CN116106307B - Image recognition-based detection result evaluation method of intelligent cash dispenser - Google Patents

Image recognition-based detection result evaluation method of intelligent cash dispenser Download PDF

Info

Publication number
CN116106307B
CN116106307B CN202310330833.3A CN202310330833A CN116106307B CN 116106307 B CN116106307 B CN 116106307B CN 202310330833 A CN202310330833 A CN 202310330833A CN 116106307 B CN116106307 B CN 116106307B
Authority
CN
China
Prior art keywords
gold
purity
golden
section
generate
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
CN202310330833.3A
Other languages
Chinese (zh)
Other versions
CN116106307A (en
Inventor
黄仕峰
刘杰
刘伟程
陈超峰
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.)
Shenzhen Shangshan Intelligent Co ltd
Original Assignee
Shenzhen Shangshan Intelligent 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 Shenzhen Shangshan Intelligent Co ltd filed Critical Shenzhen Shangshan Intelligent Co ltd
Priority to CN202310330833.3A priority Critical patent/CN116106307B/en
Publication of CN116106307A publication Critical patent/CN116106307A/en
Application granted granted Critical
Publication of CN116106307B publication Critical patent/CN116106307B/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
    • 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
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/20—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by using diffraction of the radiation by the materials, e.g. for investigating crystal structure; by using scattering of the radiation by the materials, e.g. for investigating non-crystalline materials; by using reflection of the radiation by the materials
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/04—Analysing solids
    • G01N29/045—Analysing solids by imparting shocks to the workpiece and detecting the vibrations or the acoustic waves caused by the shocks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G01N29/4454—Signal recognition, e.g. specific values or portions, signal events, signatures
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/66—Trinkets, e.g. shirt buttons or jewellery items
    • 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
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Immunology (AREA)
  • Biochemistry (AREA)
  • Artificial Intelligence (AREA)
  • Pathology (AREA)
  • Analytical Chemistry (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Crystallography & Structural Chemistry (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Signal Processing (AREA)
  • Acoustics & Sound (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

The invention relates to the technical field of image recognition, in particular to a detection result evaluation method of an intelligent cash dispenser based on image recognition. The method comprises the following steps: image sampling is carried out on the section of the gold ornament so as to generate a first gold section image set, and the first gold section image set is identified through a preset gold detection neural network model so as to generate a first gold purity; performing gold detection on the gold ornaments by using X rays to generate second purity data; image sampling is carried out on the section of the golden ingot so as to generate a second golden section image set, and the second golden section image set is identified through a preset golden detection neural network model to obtain second yellow Jin Chundu; obtaining first purity data according to the first gold purity and the second gold purity; the first purity data and the second purity data are analytically compared to generate an actual purity error. The invention provides a detection result with accurate depth by carrying out image detection and identification on the golden section.

