtering equipment according to the central coordinate and the current pixel average value to obtain a tearing length image.
Optionally, the segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target region includes:
acquiring the acquisition direction of the current tearing image;
averagely dividing the tearing length image according to the acquisition direction to obtain a target number tearing length image;
respectively calculating the tearing length images of the target quantity to obtain pixel Lorentz information measure;
and when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area.
Optionally, when the lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tear length image according to a preset laser stripe segmentation algorithm to obtain a tear stripe image of the target region, including:
when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold value, obtaining the image gray probability according to the tearing length image;
calculating the image gray probability through a preset weighting strategy to obtain target linear weighting;
determining a corresponding segmentation threshold according to the target linear weighting;
and segmenting the tearing length image through a preset laser stripe segmentation algorithm according to the segmentation threshold to obtain a tearing stripe image of the target area.
Optionally, the obtaining of the current tearing characteristic according to the preset matrix tracking algorithm and the target region tearing stripe image includes:
detecting the direction of the tearing stripe of the target area according to a preset matrix tracking algorithm to obtain the normal direction of the current tearing stripe;
obtaining pixel point coordinates according to the tearing stripe image of the target area;
determining a tear stripe central line according to the current tear stripe normal direction and the pixel point coordinates;
and tracking the center line of the tearing stripe through a preset matrix tracking algorithm to obtain the current tearing characteristic.
Optionally, the measuring the current tear characteristic through a target deep learning model to obtain the tear of the transmission belt includes:
obtaining a corresponding tearing image point according to the current tearing characteristic;
performing linear fitting on the tearing image points to obtain a current tearing curve;
and when the curvature of the current tearing curve is larger than a preset curvature threshold value, measuring the current tearing curve through a target deep learning model to obtain the tearing strength of the transmission belt.
Further, to achieve the above object, the present invention also proposes a deep-learning belt tear measuring device including:
the extraction module is used for acquiring a current tearing image of the transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image;
the segmentation module is used for segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area;
the acquisition module is used for acquiring the current tearing characteristics according to a preset matrix tracking algorithm and the tearing stripe image of the target area;
and the measurement module is used for measuring the current tearing characteristics through a target deep learning model to obtain the tearing degree of the transmission belt.
Further, to achieve the above object, the present invention also proposes a deep-learning belt tear measuring apparatus comprising: a memory, a processor, and a deep-learned belt tear metric program stored on the memory and executable on the processor, the deep-learned belt tear metric program configured to implement a deep-learned belt tear metric method as described above.
Further, to achieve the above object, the present invention also proposes a storage medium having stored thereon a deep-learned belt tear measurement program which, when executed by a processor, implements the deep-learned belt tear measurement method as described above.
The belt tearing strength measuring method for deep learning comprises the steps of obtaining a current tearing image of a transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image; segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area; obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image; measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt; compared with the prior art that the tearing degree of the belt is determined manually, the accuracy of measuring the tearing of the belt can be effectively improved, and the transmission danger is reduced.
Drawings
FIG. 1 is a schematic diagram of a deep learning belt tear metrology device for a hardware operating environment in accordance with an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a first embodiment of a deep learned method of belt tear of the present invention;
FIG. 3 is a schematic flow chart of a second embodiment of a deep learned method of belt tear of the present invention;
FIG. 4 is a schematic flow chart of a third embodiment of a deep learned method of belt tear of the present invention;
FIG. 5 is a functional block diagram of a deep learned belt tear measurement apparatus according to a first embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a deep learning belt tearing measuring apparatus for a hardware operating environment according to an embodiment of the present invention.
As shown in fig. 1, the deep-learning belt tear measurement apparatus may include: a processor 1001, such as a Central Processing Unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a Wireless interface (e.g., a Wireless-Fidelity (Wi-Fi) interface). The Memory 1005 may be a Random Access Memory (RAM) Memory, or may be a Non-Volatile Memory (NVM), such as a disk Memory. The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the configuration shown in FIG. 1 does not constitute a limitation of a deep-learned belt tear measuring device, and may include more or fewer components than shown, or some components in combination, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a storage medium, may include therein an operating system, a network communication module, a user interface module, and a deep-learning belt tear measurement program.
