CN109613108A - Threshold selection method and equipment - Google Patents
Threshold selection method and equipment Download PDFInfo
- Publication number
- CN109613108A CN109613108A CN201811610109.1A CN201811610109A CN109613108A CN 109613108 A CN109613108 A CN 109613108A CN 201811610109 A CN201811610109 A CN 201811610109A CN 109613108 A CN109613108 A CN 109613108A
- Authority
- CN
- China
- Prior art keywords
- product
- sampled data
- group
- metalliferous
- metal
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
- G01N27/72—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
- G01V3/38—Processing data, e.g. for analysis, for interpretation, for correction
-
- 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
Abstract
The invention discloses a kind of threshold selection method and equipment, comprising: obtains one group of metal-free product sampled data and at least one set of metalliferous product sampled data;According to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data, the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous corresponding omission factor group of product sampled data is calculated;According to the requirement of product false detection rate group, every group of metalliferous product sampled data corresponding omission factor group and erroneous detection omission factor, select corresponding numerical value as threshold value.The present invention is according to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data, suitable threshold value is selected according to the requirement of false detection rate, omission factor, the selection of threshold value is set to obtain effective quantized data support, the threshold value of selection more meets actual use demand, it reduces false detection rate, guarantee omission factor, improve the monitoring to product quality.
Description
Technical field
The present invention relates to metal detector field more particularly to a kind of threshold selection method and equipment.
Background technique
Many fields (such as: food, pharmacy, rubber, chemical industry etc.) all can may be mixed into production in monitor production process now
Metallic foreign body in product, principle are: when being mixed with the metallic foreign body of certain size in product, metal detector will generate one
A bigger signal makes the metal detector different to metal to realize by the signal threshold value for setting different when users use
The monitoring and judgement of object.
For example, the signal value of normal product may beat within 100, (the ratio when being mixed with the metallic foreign body of certain size
Such as 1.5mm), by generate it is one bigger may be 300 or even 500 signal.At this point, if the threshold value of setting is 300,
Any signal value is more than 300 product, is considered to unqualified, i.e., metal detector determines that it has metal;Meanwhile it is because various
The influence of reason, even if the same foreign matter iterates through metal detector, the signal value generated every time is also different.
Present threshold value setting is usually all the experience self-setting according to user, but this threshold selected by rule of thumb
Value in use because the precision of metal detector is different, use environment is different and everyone feel it is different etc. it is various not
It is determining, be difficult to the factor quantified influence, there is very big unstability, without specific, quantifiable foundation, be unable to satisfy
Low false detection rate, low omission factor high-quality monitoring demand.
Summary of the invention
The object of the present invention is to provide a kind of threshold selection method and equipment, select threshold value according to reliable quantized data,
Not only meet low false detection rate, but also realize the demand of the angles of science monitoring of low omission factor.
Technical solution provided by the invention is as follows:
A kind of threshold selection method, comprising: obtain one group of metal-free product sampled data and at least one set of containing metal
Product sampled data;According to the distribution of metal-free product sampled data and the metalliferous product sampled data
Rule, is calculated the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product is adopted
The corresponding omission factor group of sample data;It is respectively right according to the product false detection rate group, every group of metalliferous product sampled data
The requirement of the omission factor group and erroneous detection omission factor answered, selects corresponding numerical value as threshold value.
In the above-mentioned technical solutions, according to point of metal-free product sampled data and metalliferous product sampled data
Cloth rule selects suitable threshold value according to the requirement of false detection rate, omission factor, the selection of threshold value is made to obtain effective quantized data branch
Support, the threshold value of selection more meet actual use demand, reduce false detection rate, omission factor, improve the monitoring efficiency and product of product quality
The quality of matter management work.
Further, point according to metal-free product sampled data and the metalliferous product sampled data
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampling is calculated in cloth rule
The corresponding omission factor group of data includes: when metal-free product sampled data and the metalliferous product sampling
When data fit normal distribution law, according to default normal state formula, preset bounds parameter, metal-free product hits
According to each metalliferous product sampled data, the corresponding product of metal-free product sampled data is calculated and misses
Inspection rate group and every group of metalliferous corresponding omission factor group of product sampled data.
In the above-mentioned technical solutions, when meeting normal distribution law, default normal state formula can be used and calculate omission factor group
With false detection rate group.
Further, the basis presets normal state formula, preset bounds parameter, metal-free product sampled data
With each metalliferous product sampled data, the corresponding product erroneous detection of metal-free product sampled data is calculated
Rate group and every group of metalliferous corresponding omission factor group of product sampled data include: to be adopted according to metal-free product
Sample data, calculate corresponding product (signal) average value and product standard is poor;The metalliferous product sampling according to every group
Data, calculate every group described in the metalliferous corresponding metal of product sampled data (signal) average value and metal master
Difference;According to the product average value, the product standard is poor, the preset bounds parameter and default normal state formula, is calculated
The corresponding product false detection rate group of metal-free product sampled data;It is public according to the preset bounds parameter, default normal state
The corresponding metal average value of metalliferous product sampled data described in formula and every group and metal master are poor, calculate separately to obtain
The corresponding omission factor group of every group of metalliferous product sampled data.
In the above-mentioned technical solutions, it specifies detailed calculation when each sampled data meets normal distribution law, protects
The accuracy for having demonstrate,proved data further ensures the reasonability of threshold value.
Further, the preset bounds parameter includes: minimum threshold, max-thresholds and threshold interval value.
In the above-mentioned technical solutions, the setting of maximum/small threshold value can prevent unnecessary data calculating, reduce and calculate
Amount.
Further, described corresponding according to the product false detection rate group, every group of metalliferous product sampled data
Omission factor group and the requirement of erroneous detection omission factor, selecting corresponding numerical value to be set as threshold value includes: according to every group of metalliferous product sampling
The corresponding omission factor group of data and the product false detection rate group, draw corresponding curve graph;It is required according to erroneous detection omission factor,
Select corresponding numerical value as threshold value from the curve graph.