Description

Image recognition-based detection result evaluation method of intelligent cash dispenser
Technical Field
The invention relates to the technical field of image recognition, in particular to a detection result evaluation method of an intelligent cash dispenser based on image recognition.
Background
Several recovery channels are currently common for gold recovery: 01. in exchange for old: the gold first decorations are replaced with old ones, and the discount is large, and the replacement cost is about 20% -30%. 02. Department store: is recycled with degradation. 03. Classical: the recovery price is low. 04. Banking/gold coin company, etc.: gold is generally recovered only from home. The appearance of the intelligent gold recycling machine can effectively solve the pain and difficulty, not only enables the whole closed loop of the gold industry to be opened, but also derives a plurality of new projects and new services for the gold industry. A gold ecological platform can be built around the machine, and a plurality of gold-related businesses can be developed.
The gold-charging machine can effectively solve the problems of poor fluidity, inconvenient buyback experience, insufficient public confidence, high buyback cost and the like of gold in real objects, so that gold can be circulated. Specifically, the intelligent gold-charging machine is formed by developing more than 2000 precise parts, is formed by combining an identity authentication technology, an X-ray detection technology, a smelting cooling technology, a measurement weighing technology, a noble metal storage technology and the like, provides intelligent, convenient, safe and efficient gold recovery service for industry clients and terminal consumers, and redefines new standards of the gold recovery industry. However, in practical application, when gold is recovered by the gold blending machine, the purity of the gold surface can only be detected by the X-ray detection technology, and the detection precision of gold ornaments with uneven purity is poor.
Disclosure of Invention
The invention provides a detection result evaluation method of an intelligent cash dispenser based on image recognition to solve at least one technical problem.
An intelligent gold-charging machine detection result evaluation method based on image recognition comprises the following steps:
step S1: cutting the gold ornament to form a section, performing image sampling on the section of the gold ornament to generate a first gold section image set, and identifying the first gold section image set through a preset gold detection neural network model to generate a first gold purity;
step S2: performing a high temperature melting operation on the gold ornament to melt the gold ornament and cool the gold ornament to form a gold ingot, performing a gold detection operation by performing multipoint sampling on the surface of the gold ingot by X-rays, and generating second purity data;
step S3: cutting a gold ingot to form a section, performing image sampling on the section of the gold ingot to generate a second gold section image set, and identifying the second gold section image set through a preset gold detection neural network model to obtain second yellow Jin Chundu;
step S4: obtaining first purity data according to the first gold purity and the second gold purity;
Step S5: and analyzing and comparing the first purity data and the second purity data to generate an actual purity error, generating a gold detection report when the actual purity error is smaller than a preset threshold value, and generating a final measurement application request of the fire test method when the actual purity error is larger than or equal to the preset threshold value.
According to the embodiment, a plurality of sections of gold in different states are selected for section image acquisition for multiple times, a large number of section images are uploaded to a system to construct a training set of a neural network, the section images in the training set are provided with labels to show the purity, a neural network model for gold purity detection is trained through the training set, purity evaluation is carried out on the section images of gold by using the neural network model to obtain purity data, the purity data are compared with the purity data after secondary detection, when the difference value of the purity data and the purity data exceeds a fault tolerance threshold value, warning information is given to provide a medium-level measurement application request of a fire test method, so that potential economic loss is reduced, meanwhile, comprehensive analysis and comparison are carried out by combining with the surface scanning of X-rays and gold textures of sections formed after cutting, so that the problem that single detection is caused by insufficient detection precision is avoided, and accurate detection data are provided for a tester or a tester.
In one embodiment of the present disclosure, the step of cutting the gold ornament to form a cross section and image-sampling the cross section of the gold ornament to generate the first gold cross section image set in step S1 includes the steps of:
step S11: cutting the gold ornament at a first cutting angle to generate a first gold section;
step S12: carrying out image acquisition on a first golden section for multiple times through a first shooting angle to generate a first section image set;
step S13: cutting the gold ornament at a second cutting angle to generate second yellow Jin Qiemian;
step S14: performing multiple image acquisition on the second yellow Jin Qiemian through a second shooting angle to generate a second section image set;
step S15: and combining the first section image set and the second section image set to generate a first golden section image set.
According to the embodiment, the image acquisition is carried out through different cutting angles, so that errors caused by cutting at a single angle are avoided, the accuracy of data is improved, the shot image is possibly inaccurate due to cutting at a specific angle, the image acquisition is carried out through cutting at different angles, and therefore the problems of inaccurate and under-fitting of the data caused by impurity textures and the single angle are solved.
In one embodiment of the present specification, the step of constructing the preset gold purity detection neural network model includes the steps of:
step S01: acquiring a standard gold section image set and standard gold purity corresponding to the standard gold section image set;
step S02: constructing a Gaussian convolution kernel, and convolving the first golden section image set according to the Gaussian convolution kernel to generate a golden section characteristic value set;
step S03: performing dimension reduction pooling calculation on the golden section feature set to generate a golden section pooling value set;
step S04: calculating through a full connection layer according to the golden section pooling value set to generate a golden section identification value set;
step S05: and marking the golden section identification value set by the standard golden purity, thereby constructing a golden purity detection nerve model.
According to the embodiment, the golden section image set is convolved by constructing the Gaussian convolution check, so that corresponding characteristic collection is performed, data accuracy is improved through Gaussian distribution calculation, reliable data support is provided for pooling calculation and full-connection calculation, the problem of data under fitting is solved, and therefore an accurate and reliable golden model is provided.
In one embodiment of the present disclosure, the step of cutting the gold ingot to form a cross-section, and image-sampling the cross-section of the gold ingot to generate the second gold cross-section image set in step S3 includes the steps of:
Step S31: cutting the gold melt at a first cutting angle to generate a third golden section;
step S32: performing multiple image acquisition on the third golden section through a first shooting angle to generate a third section image set;
step S33: cutting the gold melt at the second cutting angle to generate a fourth golden section;
step S34: carrying out multiple image acquisition on the fourth golden section through a second shooting angle to generate a fourth section image set;
step S35: and combining the third section image set and the fourth section image set to generate a second gold section image set.
According to the embodiment, gold cutting is performed through different cutting angles, so that the influence caused by uniform impurity distribution is reduced, and accurate data support is provided for image recognition.
In one embodiment of the present specification, step S4 includes the steps of:
calculating according to the first gold purity and the second gold purity through a purity texture calculation formula to generate first purity data;
the purity texture calculation formula specifically comprises:
Figure SMS_1
Figure SMS_5
for the first purity data, < >>
Figure SMS_8
Is the first gold purity +.>
Figure SMS_11
Gold purity (I)>
Figure SMS_4
Is the first gold purity +.>
Figure SMS_9
Weight information of gold purity, +. >
Figure SMS_14
Is the +.f. in the second gold purity>
Figure SMS_16
Gold purity (I)>
Figure SMS_2
Is the +.f. in the second gold purity>
Figure SMS_6
Weight information of gold purity, +.>
Figure SMS_10
Mean error term for golden texture, < >>
Figure SMS_15
An adjustment term for the golden texture mean error term, < ->
Figure SMS_3
For adjusting the weight information of the item +.>
Figure SMS_7
For adjusting items->
Figure SMS_12
For the golden texture bias term, ++>
Figure SMS_13