In the deep-learning belt tear metrology device shown in FIG. 1, the network interface 1004 is primarily used for data communication with the network-integrated platform workstation; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 of the deep-learning belt tear measurement apparatus of the present invention may be provided in a deep-learning belt tear measurement apparatus, which calls a deep-learning belt tear measurement program stored in the memory 1005 through the processor 1001 and executes the deep-learning belt tear measurement method provided by the embodiment of the present invention.
Based on the hardware structure, the embodiment of the belt tearing strength measuring method for deep learning is provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of a first embodiment of a deep learned belt tear measurement method of the present invention.
In a first embodiment, the deep learned belt tearability measurement method comprises the steps of:
and step S10, acquiring a current tearing image of the transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image.
It should be noted that the execution subject of the present embodiment is a belt tearing measuring device with deep learning, and may also be other devices that can implement the same or similar functions, such as a measuring server.
It should be understood that the current torn image refers to an image of the transmission belt when the transmission belt is torn, and the current torn image is acquired through a machine vision system, where the machine vision system includes a vision sensor, an image acquisition card and a linear laser device, specifically, the linear laser device projects a laser stripe to the transmission belt, and then the vision sensor acquires the current torn image of the transmission belt in real time and transmits the current torn image to the measurement server in real time through the image acquisition card.
It can be understood that the tear length image refers to a length image Of a tear opening Of a transmission belt, and is obtained by extracting a current tear image through a preset image ROI (region Of interest) strategy, the preset image ROI (region Of interest) strategy refers to a strategy for extracting a region Of interest in an image, and the difficulty in processing the image can be effectively reduced through the preset image ROI strategy.
And step S20, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area.
It can be understood that the preset laser stripe segmentation algorithm refers to an algorithm for segmenting the tear stripe image of the target region from the tear length image, and the preset laser stripe segmentation algorithm may be a lorentz measure segmentation algorithm or other algorithms, which is not limited in this embodiment.
It should be understood that the target region tearing stripe image refers to an image of a tearing stripe of the region to be measured, and the tearing length image is divided into a target region tearing stripe image and a background region image, that is, the target region tearing stripe image and the background region image are separated by a preset laser stripe segmentation algorithm.
And step S30, obtaining the current tearing characteristics according to the preset matrix tracking algorithm and the target area tearing stripe image.
It should be understood that the preset matrix tracking algorithm refers to an algorithm for tracking a tear stripe image of a target region in real time according to a laser stripe direction, the preset matrix tracking algorithm may be a black plug matrix tracking algorithm or other matrix tracking algorithms, and this embodiment does not limit this, and the current tear feature refers to a feature capable of uniquely identifying the tear stripe image of the target region, and the current tear feature may be a pixel light stripe central point.
Further, step S30 includes: detecting the direction of the tearing stripe of the target area according to a preset matrix tracking algorithm to obtain the normal direction of the current tearing stripe; obtaining pixel point coordinates according to the tearing stripe image of the target area; determining a tear stripe central line according to the current tear stripe normal direction and the pixel point coordinates; and tracking the center line of the tearing stripe through a preset matrix tracking algorithm to obtain the current tearing characteristic.
It will be understood that, the current tear stripe normal direction refers to the normal direction of the target area tear stripe, the normal direction of the current tearing stripe is determined by a black plug matrix of a preset matrix tracking algorithm, the pixel point coordinate refers to the pixel position of the stripe center in the tearing stripe image of the target area, since the tearing stripe image of the target area is in Gaussian distribution in the normal direction of the current tearing stripe, namely, the larger the gray value in the tearing stripe image of the target area is, the closer the tearing stripe image is to the central line of the tearing stripe, and the preset matrix tracking algorithm can only process the pixel-level image, the pixel point coordinates need to be rounded, then the center line of the torn stripe is determined according to the normal direction of the current torn stripe and the pixel point coordinates, and tracking points on the center line of the torn stripe is repeatedly tracked through a preset matrix tracking algorithm, so that the current torn characteristic is obtained.
And step S40, measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt.
It can be understood that the tearing degree refers to the tearing degree of the transmission belt, after the current tearing characteristics are obtained, the current tearing characteristics are measured through a target deep learning model to obtain the tearing degree of the transmission belt, the target deep learning model is obtained through a deep learning network training historical tearing sample data set, and the historical tearing sample data set is composed of the tearing characteristics and the tearing degree when the transmission belt of the same type is torn.