In the above-mentioned technical solutions, it is presented in a manner of curve graph, allows user is more intuitive, threshold value is well understood to select
The reason of.
Further, the step S1 includes: to obtain one group of metal-free product sampling when reaching certain time interval
Data and at least one set of metalliferous product sampled data;The step S2 includes: according to certain time interval acquisition
The regularity of distribution of metal-free product sampled data and the metalliferous product sampled data is calculated described without gold
The corresponding product false detection rate group of the product sampled data of category and every group of metalliferous corresponding omission factor of product sampled data
Group.
In the above-mentioned technical solutions, sampled data is periodically obtained, threshold value is periodically monitored, metal is visited
The threshold value for surveying device meets actual production demand.
Further, the step S3 further comprises: when the threshold value being arranged meets the requirement of the erroneous detection omission factor, still
Using this threshold value;When the threshold value being arranged does not meet the requirement of the erroneous detection omission factor, then according to the product false detection rate group, every
The requirement of group metalliferous product sampled data corresponding omission factor group and erroneous detection omission factor, reselects corresponding numerical value
It updates or prompt updates the threshold value.
In the above-mentioned technical solutions, when the threshold value being arranged does not meet current situation, it can be adjusted, visit metal
The testing result moment for surveying device meets production requirement.
The present invention also provides a kind of threshold values to select equipment, comprising: module is obtained, for obtaining one group of metal-free product
Sampled data and at least one set of metalliferous product sampled data;Computing module, for being adopted according to metal-free product
Metal-free product sampled data is calculated in the regularity of distribution of sample data and the metalliferous product sampled data
Corresponding product false detection rate group and every group of metalliferous corresponding omission factor group of product sampled data;Selecting module is used for
According to the product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and erroneous detection omission factor
It is required that selecting corresponding numerical value as threshold value.
In the above-mentioned technical solutions, according to point of metal-free product sampled data and metalliferous product sampled data
Cloth rule requires to select suitable threshold value according to erroneous detection omission factor, and the selection of threshold value is made to obtain effective quantized data support, choosing
The threshold value selected more meets actual use demand, reduces false detection rate and omission factor, improves the monitoring of product quality.
Further, the computing module, for according to metal-free product sampled data and described metalliferous
The regularity of distribution of product sampled data, be calculated the corresponding product false detection rate group of metal-free product sampled data and
Every group of metalliferous corresponding omission factor group of product sampled data includes: the computing module, when described metal-free
When product sampled data and the metalliferous product sampled data meet normal distribution law, according to default normal state formula, in advance
If limit parameter, metal-free product sampled data and each metalliferous product sampled data, are calculated institute
It states the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data respectively corresponds to
Omission factor group.
Further, the computing module, according to default normal state formula, preset bounds parameter, metal-free product
Sampled data and each metalliferous product sampled data, it is corresponding to be calculated metal-free product sampled data
Product false detection rate group and every group of metalliferous corresponding omission factor group of product sampled data include: the computing module, root
According to metal-free product sampled data, calculates corresponding product average value and product standard is poor;And according to every group of institute
State metalliferous product sampled data, calculate every group described in the metalliferous corresponding metal average value of product sampled data and
Metal master is poor;And according to the product average value, the product standard is poor, the preset bounds parameter and default normal state
The corresponding product false detection rate group of metal-free product sampled data is calculated in formula;And according to the default boundary
Limit the corresponding metal average value of metalliferous product sampled data described in parameter, default normal state formula and every group and metal mark
It is quasi- poor, calculate separately to obtain the corresponding omission factor group of every group of metalliferous product sampled data.
Further, the selecting module, for according to the product false detection rate group, every group of metalliferous product sampled data
The requirement of corresponding omission factor group and erroneous detection omission factor, selecting corresponding numerical value as threshold value includes: drawing submodule, is used
According to every group of metalliferous corresponding omission factor group of product sampled data and the product false detection rate group, draw corresponding
Curve graph;Submodule is selected to select corresponding numerical value as threshold from the curve graph for the requirement according to erroneous detection omission factor
Value.
Further, further includes: photographing module, for obtaining the working condition of metal detector in real time;The acquisition module,
It is further used for identifying one group of metal-free product sampled data according to the working condition of the metal detector obtained in real time
With at least one set of metalliferous product sampled data.
In the above-mentioned technical solutions, by the working condition of photographing module acquisition metal detector to obtain sampled data,
Application range is more extensive, and compatibility is higher.
Further, the acquisition module, is further used for when reaching certain time interval, and one group of acquisition is metal-free
Product sampled data and at least one set of metalliferous product sampled data.
Further, the selecting module, the threshold value for being further used for be arranged meet the requirement of the erroneous detection omission factor,
Still use this threshold value;And when the threshold value being arranged does not meet the requirement of the erroneous detection omission factor, then according to the product erroneous detection
The requirement of rate group, every group of metalliferous product sampled data corresponding omission factor group and erroneous detection omission factor, reselects phase
The numerical value answered updates or prompt updates the threshold value.
Compared with prior art, threshold selection method of the invention and equipment beneficial effect are:
The present invention according to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data, according to
The requirement of false detection rate, omission factor selects suitable threshold value, and the selection of threshold value is made to obtain effective quantized data support, the threshold of selection
Value more meets actual use demand, reduces false detection rate, control omission factor, improves the monitoring of product quality.
Detailed description of the invention
Below by clearly understandable mode, preferred embodiment is described with reference to the drawings, to a kind of threshold selection method and
Above-mentioned characteristic, technical characteristic, advantage and its implementation of equipment are further described.