Is a correction term.
According to the embodiment, the first gold purity and the second gold purity are calculated through the purity texture calculation formula, wherein the influence caused by the high-purity gold texture is fully considered, when the current detected gold purity is not enough, the result error is further increased, and further, the deep connection of information between the current gold section information and the preset standard gold section texture is provided, so that accurate and reliable data support is provided, and the precondition preparation work is carried out for the next step.
The present embodiment provides a purity texture calculation formula that fully considers the first gold purity
Figure SMS_19
Gold purity->
Figure SMS_26
First gold purity->
Figure SMS_30
Weight information of individual gold purity +.>
Figure SMS_18
The>
Figure SMS_21
Purity of gold
Figure SMS_24
The>
Figure SMS_27
Weight information of individual gold purity +.>
Figure SMS_17
Gold texture mean error term- >
Figure SMS_23
Adjusting item of golden texture average error item +.>
Figure SMS_25
Weight information of adjustment item +.>
Figure SMS_31
Regulating item->
Figure SMS_20
Gold texture bias term->
Figure SMS_22
And the interaction relationship with each other to form a functional relationship +.>
Figure SMS_28
And by modifying the term->
Figure SMS_29
The correction is made to provide accurate and reliable data support.
In one embodiment of the present specification, step S2 includes the steps of:
step S21: performing a high temperature melting operation on the gold ornament to melt the gold ornament and cooling to form a gold ingot;
step S22: transmitting X-rays to the gold ingot to obtain reflected X-rays, thereby generating reflected ray energy values and reflected ray values;
step S23: and generating second purity data according to the reflected ray energy value and the reflected ray value through a preset X-ray element calculation method.
According to the embodiment, impurities in gold ornaments are removed through high-temperature smelting operation, wherein the high-temperature smelting operation comprises smelting and oscillation, so that the influence of the impurities on gold is reduced, the purity of gold ingots formed after cooling is improved through the characteristics of different melting points and relatively stable chemical stability of gold in an actual smelting link, reflected X rays are emitted and obtained, and accurate and reliable gold purity data are provided through the different characteristics of different chemical substances on the absorption and reflection quantity of the X rays.
In one embodiment of the present specification, step S5 includes the steps of:
step S51: marking the difference between the first purity data and the second purity data to generate an actual purity error;
step S52: judging whether the actual purity error is smaller than a preset purity error threshold value or not;
step S53: when the actual purity error is determined to be smaller than a preset purity error threshold, generating a qualified gold detection report for detection result evaluation;
step S54: when the actual purity error is determined to be greater than or equal to a preset purity error threshold, generating a warning gold detection report so as to provide a final measurement application request of the fire test method.
According to the embodiment, the purity data obtained through different testing modes are compared, so that interference of impurities or inaccurate data testing results on the data is reduced, the reliability of the data is improved, image information is simply considered, the situation that the impurities are evenly distributed on the data caused by the notch is likely to be inaccurate, the result information caused by smelting is considered, the weight problem caused by the impurities is likely to be interfered, and accurate and reliable data support is provided through comparison of data difference between the impurities.
In one embodiment of the present specification, step S5 is followed by the steps of:
Step S501: the solid mass body is controlled to collide with the gold ornament and carry out audio sampling to generate gold detection audio information;
step S502: framing the golden detection audio information to generate golden detection framing audio information;
step S503: noise reduction is carried out on the gold detection framing audio information, and gold detection noise reduction information is generated;
step S504: feature extraction is carried out on the gold detection noise reduction information, and gold detection feature information is generated;
step S505: calculating the golden detection characteristic information and preset standard golden characteristic information through a golden identification detection calculation formula to generate a golden detection index;
step S506: and carrying out gold estimation report according to the gold detection index.
According to the embodiment, the solid mass is controlled to collide with the gold ornament, the audio information generated by the gold oscillation is sampled and the characteristics are extracted, the gold detection index is generated by calculating the gold detection characteristic information and the preset standard gold characteristic information through the gold identification detection calculation formula, the characteristics of the collected gold audio information and the standard gold audio information are fully considered by the gold identification detection calculation formula, and the gold detection index is accurately and reliably provided through the range of the allowable difference between the gold detection formula and the standard gold audio information, so that reliable data support of a tester or a testing machine is provided.
In one embodiment of the present specification, the golden recognition detection calculation formula is specifically:
Figure SMS_32
Figure SMS_36
for the golden detection index, < >>
Figure SMS_38
For gold detecting feature information +.>
Figure SMS_41
Gold detection feature information->
Figure SMS_35
For gold detecting feature information +.>
Figure SMS_39
Weight information of individual characteristic information, +.>
Figure SMS_43
Is +.>
Figure SMS_45
Personal characteristic information->
Figure SMS_33
Is +.>
Figure SMS_40
Weight information of individual characteristic information, +.>
Figure SMS_44
For the average change rate of golden concussion->
Figure SMS_47
For the adjustment and correction term of the mean change rate of golden oscillation, < ->
Figure SMS_34
Weight information of adjustment correction item for average change rate of golden oscillation, < ->
Figure SMS_37
For adjusting correction items->
Figure SMS_42
For adjusting items->
Figure SMS_46
Is a correction term.
The embodiment provides a golden recognition detection calculation formula which fully considers the first gold detection characteristic information
Figure SMS_51
Gold detection feature information->
Figure SMS_53
The gold detection feature information is +.>
Figure SMS_57
Weight information of individual characteristic information +.>
Figure SMS_50
The>
Figure SMS_54
Personal characteristic information->
Figure SMS_58
The>
Figure SMS_61
Weight information of individual characteristic information +.>
Figure SMS_48
Mean rate of change of golden concussion->
Figure SMS_56
Adjusting and correcting term of mean change rate of golden oscillation +.>
Figure SMS_59
Weight information of adjustment correction item of mean change rate of golden oscillation- >
Figure SMS_62
Adjusting correction term->
Figure SMS_49
Regulating item->
Figure SMS_52
And the interaction relationship with each other to form a functional relationship
Figure SMS_55
And by modifying the term->
Figure SMS_60
The correction is made to provide accurate data support.
In one embodiment of the present specification, the generating step of the preset standard golden characteristic information includes the following steps:
step S507: the solid mass body is controlled to collide with the standard gold and carry out audio sampling to generate standard gold audio information;
step S508: framing the standard golden audio information to generate standard golden frame information;
step S509: noise reduction is carried out on the standard golden frame information, and standard golden noise reduction information is generated;
step S5010: and extracting features according to the standard gold noise reduction information to generate standard gold feature information.
According to the embodiment, the solid mass body is controlled to collide with the standard gold and perform corresponding standard gold audio sampling, so that framing, noise reduction and feature extraction are performed, and reliable data support is provided for collecting and identifying the tested gold impact audio.
According to the embodiment, a plurality of sections of gold in different states are selected for section image acquisition for multiple times, a large number of section images are uploaded to a system to construct a training set of a neural network, the section images in the training set are provided with labels to show the purity of the gold, a neural network model for gold purity detection is trained through the training set, purity evaluation is carried out on the section images of gold by using the neural network model to obtain purity data, the purity data are compared with the purity data after secondary detection, when the difference value of the purity data and the purity data exceeds a fault tolerance threshold value, warning information is given to provide a medium-level measurement application request of a fire test method, potential economic loss is reduced, the interference of impurity distribution is reduced by sampling through different cutting angles during data sampling, the solid mass is controlled to collide with gold, so that generated audio information is sampled and identified, and the reliability of the gold is accurately identified and evaluated by utilizing different characteristics of audio information generated by different substances to collide with a single substance.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the detailed description of non-limiting implementations made with reference to the following drawings in which:
FIG. 1 is a flow chart showing the steps of a method for evaluating the detection result of an intelligent cash dispenser based on image recognition according to an embodiment;
FIG. 2 is a flowchart illustrating steps of a first golden section image collection sampling operation of an embodiment;
FIG. 3 is a flow chart showing the steps of a method for constructing a gold purity detection neural model according to an embodiment;
FIG. 4 is a flowchart showing steps of a second gold slice image set sampling operation of an embodiment;
FIG. 5 shows a flow chart of steps of a second purity data generation method of an embodiment;
FIG. 6 is a flow chart illustrating steps of a gold detection report generation method of an embodiment;
FIG. 7 is a flowchart illustrating steps of a golden audio detection operation according to one embodiment;
FIG. 8 is a flow chart illustrating the steps of a standard golden audio feature information generation method of an embodiment.
Detailed Description
The following is a clear and complete description of the technical method of the present patent in conjunction with the accompanying drawings, and it is evident that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
Furthermore, the drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. The functional entities may be implemented in software or in one or more hardware modules or integrated circuits or in different networks and/or processor methods and/or microcontroller methods.
It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
Referring to fig. 1 to 8, the method for evaluating the detection result of the intelligent cash dispenser based on image recognition comprises the following steps:
step S1: cutting the gold ornament to form a section, performing image sampling on the section of the gold ornament to generate a first gold section image set, and identifying the first gold section image set through a preset gold detection neural network model to generate a first gold purity;
specifically, for example, a golden ornament is cut to form a section, the section of the golden ornament is subjected to image sampling to generate a first golden section image set, the first golden section image set is identified through a preset golden detection neural network model, and first golden purity is generated, wherein the preset golden detection neural network model is calculated through a convolution neural network algorithm through image information with standard golden purity.
Step S2: performing a high temperature melting operation on the gold ornament to melt the gold ornament and cool the gold ornament to form a gold ingot, performing a gold detection operation by performing multipoint sampling on the surface of the gold ingot by X-rays, and generating second purity data;
specifically, for example, gold smelting is performed at a high temperature of 1200 ℃, and vibration is performed at a specific frequency in the smelting process, so that noble metal cooling such as air drying or natural cooling is performed in a preset time range, X-rays are emitted and received through an X-ray detector, and therefore calculation is performed in a preset calculation mode according to the quantity of the emitted and received X-rays, and second purity information is generated.
Step S3: cutting a gold ingot to form a section, performing image sampling on the section of the gold ingot to generate a second gold section image set, and identifying the second gold section image set through a preset gold detection neural network model to obtain second yellow Jin Chundu;
specifically, for example, a gold ingot is cut to form a section, a camera is controlled, a lighting lamp is activated to supplement light, an image is sampled to generate a second gold section image machine, the second gold section image machine is identified through a preset gold detection neural network model, and the second gold purity is generated, wherein the preset gold detection neural network model is constructed through a deep neural network algorithm through standard gold section image information with labels, such as convolutional kernel construction, pooling layer calculation, full connection model and weight sequence calculation, so that an accurate and reliable gold detection neural network model is provided.
Step S4: obtaining first purity data according to the first gold purity and the second gold purity;
specifically, for example, the first purity data is generated by performing weighted average calculation based on the first gold purity and the second gold purity.
Step S4: and analyzing and comparing the first purity data and the second purity data to generate an actual purity error, generating a gold detection report when the actual purity error is smaller than a preset threshold value, and generating a final measurement application request of the fire test method when the actual purity error is larger than or equal to the preset threshold value.
Specifically, for example, the difference between the first purity data and the second purity data is compared and analyzed, and the difference is marked as an actual purity error, whether the actual purity error is within a preset allowable error range or not is determined, a gold detection report is generated when the actual purity error is determined to be within the preset allowable error range, and a request for a medium-level measurement application of the fire test method is generated when the actual purity error is determined to be not within the allowable error range or more.
According to the embodiment, a plurality of sections of gold in different states are selected for section image acquisition for multiple times, a large number of section images are uploaded to a system to construct a training set of a neural network, the section images in the training set are provided with labels to show the purity, a neural network model for gold purity detection is trained through the training set, purity evaluation is carried out on the section images of gold by using the neural network model to obtain purity data, the purity data are compared with the purity data after secondary detection, when the difference value of the purity data and the purity data exceeds a fault tolerance threshold value, warning information is given to provide a medium-level measurement application request of a fire test method, so that potential economic loss is reduced, meanwhile, comprehensive analysis and comparison are carried out by combining with the surface scanning of X-rays and gold textures of sections formed after cutting, so that the problem that single detection is caused by insufficient detection precision is avoided, and accurate detection data are provided for a tester or a tester.
In one embodiment of the present disclosure, the step of cutting the gold ornament to form a cross section and image-sampling the cross section of the gold ornament to generate the first gold cross section image set in step S1 includes the steps of:
step S11: cutting the gold ornament at a first cutting angle to generate a first gold section;
specifically, for example, the golden ornament is cut at a first cutting angle, such as an angle of 60 degrees with respect to the horizontal plane, to produce a first golden section.
Step S12: carrying out image acquisition on a first golden section for multiple times through a first shooting angle to generate a first section image set;
specifically, for example, the camera is controlled to perform multi-point image acquisition at an angle congruent with the first cutting angle or perpendicular to the first golden section, and a first section image set is generated.
Step S13: cutting the gold ornament at a second cutting angle to generate second yellow Jin Qiemian;
specifically, for example, the gold jewelry is cut at a second cutting angle, such as 90 degrees, resulting in a second yellow Jin Qiemian.
Step S14: performing multiple image acquisition on the second yellow Jin Qiemian through a second shooting angle to generate a second section image set;
specifically, for example, the camera is controlled to perform multi-point image acquisition at an angle congruent with the second cutting angle or perpendicular to the second yellow Jin Qiemian, generating a second cut-surface image set.
Step S15: and combining the first section image set and the second section image set to generate a first golden section image set.
Specifically, for example, the first slice image set and the second slice image set are randomly arranged in a sequential order or by a random number to generate the first golden section image set.
According to the embodiment, the image acquisition is carried out through different cutting angles, so that errors caused by cutting at a single angle are avoided, the accuracy of data is improved, the shot image is possibly inaccurate due to cutting at a specific angle, the image acquisition is carried out through cutting at different angles, and therefore the problems of inaccurate and under-fitting of the data caused by impurity textures and the single angle are solved.
In one embodiment of the present specification, the step of constructing the preset gold purity detection neural network model includes the steps of:
step S01: acquiring a standard gold section image set and standard gold purity corresponding to the standard gold section image set;
specifically, for example, a standard golden section image set and standard golden purities corresponding thereto, such as 99%, 99.9% and 99.99% standard golden purities, are acquired.
Step S02: constructing a Gaussian convolution kernel, and convolving the first golden section image set according to the Gaussian convolution kernel to generate a golden section characteristic value set;
Specifically, for example, a gaussian convolution kernel is constructed, for example, (1/2, 1/3,1/2, 1/3), where the offset is adapted to the image size, for example, when the resolution is less than 1024×720, the calculated offset is set to 1, and when the resolution is greater than 1024×720, the calculated offset is set to 2.
Step S03: performing dimension reduction pooling calculation on the golden section feature set to generate a golden section pooling value set;
specifically, the pooling layer is set to (0, 1/9,0,1/9,1,1/9,0,1/9, 0), for example, for calculation.
Step S04: calculating through a full connection layer according to the golden section pooling value set to generate a golden section identification value set;
specifically, for example, according to the golden section pooling value set, the golden section identification value set is generated by calculating through a preset full-connection layer weight sequence.
Step S05: and marking the golden section identification value set by the standard golden purity, thereby constructing a golden purity detection nerve model.
Specifically, for example, standard gold purity such as gold, impurities, or gold purity information such as 99%, 99.9%, and 99.99% are labeled with a gold section identification value set, thereby constructing a gold purity detection neural model.
According to the embodiment, the golden section image set is convolved by constructing the Gaussian convolution check, so that corresponding characteristic collection is performed, data accuracy is improved through Gaussian distribution calculation, reliable data support is provided for pooling calculation and full-connection calculation, the problem of data under fitting is solved, and therefore an accurate and reliable golden model is provided.
In one embodiment of the present disclosure, the step of cutting the gold ingot to form a cross-section, and image-sampling the cross-section of the gold ingot to generate the second gold cross-section image set in step S3 includes the steps of:
step S31: cutting the gold melt at a first cutting angle to generate a third golden section;
specifically, for example, the golden ornament is cut at a first cutting angle, such as an angle of 60 degrees with respect to the horizontal plane, to produce a third golden section.
Step S32: performing multiple image acquisition on the third golden section through a first shooting angle to generate a third section image set;
specifically, for example, the camera is controlled to perform multi-point image acquisition at an angle congruent with the first cutting angle or perpendicular to the first golden section, and a first section image set is generated.
Step S33: cutting the gold melt at the second cutting angle to generate a fourth golden section;
specifically, the golden melt is cut, for example, at a second cutting angle, such as 90 degrees, to create a fourth golden section.
Step S34: carrying out multiple image acquisition on the fourth golden section through a second shooting angle to generate a fourth section image set;
specifically, for example, the camera is controlled to perform multi-point image acquisition at an angle which is congruent with the second cutting angle or perpendicular to the fourth golden section, and a fourth section image set is generated.
Step S35: and combining the third section image set and the fourth section image set to generate a second gold section image set.
Specifically, for example, the third section image set and the fourth section image set are randomly arranged in a sequential order or by a random number to generate the first golden section image set.
According to the embodiment, gold cutting is performed through different cutting angles, so that the influence caused by uniform impurity distribution is reduced, and accurate data support is provided for image recognition.
In one embodiment of the present specification, step S4 includes the steps of:
calculating according to the first gold purity and the second gold purity through a purity texture calculation formula to generate first purity data;
specifically, the first purity data is generated, for example, by calculation through a purity texture calculation formula based on a first gold purity, e.g., 99.65%, and a second gold purity, e.g., 99.87%.
The purity texture calculation formula specifically comprises:
Figure SMS_63
Figure SMS_66
for the first purity data, < >>
Figure SMS_68
Is the first gold purity +.>
Figure SMS_72
Gold purity (I)>
Figure SMS_67
Is the first gold purity +.>
Figure SMS_69
Weight information of gold purity, +.>
Figure SMS_75
Is the +.f. in the second gold purity>
Figure SMS_77
Gold purity (I)>
Figure SMS_64
Is the +.f. in the second gold purity >
Figure SMS_71
Weight information of gold purity, +.>
Figure SMS_73
Mean error term for golden texture, < >>
Figure SMS_74
An adjustment term for the golden texture mean error term, < ->
Figure SMS_65
For adjusting the weight information of the item +.>
Figure SMS_70
For adjusting items->
Figure SMS_76
For the golden texture bias term, ++>
Figure SMS_78
Is a correction term.
According to the embodiment, the first gold purity and the second gold purity are calculated through the purity texture calculation formula, wherein the influence caused by the high-purity gold texture is fully considered, when the current detected gold purity is not enough, the result error is further increased, and further, the deep connection of information between the current gold section information and the preset standard gold section texture is provided, so that accurate and reliable data support is provided, and the precondition preparation work is carried out for the next step.
The present embodiment provides a purity texture calculation formula that fully considers the first gold purity
Figure SMS_80
Gold purity->
Figure SMS_83
First gold purity->
Figure SMS_87
Weight information of individual gold purity +.>
Figure SMS_82
The>
Figure SMS_85
Purity of gold
Figure SMS_86
The>
Figure SMS_90
Weight information of individual gold purity +.>
Figure SMS_79
Gold texture mean error term->
Figure SMS_84
Adjusting item of golden texture average error item +.>
Figure SMS_89
Weight information of adjustment item +. >
Figure SMS_92
Regulating item->
Figure SMS_81
Gold texture bias term->
Figure SMS_88
And the interaction relationship with each other to form a functional relationship +.>
Figure SMS_91
And by modifying the term->
Figure SMS_93
The correction is made to provide accurate and reliable data support.
In one embodiment of the present specification, step S2 includes the steps of:
step S21: performing a high temperature melting operation on the gold ornament to melt the gold ornament and cooling to form a gold ingot;
specifically, a high-temperature melting operation is performed on gold ornaments, for example, by a high temperature of 2100 degrees to melt the gold ornaments, and cooled by natural air drying to form gold ingots.
Step S22: transmitting X-rays to the gold ingot to obtain reflected X-rays, thereby generating reflected ray energy values and reflected ray values;
specifically, for example, X-rays are emitted to a gold ingot to obtain reflected X-rays, thereby generating reflected ray energy values and reflected ray values.
Step S23: and generating second purity data according to the reflected ray energy value and the reflected ray value through a preset X-ray element calculation method.
Specifically, the second purity data is generated, for example, from the reflected radiation energy value and the reflected radiation value, displayed and analyzed by an automatic display numerical instrument.
According to the embodiment, impurities in gold ornaments are removed through high-temperature smelting operation, wherein the high-temperature smelting operation comprises smelting and oscillation, so that the influence of the impurities on gold is reduced, the purity of gold ingots formed after cooling is improved through the characteristics of different melting points and relatively stable chemical stability of gold in an actual smelting link, reflected X rays are emitted and obtained, and accurate and reliable gold purity data are provided through the different characteristics of different chemical substances on the absorption and reflection quantity of the X rays.
In one embodiment of the present specification, step S5 includes the steps of:
step S51: marking the difference between the first purity data and the second purity data to generate an actual purity error;
specifically, for example, the difference between the first purity data, e.g., 99.57%, and the second purity data, e.g., 99.68%, is marked to produce an actual purity error, e.g., 0.11%.
Step S52: judging whether the actual purity error is smaller than a preset purity error threshold value or not;
specifically, for example, it is judged whether or not an actual purity error such as 0.11% is smaller than a preset purity error threshold such as 0.5%.
Step S53: when the actual purity error is determined to be smaller than a preset purity error threshold, generating a qualified gold detection report for detection result evaluation;