Further, step S40 includes: obtaining a corresponding tearing image point according to the current tearing characteristic; performing linear fitting on the tearing image points to obtain a current tearing curve; and when the curvature of the current tearing curve is larger than a preset curvature threshold value, measuring the current tearing curve through a target deep learning model to obtain the tearing strength of the transmission belt.
It should be understood that the current tear curve refers to a curve obtained by linear fitting of tear image points, the size of a tear gap of the transmission belt is determined through the current tear curve, the preset curvature threshold refers to a maximum tear threshold of the transmission belt, when the current tear curve is smaller than the preset curvature threshold, it indicates that the transmission belt is slightly torn, and the whole transportation work is not affected, and when the current tear curve is larger than the preset curvature threshold, the current tear curve needs to be measured through a target deep learning model at this time to obtain the tear degree of the transmission belt.
According to the method, a current tearing image of a transmission belt is obtained, and the current tearing image is extracted according to a preset image ROI strategy to obtain a tearing length image; segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area; obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image; measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt; in the embodiment, the current tearing image is extracted through a preset image ROI strategy, the tearing length image is segmented according to a preset laser stripe segmentation algorithm, and then the current tearing characteristic is measured through a target deep learning model.
In an embodiment, as shown in fig. 3, a second embodiment of the deep learning belt tearing strength method according to the present invention is provided based on the first embodiment, and the step S10 includes:
and S101, acquiring a current tearing image and a target belt characteristic set of the transmission belt.
It should be understood that the current torn image refers to an image when the transmission belt is torn, and the target belt feature set refers to a set formed by various features of the image of the transmission belt, including the number of bits, the mode, the content and the like of the image, for example, the number of bits of the image is 24 bits, the mode is an RGB color mode, and the content is three times that of a single-channel image.
And S102, carrying out image extraction on the current tearing image through a preset image ROI strategy according to the target belt feature set.
It can be understood that after the current torn image and the target belt feature set are obtained, the current torn image is extracted through the preset image ROI strategy image according to the target belt feature set, namely the current torn image is cut to the minimum area of the package torn image, so that the number of processed images can be effectively reduced, and interference is avoided.
And S103, performing single-channel extraction on the extracted current tearing image to obtain a target channel tearing image.
It should be understood that the target channel tearing image refers to a single-channel tearing image, and since the subsequent processes of filtering the tearing image and segmenting the tearing-length image are performed under the single-channel image, in order to reduce the amount of operation processing and preserve the features of the current tearing image, the single-channel extraction of the extracted current tearing image is continued after the current extracted tearing image is obtained.
And step S104, filtering the target channel tearing image through preset average filtering equipment to obtain a tearing length image.
It can be understood that the tear length image refers to an image corresponding to a tear length when the transmission belt is torn, the preset average filtering device refers to a device for filtering the tear image of the target channel, the preset average filtering device may be an arithmetic average filter, or may be another average filtering device.
Further, step S104 includes: obtaining a corresponding window specification and a center coordinate according to the target channel tearing image; carrying out average calculation on image pixel points in the window specification to obtain the average value of the current pixels; and filtering the target channel tearing image through preset average filtering equipment according to the central coordinate and the current pixel average value to obtain a tearing length image.
It should be understood that the window specification refers to the size of a window formed by a target channel tearing image, the window specification is formed by the length and the width of the target channel tearing image, the current pixel average value refers to an average value of image pixel points in the window specification, and after obtaining a center coordinate and the current pixel average value, the target channel tearing image is filtered through a preset average value filtering device, specifically:
wherein S isxyIs the window specificationAnd m is n, x is a central horizontal coordinate, y is a central vertical coordinate, f (x, y) is the average value of the current pixel, and (s, t) is a pixel coordinate point in the window specification.