Fig. 1 is the flow chart of threshold selection method one embodiment of the present invention;
Fig. 2 is the flow chart of another embodiment of threshold selection method of the present invention;
Fig. 3 is the showing interface figure of sampled data and preset bounds parameter one embodiment of the present invention;
Fig. 4 is that the present invention is drawn according to product false detection rate group and the corresponding omission factor group of three groups of metalliferous product sampled datas
The curve graph of system;
Fig. 5 is the schematic diagram of one embodiment of lowest limit value when selecting threshold value;
The schematic diagram of highest boundary value when Fig. 6 is Fig. 5 selection threshold value;
Fig. 7 is the structural schematic diagram of threshold value selection equipment one embodiment of the present invention;
Fig. 8 is the structural schematic diagram of threshold value selection another embodiment of equipment of the present invention;
Fig. 9 is the flow chart of another embodiment of threshold selection method of the present invention.
Drawing reference numeral explanation:
10. obtaining module, 20. computing modules, 30. selecting modules, 31. drawing submodules, 32. selection submodules, 40. are taken the photograph
As module.
Specific embodiment
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, Detailed description of the invention will be compareed below
A specific embodiment of the invention.It should be evident that drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing, and obtain other embodiments.
To make simplified form, part related to the present invention is only schematically shown in each figure, they are not represented
Its practical structures as product.In addition, there is identical structure or function in some figures so that simplified form is easy to understand
Component only symbolically depicts one of those, or has only marked one of those.Herein, "one" is not only indicated
" only this ", can also indicate the situation of " more than one ".
The working principle of metal detector is electromagnetic induction technology, there is the electromagnetism of a special designing inside metal detector
Coil will cause the variation of electromagnetic field when prill passes through, and receives circuit part and detects this faint signal intensity,
In conjunction with certain technological means, by certain analysis and processing, finally judge to whether there is metallic foreign body in product, and threshold value
Setting be exactly metal detector judgement foundation.False detection rate refers to when threshold value is too low, and normal product is determined as containing metal
Ratio.Omission factor is exactly that the product containing metal is determined as normal product when threshold value is too high, thus the ratio of missing inspection.
Therefore, the reasonable set of threshold value is extremely important.
With sensitivity increasingly higher demands, the threshold value of setting is lower and lower, this threshold value always to product just
Regular signal value region is close, such as 200, and 190,180,130,120,100,90 etc., after threshold value gradually decreases, it will lead to erroneous detection
Increase, excessively high false detection rate can bring biggish economic loss, and only selection threshold value is unable to meet production demand again, improves by rule of thumb
Quality.
Threshold selection method is equivalent to the program of an operation, can be installed on metal detector, can also install
In being each metal detector service in another independent equipment.
Fig. 1 shows one embodiment of the present of invention, a kind of threshold selection method, comprising:
S101 obtains one group of metal-free product sampled data and at least one set of metalliferous product sampled data.
Specifically, metal-free product sampled data refers to that product itself crosses the data that metal detector obtains.
The selection of threshold value and application are closely related, and therefore, the group number of metalliferous product sampled data is according to desired
Depending on the amount of metal of detection.Metalliferous product sampled data, which refers to contain associated metal in product and cross metal detector, to be obtained
The data arrived.
Such as: think whether there can be iron in testing product, then can obtain one group of iron-containing product sampled data, concrete mode
Metal detector acquisition is crossed to contain iron ball in product;Alternatively, thinking whether can there is iron in testing product, the metal of non-ferric, no
Become rusty steel, then needs to obtain three groups of metalliferous product sampled datas, respectively iron, non-ferric metal and stainless steel, implementation
The corresponding metalliferous product sampled data of each group is obtained for three kinds of different metals are put into test in product respectively.
Every group of sampled data number is arranged according to available accuracy, can be any number.Such as: as shown in figure 3, having 4
Group sampled data, wherein one group is metal-free product sampled data, in addition three groups are metalliferous product sampled data, often
Group has 100 data.
When the metalliferous product sampled data of general test, be by prill (such as: the iron shot of 1.5mm diameter,
The stainless shot of the non-ferrous bead of 2.0mm, 2.5mm) it is put into product, when they pass through metal detector, due to each
The influence of kind complicated, inevitable objective factor, such as by when speed, the position with respect to machine, running track, conveying
Band, vibration, product temperature, environment temperature, supply voltage, atmospheric pressure, electrostatic, the electromagnetic environment of workshop complexity etc., lead to gold
When category bead passes through every time, signal is almost impossible every time identical, and therefore, the quantity of sampled data is as more as possible, example
Such as: every group setting 200,150 etc., to improve accuracy.
It should be noted that in view of metal detector because present technical reason when detecting can be because of metallic foreign body
Shape it is different and keep result distinct, such as: diameter 1mm, length are the wire of 10mm, may can be examined in an angle
It measures, but after rotating by a certain angle (such as: 90 degree) it possibly can not detect, therefore, it is suggested that use spherical as standard.
Sampled data is obtained by metal detector, when the corresponding program of threshold selection method is run on metal detector
When, it calls directly;When this program is run in the equipment other than metal detector, can be sent by metal detector
It arrives, can also voluntarily be identified and be obtained by the test process that photographing module monitors metal detector.Sampled data is not limited herein
Specific acquisition modes.
S102 is calculated according to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data
It is respectively corresponded to the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data
Omission factor group.
Specifically, after obtaining metal-free product sampled data and each group metalliferous product sampled data, analysis
Determine the regularity of distribution of sampled data, it is subsequently selected that the suitable formula of reselection, which calculates product false detection rate group and omission factor group,
Suitable threshold value lays the foundation.
Product false detection rate group be by metal-free product sampled data in different threshold values corresponding product false detection rate group
At;One group of omission factor group refers to one group of metalliferous product sampled data corresponding omission factor composition in different threshold values.
S103 is according to product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and reality
Erroneous detection omission factor requirement, select corresponding numerical value as threshold value.
Specifically, erroneous detection omission factor requires to determine according to the actual situation, and such as: false detection rate requires to be lower than 0.1%, leaks simultaneously
Inspection rate will also be lower than 0.1%, then select from product false detection rate group and each group omission factor group while meeting the numerical value lower than 0.1%
As threshold value.