Specifically, for example, when it is determined that the actual purity error, e.g., 0.11%, is less than the preset purity error threshold, e.g., 0.5%, then a qualified gold test report is generated for evaluation of the test results.
Step S54: when the actual purity error is determined to be greater than or equal to a preset purity error threshold, generating a warning gold detection report so as to provide a final measurement application request of the fire test method.
Specifically, for example, when it is determined that the actual purity error, such as 2.1%, is greater than or equal to a preset purity error threshold, such as 0.5%, a warning gold detection report is generated to make a final measurement application request of the fire test method.
According to the embodiment, the purity data obtained through different testing modes are compared, so that interference of impurities or inaccurate data testing results on the data is reduced, the reliability of the data is improved, image information is simply considered, the situation that the impurities are evenly distributed on the data caused by the notch is likely to be inaccurate, the result information caused by smelting is considered, the weight problem caused by the impurities is likely to be interfered, and accurate and reliable data support is provided through comparison of data difference between the impurities.
In one embodiment of the present specification, step S5 is followed by the steps of:
Step S501: the solid mass body is controlled to collide with the gold ornament and carry out audio sampling to generate gold detection audio information;
specifically, for example, a solid body such as standard 1KG iron is controlled to collide with gold ornaments and perform audio sampling through a microphone or an audio sampling device, so as to generate gold detection audio information.
Step S502: framing the golden detection audio information to generate golden detection framing audio information;
specifically, for example, the golden detection frame audio information is generated by framing golden detection audio information in one frame of 20 ms.
Step S503: noise reduction is carried out on the gold detection framing audio information, and gold detection noise reduction information is generated;
specifically, for example, a noise reducer is constructed to perform noise reduction operation on the golden detection framing audio information to generate golden detection noise reduction information.
Step S504: feature extraction is carried out on the gold detection noise reduction information, and gold detection feature information is generated;
specifically, the gold detection noise reduction information is subjected to feature extraction, for example, by an MFCC algorithm, and gold detection feature information is generated.
Step S505: calculating the golden detection characteristic information and preset standard golden characteristic information through a golden identification detection calculation formula to generate a golden detection index;
Specifically, for example, the golden detection characteristic information and the preset standard golden characteristic information are calculated by the golden recognition detection calculation formula in the other embodiments, so as to generate a golden detection index, such as 4.35.
Step S506: and carrying out gold estimation report according to the gold detection index.
Specifically, for example, the greater the golden test index, the less pure the golden component, and when the golden test index is determined to be less than or equal to a predetermined golden test threshold index, such as 3.5, such as 2.25, then a qualified golden test report is generated.
According to the embodiment, the solid mass is controlled to collide with the gold ornament, the audio information generated by the gold oscillation is sampled and the characteristics are extracted, the gold detection index is generated by calculating the gold detection characteristic information and the preset standard gold characteristic information through the gold identification detection calculation formula, the characteristics of the collected gold audio information and the standard gold audio information are fully considered by the gold identification detection calculation formula, and the gold detection index is accurately and reliably provided through the range of the allowable difference between the gold detection formula and the standard gold audio information, so that reliable data support of a tester or a testing machine is provided.
In one embodiment of the present specification, the golden recognition detection calculation formula is specifically:
Figure SMS_94
Figure SMS_98
For the golden detection index, < >>
Figure SMS_102
For gold detecting feature information +.>
Figure SMS_108
Gold detection feature information->
Figure SMS_97
For gold detecting feature information +.>
Figure SMS_101
Weight information of individual characteristic information, +.>
Figure SMS_103
Is +.>
Figure SMS_107
Personal characteristic information->
Figure SMS_95
Is +.>
Figure SMS_99
Weight information of individual characteristic information, +.>
Figure SMS_104
For the average change rate of golden concussion->
Figure SMS_106
For the adjustment and correction term of the mean change rate of golden oscillation, < ->
Figure SMS_96
Weight information of adjustment correction item for average change rate of golden oscillation, < ->
Figure SMS_100
For adjusting correction items->
Figure SMS_105
For adjusting items->
Figure SMS_109
Is a correction term.
The embodiment provides a golden recognition detection calculation formula which fully considers the first gold detection characteristic information
Figure SMS_111
Gold detection feature information->
Figure SMS_117
The gold detection feature information is +.>
Figure SMS_121
Weight information of individual characteristic information +.>
Figure SMS_113
The>
Figure SMS_114
Personal characteristic information->
Figure SMS_118
The>
Figure SMS_124
Weight information of individual characteristic information +.>
Figure SMS_110
Mean rate of change of golden concussion->
Figure SMS_116
Adjusting and correcting term of mean change rate of golden oscillation +.>
Figure SMS_119
Weight information of adjustment correction item of mean change rate of golden oscillation->
Figure SMS_123
Adjusting correction term->
Figure SMS_112
Regulating item->
Figure SMS_115
And the interaction relationship with each other to form a functional relationship
Figure SMS_120
And by modifying the term->
Figure SMS_122
The correction is made to provide accurate data support.
In one embodiment of the present specification, the generating step of the preset standard golden characteristic information includes the following steps:
step S507: the solid mass body is controlled to collide with the standard gold and carry out audio sampling to generate standard gold audio information;
specifically, for example, a solid body such as a standard 1KG iron ball is controlled to collide with a standard gold body and perform audio sampling through a microphone or other audio sampling devices, so as to generate standard gold audio information.
Step S508: framing the standard golden audio information to generate standard golden frame information;
specifically, for example, standard golden audio information is framed in one frame of 20ms to generate standard golden section frame information.
Step S509: noise reduction is carried out on the standard golden frame information, and standard golden noise reduction information is generated;
specifically, for example, a noise reducer such as linear noise reduction is constructed, and standard golden frame information is noise-reduced by the noise reducer to generate standard golden noise reduction information.
Step S5010: and extracting features according to the standard gold noise reduction information to generate standard gold feature information.
Specifically, feature extraction is performed through mfcc algorithm according to standard golden noise reduction information, for example, so as to generate standard golden feature information.
According to the embodiment, the solid mass body is controlled to collide with the standard gold and perform corresponding standard gold audio sampling, so that framing, noise reduction and feature extraction are performed, and reliable data support is provided for collecting and identifying the tested gold impact audio.
According to the embodiment, a plurality of sections of gold in different states are selected for section image acquisition for multiple times, a large number of section images are uploaded to a system to construct a training set of a neural network, the section images in the training set are provided with labels to show the purity of the gold, a neural network model for gold purity detection is trained through the training set, purity evaluation is carried out on the section images of gold by using the neural network model to obtain purity data, the purity data are compared with the purity data after secondary detection, when the difference value of the purity data and the purity data exceeds a fault tolerance threshold value, warning information is given to provide a medium-level measurement application request of a fire test method, potential economic loss is reduced, the interference of impurity distribution is reduced by sampling through different cutting angles during data sampling, the solid mass is controlled to collide with gold, so that generated audio information is sampled and identified, and the reliability of the gold is accurately identified and evaluated by utilizing different characteristics of audio information generated by different substances to collide with a single substance.
The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (6)