The method comprises the steps of obtaining a current tearing image and a target belt characteristic set of a transmission belt; performing image extraction on the current tearing image through a preset image ROI strategy according to the target belt feature set; performing single-channel extraction on the extracted current tearing image to obtain a target channel tearing image; filtering the target channel tearing image through preset average filtering equipment to obtain a tearing length image; according to the method, the current tearing image is extracted through the preset image ROI strategy, then the current tearing image is extracted through a single channel again to obtain the target channel tearing image, and the target channel tearing image is filtered according to the preset mean value filtering equipment, so that the accuracy of obtaining the tearing length image can be effectively improved.
In an embodiment, as shown in fig. 4, a third embodiment of the deep learning belt tearing strength measuring method according to the present invention is provided based on the first embodiment, and the step S20 includes:
step S201, acquiring the acquisition direction of the current tearing image.
It is understood that the collecting direction refers to a direction in which the machine vision system collects the current tearing image, and the collecting direction is adapted to the position of the machine vision system, for example, when the position of the machine vision system is in a horizontal setting, the collecting direction is in a horizontal direction.
Step S202, the tearing length images are divided averagely according to the acquisition direction, and a target number tearing length image is obtained.
It should be understood that the target number of tear-length images refers to dividing the tear-length image into the target number of tear-length images, where the target number N may be 5, or may be another number, and this embodiment does not limit this, and after obtaining the acquisition direction of the current tear-length image, in order to effectively improve the segmentation accuracy, the tear-length image is divided according to the acquisition direction to obtain the target number of tear-length images.
And step S203, calculating the tearing length images of the target quantity respectively to obtain the Lorentz information measure of the pixels.
It can be understood that the pixel lorentz information measure refers to an image information measure of a pixel in the target quantity tearing length image, and the information type in the target quantity tearing length image is determined through the pixel lorentz information measure, namely the pixel lorentz information measure is positively correlated with the information type.
And S204, when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area.
It should be understood that the preset information measure threshold refers to a maximum value of lorentz information measure, when the pixel lorentz information measure is larger than the preset information measure threshold, it is indicated that the information types in the tear-length image are more, at this time, the tear-length image needs to be re-divided, the number of divided targets is N +1, and when the pixel lorentz information measure is smaller than or equal to the preset information measure threshold, the tear-length image is divided through a preset laser stripe division algorithm to obtain a tear stripe image of the target area.
Further, step S204 includes: when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold value, obtaining the image gray probability according to the tearing length image; calculating the image gray probability through a preset weighting strategy to obtain target linear weighting; determining a corresponding segmentation threshold according to the target linear weighting; and segmenting the tearing length image through a preset laser stripe segmentation algorithm according to the segmentation threshold to obtain a tearing stripe image of the target area.
It can be understood that the image gray level probability refers to the probability of the gray level in the tear-length image, and the preset weighting policy refers to a policy of calculating the image gray level probability to obtain the target linear weighting, specifically:
wherein p isiFor linear weighting of the object, p(i+n)(n-2) p for image gray scale probability(i+1)Is a reaction of with p(i+n)Adjacent image gray scale probability, (n-1) piIs a group of a general formula of (n-2) p(i+1)Adjacent image gray scale probability, (n-2) p(i-1)Is a group of a general formula of (n-1) piAdjacent image gray level probability, p(i-n)Is a group of a general formula of (n-2) p(i-1)Adjacent image gray scale probabilities.
It should be understood that after the target linear weighting is obtained, the segmentation threshold is determined by the target linear weighting, specifically:
wherein T is a division threshold value,
in order to be the weight, the weight is,
is the inter-class variance of the threshold t.
The embodiment acquires the acquisition direction of the current tearing image; averagely dividing the tearing length image according to the acquisition direction to obtain a target number tearing length image; respectively calculating the tearing length images of the target quantity to obtain pixel Lorentz information measure; when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area; according to the embodiment, the tearing length image is divided averagely through the acquisition direction, and then when the Lorentz information measure of the pixel is less than or equal to the preset information measure threshold, the tearing length image is divided according to the preset laser stripe division algorithm, so that the accuracy of obtaining the tearing stripe image of the target area is effectively improved.
Furthermore, an embodiment of the present invention further provides a storage medium having a deep-learned belt tear measurement program stored thereon, which when executed by a processor implements the steps of the deep-learned belt tear measurement method as described above.
Since the storage medium adopts all technical solutions of all the embodiments, at least all the beneficial effects brought by the technical solutions of the embodiments are achieved, and no further description is given here.