Meet false detection rate most of the time, omission factor is required of a numberical range, can arbitrarily select this this when
A numerical value in numberical range is as threshold value.
Preferably, S103 is according to product false detection rate group, every group of metalliferous corresponding omission factor of product sampled data
The requirement of group and erroneous detection omission factor, selecting corresponding numerical value as threshold value includes:
According to the corresponding omission factor group of every group of metalliferous product sampled data and product false detection rate group, draws and correspond to
Curve graph;It is required according to erroneous detection omission factor, selects corresponding numerical value as threshold value from curve graph.
Specifically, each omission factor group being calculated and product false detection rate group are depicted as corresponding curve graph, use is allowed
Person intuitively understands the corresponding false detection rate of each numerical value, omission factor, easily and quickly understands the selection reason of threshold value.
As shown in Figure 5,6, product false detection rate group has been depicted as the curve on the left side, three groups of metalliferous product sampled datas pair
The omission factor group answered has been depicted as three curves in the right, when false detection rate requires to be 1 ‰ or less, as long as can be seen that really from Fig. 5,6
When protecting the setting of t value between 233-327, by the missing inspection of the product false detection rate and metal that can ensure to be lower than 1 ‰ simultaneously
Rate guarantees good production.When being that the first factor considers with quality, product erroneous detection should not only ensure that in this way as close as possible to 233
Rate is not above 1 ‰, while guaranteeing the maximum Detection capability of metal detector;When in production due to interference, temperature, live electromagnetism
When environment starts slowly to influence equipment that signal is caused to increase, it can suitably increase threshold value while guarantee the safety detection for meeting metal
Rate, as long as i.e. live adjustable extent allows in 233-327, rather than leans on the feeling of people, amount because being no more than 327
Change numberical range and improves product control standard and quality.
According to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data in the present embodiment,
Suitable threshold value is selected according to the requirement of false detection rate, omission factor, the selection of threshold value is made to obtain effective quantized data support, selection
Threshold value more meet actual use demand, reduce false detection rate, omission factor, improve the more acurrate monitoring to product quality.
Fig. 2 shows another embodiment of the invention, a kind of threshold selection method, comprising:
S201 obtains one group of metal-free product sampled data and at least one set of metalliferous product sampled data.
S202 is calculated according to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data
It is respectively corresponded to the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data
Omission factor group include:
S212 when metal-free product sampled data and metalliferous product sampled data meet normal distribution law,
According to default normal state formula, preset bounds parameter, (preset bounds parameter includes: minimum threshold, max-thresholds and threshold interval
Value), metal-free product sampled data and each metalliferous product sampled data, metal-free product sampling is calculated
The corresponding product false detection rate group of data and every group of metalliferous corresponding omission factor group of product sampled data.
Specifically, preset bounds parameter refers to the threshold range of metal detector, it is suitable to be arranged according to actually detected demand
Minimum threshold, max-thresholds and threshold interval value.
Such as: the metal if desired detected is stainless steel and iron, as shown in figure 3, max-thresholds can be set as to 1500, most
Small threshold value is 5, and threshold interval value is 1.Certainly, these data can flexibly change, such as: max-thresholds are set as 1400, minimum
Threshold value is 10, and threshold interval value is 3 etc..If max-thresholds just lose meaning beyond the corresponding value of maximum metal block, only do
Therefore corresponding preset bounds parameter can be arranged according to the detection demand of actual metal in useless extra calculating.
Normal distribution law has its corresponding statistical formula, and can be set to default normal state formula makes for subsequent calculating
With.
Preferably, according to default normal state formula, preset bounds parameter, (preset bounds parameter includes: minimum threshold, most to S212
Big threshold value and threshold interval value), metal-free product sampled data and each metalliferous product sampled data, be calculated not
The metalliferous corresponding product false detection rate group of product sampled data and every group of metalliferous corresponding leakage of product sampled data
Inspection rate group includes:
S2121 calculates corresponding product average value and product standard is poor according to metal-free product sampled data;
S2122 calculates every group of corresponding metal of metal sampled data according to every group of metalliferous product sampled data
Average value and metal master are poor.
Specifically, average value is exactly the summation of sampled data divided by total number.
Such as: there are 200 data in one group of metal-free product sampled data, then after being added this 200 data, removes
With 200, product average value is obtained.It is similarly the corresponding metal average value of every group of metalliferous product sampled data.
Standard deviation is then calculated according to the following formula:
σ is standard deviation, and μ is average value, μiFor i-th of data in a set product sampled data, N is a set product hits
According to total number.
S2123 is according to product average value, product standard is poor, preset bounds parameter and default normal state formula, is calculated not
The corresponding product false detection rate group of metalliferous product sampled data;
S2124 is respectively corresponded to according to preset bounds parameter, default normal state formula and every group of metalliferous product sampled data
Metal average value and metal master it is poor, calculate separately to obtain the corresponding omission factor group of every group of metalliferous product sampled data.
Specifically, default normal state formula are as follows:
σ is standard deviation, and μ is average value, and x is then threshold value, minimum threshold≤x≤max-thresholds, with the progress of threshold interval value
Incremental calculation, F are the probability being calculated.
Different x values is substituted into above-mentioned formula, the probability summation less than or equal to x value can be calculated, for product,
It is normal condition lower than x value, belongs to erroneous detection situation higher than x value;For metal, belong to detection higher than x value, is lower than x value category
In missing inspection.
Assuming that having 3 groups of metalliferous product sampled datas (a, b, c group) and one group of metal-free product sampled data (d
Group), by metal-free product sampled data, product standard is poor, product average value and preset bounds parameter substitute into one by one it is above-mentioned
The data group for the product probability summation that formula calculates is Fd [x], then product false detection rate group is 1-Fd [x].Such as: it calculates
Come as x=500, Fd [x]=0.9 represent product normal through probability, and its corresponding product false detection rate is then 1-0.9
=0.1;As x=1000, Fd [x]=0.98, corresponding product false detection rate is then 1-0.98=0.02 ...