1. The detection result evaluation method of the intelligent cash dispenser based on image recognition is characterized by comprising the following steps of:
step S1: cutting the gold ornament to form a section, performing image sampling on the section of the gold ornament to generate a first gold section image set, and identifying the first gold section image set through a preset gold detection neural network model to generate a first gold purity; wherein the step of cutting the golden ornament to form a cross section, and image sampling the cross section of the golden ornament to generate a first golden section image set comprises the steps of:
Step S11: cutting the gold ornament at a first cutting angle to generate a first gold section;
step S12: carrying out image acquisition on a first golden section for multiple times through a first shooting angle to generate a first section image set;
step S13: cutting the gold ornament at a second cutting angle to generate second yellow Jin Qiemian;
step S14: performing multiple image acquisition on the second yellow Jin Qiemian through a second shooting angle to generate a second section image set;
step S15: combining the first section image set and the second section image set to generate a first golden section image set;
the construction step of the preset gold purity detection neural network model comprises the following steps of:
acquiring a standard gold section image set and standard gold purity corresponding to the standard gold section image set;
constructing a Gaussian convolution kernel, and convolving the first golden section image set according to the Gaussian convolution kernel to generate a golden section characteristic value set;
performing dimension reduction pooling calculation on the golden section feature set to generate a golden section pooling value set;
calculating through a full connection layer according to the golden section pooling value set to generate a golden section identification value set;
marking the golden section identification value set by standard golden purity, thereby constructing a golden purity detection nerve model;
Step S2: performing a high temperature melting operation on the gold ornament to melt the gold ornament and cool the gold ornament to form a gold ingot, performing a gold detection operation on the gold ingot by performing multipoint sampling on the gold ingot by X-rays, and generating second purity data;
step S3: cutting a gold ingot to form a section, performing image sampling on the section of the gold ingot to generate a second gold section image set, and identifying the second gold section image set through a preset gold detection neural network model to obtain second yellow Jin Chundu; wherein the step of cutting the gold ingot to form a cross section, and image sampling the cross section of the gold ingot to generate a second gold cross section image set comprises the steps of:
step S31: cutting the gold melt at a first cutting angle to generate a third golden section;
step S32: performing multiple image acquisition on the third golden section through a first shooting angle to generate a third section image set;
step S33: cutting the gold melt at the second cutting angle to generate a fourth golden section;
step S34: carrying out multiple image acquisition on the fourth golden section through a second shooting angle to generate a fourth section image set;
step S35: combining the third section image set and the fourth section image set to generate a second gold section image set;
Step S4: calculating according to the first gold purity and the second gold purity through a purity texture calculation formula to generate first purity data;
the purity texture calculation formula specifically comprises:
Figure QLYQS_1
；
Figure QLYQS_5
for the first purity data, < >>
Figure QLYQS_6
Is the first gold purity +.>
Figure QLYQS_10
Gold purity (I)>
Figure QLYQS_4
Is the first gold purity +.>
Figure QLYQS_8
Weight information of gold purity, +.>
Figure QLYQS_11
Is the +.f. in the second gold purity>
Figure QLYQS_14
Gold purity (I)>
Figure QLYQS_2
Is the +.f. in the second gold purity>
Figure QLYQS_7
Weight information of gold purity, +.>
Figure QLYQS_13
Mean error term for golden texture, < >>
Figure QLYQS_15
An adjustment term for the golden texture mean error term, < ->
Figure QLYQS_3
For adjusting the weight information of the item +.>
Figure QLYQS_9
For adjusting items->
Figure QLYQS_12
For the golden texture bias term, ++>
Figure QLYQS_16
Is a correction term;
step S5: and analyzing and comparing the first purity data and the second purity data to generate an actual purity error, generating a gold detection report when the actual purity error is smaller than a preset threshold value, and generating a final measurement application request of the fire test method when the actual purity error is larger than or equal to the preset threshold value.
2. The method according to claim 1, wherein step S2 comprises the steps of:
performing a high temperature melting operation on the gold ornament to melt the gold ornament and cooling to form a gold ingot;
Transmitting X-rays to the gold ingot to obtain reflected X-rays, thereby generating reflected ray energy values and reflected ray values;
and generating second purity data according to the reflected ray energy value and the reflected ray value through a preset X-ray element calculation method.
3. The method according to claim 1, wherein step S5 comprises the steps of:
marking the difference between the first purity data and the second purity data to generate an actual purity error;
judging whether the actual purity error is smaller than a preset purity error threshold value or not;
when the actual purity error is determined to be smaller than a preset purity error threshold, generating a qualified gold detection report for detection result evaluation;
when the actual purity error is determined to be greater than or equal to a preset purity error threshold, generating a warning gold detection report so as to provide a final measurement application request of the fire test method.
4. The method according to claim 1, characterized in that after step S5 further comprises the steps of:
step S501: the solid mass body is controlled to collide with the gold ornament and carry out audio sampling to generate gold detection audio information;
step S502: framing the golden detection audio information to generate golden detection framing audio information;
Step S503: noise reduction is carried out on the gold detection framing audio information, and gold detection noise reduction information is generated;
step S504: feature extraction is carried out on the gold detection noise reduction information, and gold detection feature information is generated;
step S505: calculating the golden detection characteristic information and preset standard golden characteristic information through a golden identification detection calculation formula to generate a golden detection index;
step S506: and carrying out gold estimation report according to the gold detection index.
5. The method of claim 4, wherein the golden recognition detection calculation formula is specifically:
Figure QLYQS_17
；
Figure QLYQS_19
for the golden detection index, < >>
Figure QLYQS_23
For gold detecting feature information +.>
Figure QLYQS_29
Gold detection feature information->
Figure QLYQS_21
For gold detecting feature information +.>
Figure QLYQS_25
Weight information of individual characteristic information, +.>
Figure QLYQS_27
Is +.>
Figure QLYQS_32
Personal characteristic information->
Figure QLYQS_18
Is +.>
Figure QLYQS_22
Weight information of individual characteristic information, +.>
Figure QLYQS_26
For the average change rate of golden concussion->
Figure QLYQS_30
For the adjustment and correction term of the mean change rate of golden oscillation, < ->
Figure QLYQS_20
Weight information of adjustment correction item for average change rate of golden oscillation, < ->
Figure QLYQS_24
For adjusting correction items->
Figure QLYQS_28
For adjusting items->
Figure QLYQS_31
Is a correction term.
6. The method of claim 4, wherein the generating step of the predetermined standard golden characteristic information comprises the steps of:
Step S507: the solid mass body is controlled to collide with the standard gold and carry out audio sampling to generate standard gold audio information;
step S508: framing the standard golden audio information to generate standard golden frame information;
step S509: noise reduction is carried out on the standard golden frame information, and standard golden noise reduction information is generated;
step S5010: and extracting features according to the standard gold noise reduction information to generate standard gold feature information.
CN202310330833.3A 2023-03-31 2023-03-31 Image recognition-based detection result evaluation method of intelligent cash dispenser Active CN116106307B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310330833.3A CN116106307B (en) 2023-03-31 2023-03-31 Image recognition-based detection result evaluation method of intelligent cash dispenser