In addition, referring to fig. 5, an embodiment of the present invention further provides a deep-learning belt tear measurement apparatus, including:
the extraction module 10 is configured to obtain a current torn image of the transmission belt, and extract the current torn image according to a preset image ROI policy to obtain a torn length image.
And the segmentation module 20 is configured to segment the tear length image according to a preset laser stripe segmentation algorithm to obtain a tear stripe image of the target region.
And the obtaining module 30 is configured to obtain a current tearing characteristic according to a preset matrix tracking algorithm and the target region tearing stripe image.
And the measuring module 40 is used for measuring the current tearing characteristics through a target deep learning model to obtain the tearing degree of the transmission belt.
According to the method, a current tearing image of a transmission belt is obtained, and the current tearing image is extracted according to a preset image ROI strategy to obtain a tearing length image; segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area; obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image; measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt; in the embodiment, the current tearing image is extracted through a preset image ROI strategy, the tearing length image is segmented according to a preset laser stripe segmentation algorithm, and then the current tearing characteristic is measured through a target deep learning model.
It should be noted that the above-described work flows are only exemplary, and do not limit the scope of the present invention, and in practical applications, a person skilled in the art may select some or all of them to achieve the purpose of the solution of the embodiment according to actual needs, and the present invention is not limited herein.
In addition, the technical details that are not elaborated in this embodiment may be referred to the deep learning belt tearing strength method provided in any embodiment of the present invention, and are not described herein again.
In an embodiment, the extracting module 10 is further configured to obtain a current tear image and a target belt feature set of the conveying belt; performing image extraction on the current tearing image through a preset image ROI strategy according to the target belt feature set; performing single-channel extraction on the extracted current tearing image to obtain a target channel tearing image; and filtering the target channel tearing image through preset average filtering equipment to obtain a tearing length image.
In an embodiment, the extracting module 10 is further configured to obtain a corresponding window specification and a center coordinate according to the target channel tearing image; carrying out average calculation on image pixel points in the window specification to obtain the average value of the current pixels; and filtering the target channel tearing image through preset average filtering equipment according to the central coordinate and the current pixel average value to obtain a tearing length image.
In an embodiment, the segmentation module 20 is further configured to obtain an acquisition direction of the current tearing image; averagely dividing the tearing length image according to the acquisition direction to obtain a target number tearing length image; respectively calculating the tearing length images of the target quantity to obtain pixel Lorentz information measure; and when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area.
In an embodiment, the segmentation module 20 is further configured to obtain an image gray level probability according to the tear length image when the lorentz information measure of the pixel is less than or equal to a preset information measure threshold; calculating the image gray probability through a preset weighting strategy to obtain target linear weighting; determining a corresponding segmentation threshold according to the target linear weighting; and segmenting the tearing length image through a preset laser stripe segmentation algorithm according to the segmentation threshold to obtain a tearing stripe image of the target area.
In an embodiment, the obtaining module 30 is further configured to perform direction detection on the tear stripe in the target area according to a preset matrix tracking algorithm, so as to obtain a normal direction of the current tear stripe; obtaining pixel point coordinates according to the tearing stripe image of the target area; determining a tear stripe central line according to the current tear stripe normal direction and the pixel point coordinates; and tracking the center line of the tearing stripe through a preset matrix tracking algorithm to obtain the current tearing characteristic.
In an embodiment, the metric module 40 is further configured to obtain a corresponding tearing image point according to the current tearing feature; performing linear fitting on the tearing image points to obtain a current tearing curve; and when the curvature of the current tearing curve is larger than a preset curvature threshold value, measuring the current tearing curve through a target deep learning model to obtain the tearing strength of the transmission belt.
Other embodiments or methods of implementing the deep learning belt tear measurement apparatus of the present invention are described with reference to the method embodiments described above and are not intended to be exhaustive.
Further, it is to be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., Read Only Memory (ROM)/RAM, magnetic disk, optical disk), and includes several instructions for enabling a terminal device (e.g., a mobile phone, a computer, an all-in-one platform workstation, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A deep-learned belt tearability measurement method, comprising:
acquiring a current tearing image of a transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image;
segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of a target area;
obtaining current tearing characteristics according to a preset matrix tracking algorithm and a target area tearing stripe image;
and measuring the current tearing characteristics through a target deep learning model to obtain the tearing strength of the transmission belt.