By the metalliferous product sampled data of a group, metal master is poor, metal average value and preset bounds parameter substitute into one by one
The data group for the probability summation that above-mentioned formula is calculated is Fa [x], the i.e. corresponding missing inspection of the metalliferous product sampled data of a group
Rate group similarly obtains b group and the corresponding omission factor group Fb [x] of the metalliferous product sampled data of c group and Fc [x], according to 1-Fd
The curve graph that [x], Fa [x], Fb [x] and Fc [x] are drawn is as shown in Figure 4.
Although being not intended to limit to be it should be noted that the present embodiment has said the sequencing of S2121-S2124
This sequence, as long as product false detection rate group and the corresponding omission factor group of every group of metalliferous product sampled data can be calculated
?.
S203 is according to product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and reality
Erroneous detection omission factor requirement, select corresponding numerical value as threshold value.
Preferably, S203 is according to product false detection rate group, every group of metalliferous corresponding omission factor of product sampled data
Group and the requirement of erroneous detection omission factor, selecting corresponding numerical value as threshold value includes:
S213 is drawn according to the corresponding omission factor group of every group of metalliferous product sampled data and product false detection rate group
Corresponding curve graph;S223 selects corresponding numerical value as threshold value according to the requirement of erroneous detection omission factor from curve graph.
Specifically explain same as the previously described embodiments, details are not described herein.
Specific probability calculation mode is given for the sampled data for meeting normal distribution law in the present embodiment, is obtained
The data supporting of quantization, makes the numerical value as threshold value selected more meet actual monitoring demand.
In yet another embodiment of the present invention, as shown in figure 9, a kind of threshold selection method, comprising:
S301 obtains one group of (preferably, apart from current time nearest) and is free of metal when reaching certain time interval
Product sampled data and at least one set of metalliferous product sampled data.
The metal-free product sampled data and metalliferous product hits that S302 is obtained according to certain time interval
According to the regularity of distribution, the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous production is calculated
The corresponding omission factor group of product sampled data.
S303 is according to product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and erroneous detection
The requirement of omission factor selects corresponding numerical value as threshold value.
Specifically, the threshold value selection of the present embodiment can be carried out periodically in actual use, when certain time interval are set
Set it is sufficiently small when, it can be achieved that real-time monitoring.
When corresponding program is run on metal detector, metal detector is according to each sampled data obtained in real time, selection
One group of metal-free product sampled data and at least one set of metalliferous product sampled data, carry out in real time/periodically count
It calculates, selects reasonable data as threshold value.
When monitoring the threshold value of metal detector using other equipment, can real-time monitoring metal detector, obtain in real time
Each sampled data is taken, one group of metal-free product sampled data and at least one set of metalliferous product sampled data are selected, into
Row in real time/periodically calculate, select reasonable data as threshold value for metal detector.
Preferably, when the quantity of each sampled data updates, take a certain number of data nearest apart from current time into
Row analysis, makes calculated result more meet present case.
S303 is according to product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and erroneous detection
The requirement of omission factor, selecting corresponding numerical value as threshold value includes:
The threshold value that S313 ought be arranged meets the requirement of erroneous detection omission factor, still uses this threshold value;
The threshold value that S323 ought be arranged does not meet the requirement of erroneous detection omission factor, then according to product false detection rate group, every group containing gold
The requirement of the product sampled data of category corresponding omission factor group and erroneous detection omission factor, reselect corresponding numerical value update or
Prompt updates threshold value.
Specifically, illustrating that metal detector has been provided with a threshold value and (may be if real-time/periodical monitoring
One value is rule of thumb set, it is also possible to for according to a value of the result selection calculated in real time), put into actual product
In monitoring, during actual use influence metal detector since interference, temperature, live electromagnetic environment may will be slow and lead to letter
Number increase when, the threshold value of current setting may be unable to satisfy the omission factor of metal and the low false detection rate of product.
Therefore, product false detection rate group, the every group of metalliferous product sampled data that can be come out according to latest computed are respectively right
The requirement of the omission factor group and erroneous detection omission factor answered confirms whether the threshold value being arranged needs to adjust, if it meets erroneous detection missing inspection
The requirement of rate does not need then to adjust, if it does not meet, it is (specific real as new threshold value to need to reselect a data
Existing process can automatically select a new threshold value, alternatively, issuing prompt, allow staff to know and need to adjust), meet production and wants
It asks.
Optionally, according to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data, meter
Calculation obtains the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data respectively
Corresponding omission factor group includes:
When metal-free product sampled data and metalliferous product sampled data meet normal distribution law, according to
Default normal state formula, preset bounds parameter, metal-free product sampled data and each metalliferous product sampled data, calculate
It obtains the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data is respectively right
The omission factor group answered.
Optionally, according to default normal state formula, preset bounds parameter, metal-free product sampled data and respectively containing metal
Product sampled data, the corresponding product false detection rate group of metal-free product sampled data is calculated and every group is metalliferous
The corresponding omission factor group of product sampled data includes:
According to metal-free product sampled data, calculates corresponding product average value and product standard is poor;
According to every group of metalliferous product sampled data, every group of metalliferous corresponding gold of product sampled data is calculated
Belong to average value and metal master is poor;
According to product average value, product standard is poor, (preset bounds parameter includes: minimum threshold, maximum to preset bounds parameter
Threshold value and threshold interval value) and default normal state formula, the corresponding product erroneous detection of metal-free product sampled data is calculated
Rate group;
According to preset bounds parameter, default normal state formula and every group of metalliferous corresponding metal of product sampled data
Average value and metal master are poor, calculate separately to obtain the corresponding omission factor group of every group of metalliferous product sampled data.