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310330833.3A CN116106307B (en) 2023-03-31 2023-03-31 Image recognition-based detection result evaluation method of intelligent cash dispenser

Publications (2)

Publication Number Publication Date
CN116106307A CN116106307A (en) 2023-05-12
CN116106307B true CN116106307B (en) 2023-06-30

Family

ID=86262338

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310330833.3A Active CN116106307B (en) 2023-03-31 2023-03-31 Image recognition-based detection result evaluation method of intelligent cash dispenser

Country Status (1)

Country Link
CN (1) CN116106307B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116577363B (en) * 2023-07-04 2023-10-17 深圳上善智能有限公司 Gold multi-point X fluorescence detection method based on machine vision

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS58202866A (en) * 1982-05-21 1983-11-26 Tanaka Kikinzoku Kogyo Kk Judging method of gold ingot
CN104198399A (en) * 2014-08-22 2014-12-10 南京市产品质量监督检验院 Metallographic analysis method for gold detection
WO2019083476A2 (en) * 2017-05-05 2019-05-02 Sicpa Turkey Ürün Güvenli̇ği̇ Sanayi̇ Ve Ti̇caret Anoni̇m Şi̇rketi̇ Noble metal products identification and tracking system
WO2021057810A1 (en) * 2019-09-29 2021-04-01 深圳数字生命研究院 Data processing method, data training method, data identifying method and device, and storage medium
CN112630088A (en) * 2020-12-28 2021-04-09 国家珠宝检测中心(广东)有限责任公司 Low-purity gold quantitative detection method

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7312087B2 (en) * 2000-01-11 2007-12-25 Clinical Micro Sensors, Inc. Devices and methods for biochip multiplexing
DE10231228B4 (en) * 2002-07-11 2004-09-30 Guido Koschany Recovery of valuable materials from spark plugs and glow plugs
JP4266039B2 (en) * 2008-05-22 2009-05-20 京都ブラス株式会社 Method for producing lead-free free-cutting brass alloy
CN105784736A (en) * 2016-03-14 2016-07-20 江苏大学 Scanning electron microscope-energy dispersive X-ray spectroscopy (SEM-EDX) detection method for distribution characteristics of nitrogen, phosphorus and potassium in crop leaf
US20210142241A1 (en) * 2017-03-09 2021-05-13 Queen's University At Kingston Operations Management Methods and Products
CN108856753B (en) * 2018-08-22 2023-12-22 华南理工大学 Micro-texture cutter based on silicon brass tissue structure and processing method and application thereof
US11189021B2 (en) * 2018-11-16 2021-11-30 Align Technology, Inc. Machine based three-dimensional (3D) object defect detection
CN109829427B (en) * 2019-01-31 2022-05-17 福州大学 Face clustering method based on purity detection and spatial attention network
CN111046883B (en) * 2019-12-05 2022-08-23 吉林大学 Intelligent assessment method and system based on ancient coin image
US12017301B2 (en) * 2020-03-13 2024-06-25 General Electric Company Systems and methods for compression, management, and analysis of downbeam camera data for an additive machine
CN111754516B (en) * 2020-05-25 2023-06-30 沈阳工程学院 Intelligent control method for rutile monocrystal growth based on computer vision feedback
KR102218330B1 (en) * 2020-12-29 2021-02-22 주식회사 다이아앤코 method and apparatus for manufacturing high-hardness diamond simulant by cutting a gemstone into 100-sided body
US20220063135A1 (en) * 2020-09-03 2022-03-03 Dia&Co Inc. Method and apparatus for manufacturing high-hardness diamond simulant by cutting a gemstone into 100-sided body
CN113252585B (en) * 2021-04-21 2024-02-09 深圳市海谱纳米光学科技有限公司 Method and device for judging gold surface coating based on hyperspectral image
US20230023641A1 (en) * 2021-07-22 2023-01-26 X Development Llc Automated detection of chemical component of moving object
CN113947690A (en) * 2021-10-14 2022-01-18 中国平安财产保险股份有限公司 Method, device and equipment for generating service list data and storage medium
CN114769162A (en) * 2022-06-08 2022-07-22 深圳上善智能有限公司 Intelligent recovery device and method based on valuables identification

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS58202866A (en) * 1982-05-21 1983-11-26 Tanaka Kikinzoku Kogyo Kk Judging method of gold ingot
CN104198399A (en) * 2014-08-22 2014-12-10 南京市产品质量监督检验院 Metallographic analysis method for gold detection
WO2019083476A2 (en) * 2017-05-05 2019-05-02 Sicpa Turkey Ürün Güvenli̇ği̇ Sanayi̇ Ve Ti̇caret Anoni̇m Şi̇rketi̇ Noble metal products identification and tracking system
WO2021057810A1 (en) * 2019-09-29 2021-04-01 深圳数字生命研究院 Data processing method, data training method, data identifying method and device, and storage medium
CN112630088A (en) * 2020-12-28 2021-04-09 国家珠宝检测中心(广东)有限责任公司 Low-purity gold quantitative detection method

Also Published As

Publication number Publication date
CN116106307A (en) 2023-05-12

Similar Documents

Publication Publication Date Title
CN112115895B (en) Pointer type instrument reading identification method, pointer type instrument reading identification device, computer equipment and storage medium
CN116106307B (en) Image recognition-based detection result evaluation method of intelligent cash dispenser
US8750570B2 (en) Methods, systems and apparatus for automated authentication
Steele et al. White dwarfs in the UKIRT Infrared Deep Sky Survey Large Area Survey: the substellar companion fraction
Christlieb et al. The stellar content of the Hamburg/ESO survey-I. Automated selection of DA white dwarfs
Williams et al. T-RaMiSu: the Two-meter Radio Mini Survey-I. The Boötes Field
CN113674273B (en) Optical detection method and system based on product defects and readable storage medium
Kampczyk et al. Simulating the Cosmos:* The Fraction of Merging Galaxies at High Redshift
CN111738349B (en) Detection effect evaluation method and device of target detection algorithm, storage medium and equipment
Anthony-Twarog et al. WIYN open cluster study. LXXVI. Li evolution among stars of low/intermediate mass: the metal-deficient open cluster NGC 2506
CN115015286B (en) Chip detection method and system based on machine vision
Górny et al. Planetary nebulae in the direction of the Galactic bulge: on nebulae with emission-line central stars
CN106528665B (en) AOI equipment test file searching method and system
CN106204549B (en) A kind of billboard monitoring method, device and electronic equipment based on video analysis
Rasekh et al. The Lyman Alpha Reference Sample-XII. Morphology of extended Lyman alpha emission in star-forming galaxies
Tout Automatic vision system for surface inspection and monitoring: Application to wheel inspection
CN110017998A (en) Vehicle checking method, device and equipment
CN114596296A (en) High-sensitivity hot-rolled steel coil end surface defect identification system and method
CN105954536A (en) Optical flow speed measurement module and velocity measurement method
Boulyga et al. Automation system for optical counting of nuclear tracks
CN111122454A (en) Apple brittleness nondestructive detector based on optical fiber spectrum and machine vision
Fouts et al. Evolving robust gender classification features for CAESAR data
Jo et al. Inference for integer-valued AR model with zero-inflated Poisson innovations
CN117523324B (en) Image processing method and image sample classification method, device and storage medium
KR102686135B1 (en) Method of inspecting defect using artificial intelligence and apparatus for inspecting defect using artificial intelligence

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