2. The method for deeply learning belt tearability according to claim 1, wherein the obtaining a current torn image of the transmission belt, and extracting the current torn image according to a preset image ROI policy to obtain a torn length image comprises:
acquiring a current tearing image and a target belt characteristic set of a transmission belt;
performing image extraction on the current tearing image through a preset image ROI strategy according to the target belt feature set;
performing single-channel extraction on the extracted current tearing image to obtain a target channel tearing image;
and filtering the target channel tearing image through preset average filtering equipment to obtain a tearing length image.
3. The deep-learned belt tearability method of claim 2, wherein the filtering the target passageway tear image with a preset mean filtering device to obtain a tear length image comprises:
obtaining a corresponding window specification and a center coordinate according to the target channel tearing image;
carrying out average calculation on image pixel points in the window specification to obtain the average value of the current pixels;
and filtering the target channel tearing image through preset average filtering equipment according to the central coordinate and the current pixel average value to obtain a tearing length image.
4. The method for deep learning of belt tearability according to claim 1, wherein the segmenting the tear length image according to a preset laser stripe segmentation algorithm to obtain a target region tear stripe image comprises:
acquiring the acquisition direction of the current tearing image;
averagely dividing the tearing length image according to the acquisition direction to obtain a target number tearing length image;
respectively calculating the tearing length images of the target quantity to obtain pixel Lorentz information measure;
and when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold, segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area.
5. The deep-learning belt tearability measurement method according to claim 4, wherein the segmenting the tear length image according to a preset laser stripe segmentation algorithm to obtain a target area tear stripe image when the pixel lorentz information measure is less than or equal to a preset information measure threshold value comprises:
when the Lorentz information measure of the pixel is less than or equal to a preset information measure threshold value, obtaining the image gray probability according to the tearing length image;
calculating the image gray probability through a preset weighting strategy to obtain target linear weighting;
determining a corresponding segmentation threshold according to the target linear weighting;
and segmenting the tearing length image through a preset laser stripe segmentation algorithm according to the segmentation threshold to obtain a tearing stripe image of the target area.
6. The deep-learning belt tearability measurement method according to claim 1, wherein the obtaining of the current tear characteristic according to the preset matrix tracking algorithm and the target area tear stripe image comprises:
detecting the direction of the tearing stripe of the target area according to a preset matrix tracking algorithm to obtain the normal direction of the current tearing stripe;
obtaining pixel point coordinates according to the tearing stripe image of the target area;
determining a tear stripe central line according to the current tear stripe normal direction and the pixel point coordinates;
and tracking the center line of the tearing stripe through a preset matrix tracking algorithm to obtain the current tearing characteristic.
7. The deep-learned belt tearability measurement method of any one of claims 1-6, wherein the measuring the current tear characteristic by a target deep-learning model to obtain the tearability of the transmission belt comprises:
obtaining a corresponding tearing image point according to the current tearing characteristic;
performing linear fitting on the tearing image points to obtain a current tearing curve;
and when the curvature of the current tearing curve is larger than a preset curvature threshold value, measuring the current tearing curve through a target deep learning model to obtain the tearing strength of the transmission belt.
8. A deep-learned belt tear metric apparatus, comprising:
the extraction module is used for acquiring a current tearing image of the transmission belt, and extracting the current tearing image according to a preset image ROI strategy to obtain a tearing length image;
the segmentation module is used for segmenting the tearing length image according to a preset laser stripe segmentation algorithm to obtain a tearing stripe image of the target area;
the acquisition module is used for acquiring the current tearing characteristics according to a preset matrix tracking algorithm and the tearing stripe image of the target area;
and the measurement module is used for measuring the current tearing characteristics through a target deep learning model to obtain the tearing degree of the transmission belt.
9. A deep-learning belt tear metrology device, comprising: a memory, a processor, and a deep-learned belt tear metric program stored on the memory and executable on the processor, the deep-learned belt tear metric program configured with a belt tear metric method to implement the deep learning of any one of claims 1-7.