Optionally, according to product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and
The requirement of erroneous detection omission factor, selects corresponding numerical value to be set as threshold value to include:
According to the corresponding omission factor group of every group of metalliferous product sampled data and product false detection rate group, draws and correspond to
Curve graph;
According to the requirement of erroneous detection omission factor, select corresponding numerical value as threshold value from curve graph.
Details are not described herein for this implementation explanation part identical with above-mentioned each method embodiment, and it is real to refer to above-mentioned each method
Apply example.
In the present embodiment, periodically/monitoring in real time is carried out to the selection of threshold value, the moment guarantees metal detector setting
Threshold value can satisfy the demand of actual production.
Fig. 7 shows the embodiment of threshold value selection equipment of the invention, comprising:
Module 10 is obtained, is adopted for obtaining one group of metal-free product sampled data and at least one set of metalliferous product
Sample data.
Specifically, metal-free product sampled data refers to that product itself crosses the data that metal detector obtains.
The selection of threshold value and application are closely related, and therefore, the group number of metalliferous product sampled data is according to desired
Depending on the amount of metal of detection.Metalliferous product sampled data, which refers to contain associated metal in product and cross metal detector, to be obtained
The data arrived.
Such as: think whether there can be iron in testing product, then can obtain one group of iron-containing product sampled data, concrete mode
Metal detector acquisition is crossed to contain iron ball in product;Alternatively, thinking whether can there is iron in testing product, the metal of non-ferric, no
Become rusty steel, then needs to obtain three groups of metalliferous product sampled datas, respectively iron, non-ferric metal and stainless steel, implementation
The corresponding metalliferous product sampled data of each group is obtained for three kinds of different metals are put into test in product respectively.
Every group of sampled data number is arranged according to available accuracy, can be any number.Such as: as shown in figure 3, having 4
Group sampled data, wherein one group is metal-free product sampled data, in addition three groups are metalliferous product sampled data, often
Group has 100 data.
When the metalliferous product sampled data of general test, be by prill (such as: the iron shot of 1.5mm diameter,
The stainless shot of the non-ferrous bead of 2.0mm, 2.5mm.) it is put into product, allow them to obtain by metal detector.Due to each
The influence of kind complicated, inevitable objective factor, such as by when speed, the position with respect to machine, running track, conveying
Band, vibration, product temperature, environment temperature, supply voltage, atmospheric pressure, electrostatic, the electromagnetic environment of workshop complexity etc., lead to gold
When category bead passes through every time, signal is almost impossible every time identical, and therefore, the quantity of sampled data is as more as possible, example
Such as: every group setting 200,150 etc., to improve accuracy.
It should be noted that in view of metal detector because present technical reason when detecting can be because of metallic foreign body
Shape it is different and keep result distinct, such as: diameter 1mm, length are the wire of 10mm, may can be examined in an angle
It measures, but after rotating by a certain angle (such as: 90 degree) it possibly can not detect, therefore, it is suggested that use spherical as standard.
Sampled data monitors to obtain by metal detector, when the corresponding program of threshold selection method runs on metal detection
When on device, call directly;When this program is run in the equipment other than metal detector, can be sent by metal detector
It obtains, can also voluntarily be identified and be obtained by the test process that photographing module monitors metal detector.Hits is not limited herein
According to specific acquisition modes.
When the corresponding program of threshold selection method is not run on metal detector, optionally, threshold value selects equipment also
It include: photographing module, for obtaining the working condition of metal detector in real time;Module 10 is obtained, is further used for according in real time
The working condition of the metal detector of acquisition identifies that one group of metal-free product sampled data and at least one set are metalliferous
Product sampled data.
Specifically, photographing module is made of several photographic devices, each photographic device is for monitoring one or more gold
Belong to detector, shoot metal detector working condition (such as: test product, wrap metalliferous product when test knot
Fruit), identified and interpreted by acquisition module 10 (such as: the conversion of picture to numerical value is carried out according to the picture of camera shooting), it obtains
One group of metal-free product sampled data and at least one set of metalliferous product sampled data.
And the function of obtaining module 10, computing module 20 and selecting module 30 can be completed by a computer/server.It takes the photograph
As device and computer/server communication connection (such as: each camera by high-speed data communication line through the network equipment will clap
The video signal transmission taken the photograph is to computer/server, and high-speed data communication line can be Ethernet, intranet, can also be with
It is Internet, can also be the various high-speed communication lines including USB), guarantee data transmission between the two.When
So, photographic device also can work independently, it is found that when the false detection rate of metal detector, omission factor need to adjust, automatic issue is reminded.
Computing module 20, for the distribution according to metal-free product sampled data and metalliferous product sampled data
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product hits is calculated in rule
According to corresponding omission factor group.
Specifically, after obtaining metal-free product sampled data and each group metalliferous product sampled data, analysis
Determine the regularity of distribution of sampled data, it is subsequently selected that the suitable formula of reselection, which calculates product false detection rate group and omission factor group,
Suitable threshold value lays the foundation.
Selecting module 30, for according to product false detection rate group, every group of metalliferous corresponding leakage of product sampled data
The requirement of inspection rate group and actual erroneous detection omission factor, selects corresponding numerical value as threshold value.
Specifically, erroneous detection omission factor requires to determine according to the actual situation, and such as: false detection rate requires to be lower than 0.1%, leaks simultaneously
Inspection rate will also be lower than 0.1%, then select from product false detection rate group and each group omission factor group while meeting the numerical value lower than 0.1%
As threshold value.
Meet false detection rate most of the time, omission factor is required of a numberical range, then can arbitrarily select this numerical value model
A numerical value in enclosing is as threshold value.
Preferably, selecting module 30, for respectively right according to product false detection rate group, every group of metalliferous product sampled data
The requirement of the omission factor group and actual erroneous detection omission factor answered, selecting corresponding numerical value as threshold value includes:
Drawing submodule 31, for according to the corresponding omission factor group of every group of metalliferous product sampled data and product
False detection rate group draws corresponding curve graph;
Submodule 32 is selected, for being required according to erroneous detection omission factor, selects corresponding numerical value as threshold value from curve graph.
Specifically, each omission factor group being calculated and product false detection rate group are depicted as corresponding curve graph, use is allowed
Person intuitively understands the corresponding false detection rate of each numerical value, omission factor, easily and quickly understands the selection reason of threshold value.Specific example
Son can be found in corresponding embodiment of the method, and details are not described herein.
According to the regularity of distribution of metal-free product sampled data and metalliferous product sampled data in the present embodiment,
Suitable threshold value is selected according to the requirement of false detection rate, omission factor, the selection of threshold value is made to obtain effective quantized data support, selection
Threshold value more meet actual use demand, reduce false detection rate and omission factor, improve the monitoring of product quality.
Fig. 8 shows the embodiment of another threshold value selection equipment, comprising:
Module 10 is obtained, is adopted for obtaining one group of metal-free product sampled data and at least one set of metalliferous product
Sample data.
Optionally, when the corresponding program of threshold selection method is not run on metal detector, optionally, threshold value selection
Equipment further include: photographing module 40, for obtaining the working condition of metal detector in real time;Module 10 is obtained, is further used for
According to the working condition of the metal detector obtained in real time, one group of metal-free product sampled data and at least one set are identified
Metalliferous product sampled data.
Specific explanation is identical as above equipment embodiment, and details are not described herein.
Computing module 20, for the distribution according to metal-free product sampled data and metalliferous product sampled data
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product hits is calculated in rule
Include: according to corresponding omission factor group
Computing module 20, when metal-free product sampled data and metalliferous product sampled data meet normal distribution
When regular, according to default normal state formula, preset bounds parameter, (preset bounds parameter includes: minimum threshold, max-thresholds and threshold value
Spacing value), metal-free product sampled data and each metalliferous product sampled data, metal-free product is calculated
The corresponding product false detection rate group of sampled data and every group of metalliferous corresponding omission factor group of product sampled data.
Specifically, preset bounds parameter refers to the threshold range of metal detector, it is suitable to be arranged according to actually detected demand
Minimum threshold, max-thresholds and threshold interval value.
Such as: the metal if desired detected is stainless steel and iron, as shown in figure 3, max-thresholds can be set as to 1500, most
Small threshold value is 5, and threshold interval value is 1.Certainly, these data can flexibly change, such as: max-thresholds are set as 1400, minimum
Threshold value is 10, and threshold interval value is 3 etc..
Normal distribution law has its corresponding statistical formula, and can be set to default normal state formula makes for subsequent calculating
With.
Preferably, computing module 20, according to default normal state formula, preset bounds parameter, metal-free product hits
According to each metalliferous product sampled data, be calculated the corresponding product false detection rate group of metal-free product sampled data and
Every group of metalliferous corresponding omission factor group of product sampled data include:
Computing module 20 calculates corresponding product average value and product standard according to metal-free product sampled data
Difference;And according to every group of metalliferous product sampled data, calculate every group of metalliferous corresponding gold of product sampled data
Belong to average value and metal master is poor;And according to product average value, product standard is poor, preset bounds parameter and default normal state are public
The corresponding product false detection rate group of product sampled data is calculated in formula;And according to preset bounds parameter, default normal state formula
It is poor with the corresponding metal average value of every group of metalliferous product sampled data and metal master, it calculates separately to obtain every group and contains
The corresponding omission factor group of the product sampled data of metal.
Specifically, average value is exactly the summation of sampled data divided by total number.Such as: have in a set product sampled data
150 data, divided by 150, obtain product average value after being then added this 150 data.It is similarly every group of metal sampled data
Corresponding metal average value.
Standard deviation is then calculated according to the formula in above-mentioned corresponding method embodiment, refers to corresponding method and implements
Example, details are not described herein.
Default normal state formula refers to corresponding embodiment of the method, and different x values is substituted into default normal state formula, can be calculated
It obtains being less than or equal to the probability summation of x value being normal condition lower than x value for product, belongs to erroneous detection situation higher than x value;
For metal, belong to detection higher than x value, belongs to missing inspection lower than x value.Specific example can be found in corresponding method and implement
Example, details are not described herein.
Selecting module 30, for according to product false detection rate group, every group of metalliferous corresponding leakage of product sampled data
Inspection rate group and the requirement of erroneous detection omission factor, select corresponding numerical value as threshold value.
Preferably, selecting module 30, for respectively right according to product false detection rate group, every group of metalliferous product sampled data
Omission factor group and erroneous detection the omission factor requirement answered, selecting corresponding numerical value as threshold value includes:
Drawing submodule 31, for according to the corresponding omission factor group of every group of metalliferous product sampled data and product
False detection rate group draws corresponding curve graph;
Submodule 32 is selected, for being required according to erroneous detection omission factor, selects corresponding numerical value as threshold value from curve graph.
Specific implementation process is identical as the specific implementation process in above equipment embodiment, is not described in detail herein.
Specific probability calculation mode is given for the sampled data for meeting normal distribution law in the present embodiment, is obtained
The data supporting of quantization, makes the numerical value as threshold value selected more meet actual monitoring demand.
When the corresponding program of the selection method of threshold value is executed by nonmetallic detector equipment, the equipment such as camera can be passed through
For monitoring metal detector to obtain sampled data, implementation diversification meets the application demand of different occasions.
In another apparatus embodiments of the invention, a kind of threshold value selection equipment includes:
Module 10 is obtained, for when reaching certain time interval, obtains (preferably, apart from current time nearest) one
The metal-free product sampled data of group and at least one set of metalliferous product sampled data;
Computing module 20, for the distribution according to metal-free product sampled data and metalliferous product sampled data
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product hits is calculated in rule
According to corresponding omission factor group;
Selecting module 30, for according to product false detection rate group, every group of metalliferous corresponding leakage of product sampled data
The requirement of inspection rate group and erroneous detection omission factor selects corresponding numerical value as threshold value.