10. A storage medium having stored thereon a deep-learned belt tear measurement program that, when executed by a processor, implements a deep-learned belt tear measurement method as recited in any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111628686.5A CN113989286A (en) | 2021-12-29 | 2021-12-29 | Deep learning method, device, equipment and storage medium for belt tearing strength |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111628686.5A CN113989286A (en) | 2021-12-29 | 2021-12-29 | Deep learning method, device, equipment and storage medium for belt tearing strength |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113989286A true CN113989286A (en) | 2022-01-28 |
Family
ID=79734858
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111628686.5A Pending CN113989286A (en) | 2021-12-29 | 2021-12-29 | Deep learning method, device, equipment and storage medium for belt tearing strength |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113989286A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115947066A (en) * | 2023-03-14 | 2023-04-11 | 合肥金星智控科技股份有限公司 | Belt tearing detection method, device and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100236997A1 (en) * | 2009-03-18 | 2010-09-23 | Bowe Bell + Howell Company | Profile based laser cutting within a high-speed transport device |
| CN109409368A (en) * | 2018-11-06 | 2019-03-01 | 天地（常州）自动化股份有限公司 | Mine leather belt is vertical to tear detection device and detection method |
| CN113658136A (en) * | 2021-08-17 | 2021-11-16 | 燕山大学 | Conveyor belt defect detection method based on deep learning |
-
2021
- 2021-12-29 CN CN202111628686.5A patent/CN113989286A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100236997A1 (en) * | 2009-03-18 | 2010-09-23 | Bowe Bell + Howell Company | Profile based laser cutting within a high-speed transport device |
| CN109409368A (en) * | 2018-11-06 | 2019-03-01 | 天地（常州）自动化股份有限公司 | Mine leather belt is vertical to tear detection device and detection method |
| CN113658136A (en) * | 2021-08-17 | 2021-11-16 | 燕山大学 | Conveyor belt defect detection method based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 卢金龙: "基于机器视觉的皮带撕裂检测系统设计与实现", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115947066A (en) * | 2023-03-14 | 2023-04-11 | 合肥金星智控科技股份有限公司 | Belt tearing detection method, device and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111310645B (en) | Method, device, equipment and storage medium for warning overflow bin of goods accumulation | |
| CN108579094B (en) | User interface detection method, related device, system and storage medium | |
| US20140050387A1 (en) | System and Method for Machine Vision Inspection | |
| CN116152261B (en) | Visual inspection system for quality of printed product | |
| CN115684174A (en) | Agricultural product transportation conveyor belt safe operation monitoring method | |
| CN113989286A (en) | Deep learning method, device, equipment and storage medium for belt tearing strength | |
| CN114140684A (en) | Method, device and equipment for detecting coal blockage and coal leakage and storage medium | |
| CN114882010A (en) | Surface defect detection method based on picture recognition | |
| CN113850244A (en) | Coal conveying quantity monitoring method, device and equipment based on image recognition and storage medium | |
| CN116091503B (en) | Method, device, equipment and medium for discriminating panel foreign matter defects | |
| CN113192016A (en) | Method, device and equipment for detecting abnormal deformation of conveyor belt and storage medium | |
| CN116189093A (en) | Slag soil vehicle dust on-line monitoring method and system | |
| CN114937037B (en) | Product defect detection method, device and equipment and readable storage medium | |
| CN115965796A (en) | Metal corrosion detection method and system based on image recognition | |
| CN113989285B (en) | Belt deviation monitoring method, device and equipment based on image and storage medium | |
| CN112991308A (en) | Image quality determination method and device, electronic equipment and medium | |
| CN110634124A (en) | Method and equipment for area detection | |
| CN114926429A (en) | Method, device and equipment for detecting trace length and readable storage medium | |
| CN114549346A (en) | Blurred image recognition method, device, equipment and storage medium | |
| CN114332695A (en) | Method and device for identifying opening and closing of elevator door and storage medium | |
| CN113239832A (en) | Hidden danger intelligent identification method and system based on image identification | |
| CN114004832B (en) | Method, device and equipment for detecting abnormal vibration of coal conveying belt and storage medium | |
| CN106845479B (en) | Small-size license plate detection method based on color contrast rectangle features | |
| CN113508395A (en) | Method for detecting an object | |
| CN111709360A (en) | Safety rope wearing identification method and system |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220128 |