Specifically, the threshold value selection of the present embodiment can be carried out periodically in actual use, when certain time interval are set
Set it is sufficiently small when, it can be achieved that real-time monitoring.
When corresponding program is run on metal detector, metal detector is according to each sampled data obtained in real time, selection
One group of metal-free product sampled data and at least one set of metalliferous product sampled data, carry out in real time/periodically count
It calculates, selects reasonable data as threshold value.
When monitoring the threshold value of metal detector using other equipment, can real-time monitoring metal detector, obtain in real time
Each sampled data is taken, one group of metal-free product sampled data and at least one set of metalliferous product sampled data are selected, into
Row in real time/periodically calculate, provide suggestion of the reasonable data as threshold value for metal detector.
Preferably, when the quantity of each sampled data updates, take a certain number of data nearest apart from current time into
Row analysis, makes calculated result more meet present case.
Selecting module 30, for according to product false detection rate group, every group of metalliferous corresponding leakage of product sampled data
The requirement of inspection rate group and erroneous detection omission factor, selecting corresponding numerical value as threshold value includes:
Selecting module 30, the threshold value for that ought be arranged meet the requirement of erroneous detection omission factor, still use this threshold value;And
When the threshold value being arranged does not meet the requirement of erroneous detection omission factor, then according to product false detection rate group, every group of metalliferous product sampling
The requirement of data corresponding omission factor group and erroneous detection omission factor reselects corresponding numerical value and updates or prompt to update threshold
Value.
Specifically, illustrating that metal detector has been provided with a threshold value and (may be if real-time/periodical monitoring
One value is rule of thumb set, it is also possible to for according to a value of the result selection calculated in real time), put into actual product
In monitoring, during actual use influence metal detector since interference, temperature, live electromagnetic environment may will be slow and lead to letter
Number increase when, the threshold value of current setting may be unable to satisfy the omission factor of metal and the low false detection rate of product.
Therefore, product false detection rate group, the every group of metalliferous product sampled data that can be come out according to latest computed are respectively right
The requirement of the omission factor group and erroneous detection omission factor answered confirms whether the threshold value being arranged needs to adjust, if it meets erroneous detection missing inspection
The requirement of rate does not need then to adjust, if it does not meet, it is (specific real as new threshold value to need to reselect a data
Existing process can automatically select a new threshold value, alternatively, issuing prompt, allow staff to know and need to adjust), meet production and wants
It asks.
Optionally, computing module 20, for according to metal-free product sampled data and metalliferous product hits
According to the regularity of distribution, the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous production is calculated
The corresponding omission factor group of product sampled data includes:
Computing module 20, when metal-free product sampled data and metalliferous product sampled data meet normal distribution
When regular, adopted according to default normal state formula, preset bounds parameter, metal-free product sampled data and each metalliferous product
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampling is calculated in sample data
The corresponding omission factor group of data.
Optionally, computing module 20, according to default normal state formula, preset bounds parameter, metal-free product hits
According to each metalliferous product sampled data, be calculated the corresponding product false detection rate group of metal-free product sampled data and
Every group of metalliferous corresponding omission factor group of product sampled data include:
Computing module 20 calculates corresponding product average value and product standard according to metal-free product sampled data
Difference;And according to every group of metalliferous product sampled data, calculate every group of metalliferous corresponding gold of product sampled data
Belong to average value and metal master is poor;And according to product average value, product standard is poor, preset bounds parameter (preset bounds parameter
Include: minimum threshold, max-thresholds and threshold interval value) and default normal state formula, metal-free product sampling is calculated
The corresponding product false detection rate group of data;And it is adopted according to preset bounds parameter, default normal state formula and every group of metalliferous product
The corresponding metal average value of sample data and metal master are poor, calculate separately to obtain every group of metalliferous product sampled data pair
The omission factor group answered.
Optionally, selecting module 30, for respectively right according to product false detection rate group, every group of metalliferous product sampled data
The requirement of the omission factor group and erroneous detection omission factor answered, selecting corresponding numerical value as threshold value includes:
Drawing submodule 31, for according to the corresponding omission factor group of every group of metalliferous product sampled data and product
False detection rate group draws corresponding curve graph;
Submodule 32 is selected, for being required according to erroneous detection omission factor, selects corresponding numerical value as threshold value from curve graph.
Optionally, when threshold value selects equipment as the equipment of nonmetallic detector, threshold value selection equipment may also include that camera shooting
Module 40, for obtaining the working condition of metal detector in real time;Module 10 is obtained, is further used for according to the gold obtained in real time
The working condition for belonging to detector identifies one group of metal-free product sampled data and at least one set of metalliferous product sampling
Data.
Details are not described herein for this implementation explanation part identical with above-mentioned each apparatus embodiments, and it is real to refer to above-mentioned each method
Apply example.
In the present embodiment, periodically/monitoring in real time is carried out to the selection of threshold value, the moment guarantees metal detector setting
Threshold value can satisfy the demand of actual production.
It should be noted that above-described embodiment can be freely combined as needed.The above is only of the invention preferred
Embodiment, it is noted that for those skilled in the art, in the premise for not departing from the principle of the invention
Under, several improvements and modifications can also be made, these modifications and embellishments should also be considered as the scope of protection of the present invention.
Claims (14)
1. a kind of threshold selection method, which comprises the following steps:
S1 obtains one group of metal-free product sampled data and at least one set of metalliferous product sampled data;
S2 is calculated according to the regularity of distribution of metal-free product sampled data and the metalliferous product sampled data
It obtains the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product sampled data is each
Self-corresponding omission factor group;
S3 is leaked according to the product false detection rate group, the corresponding omission factor group of every group of metalliferous product sampled data and erroneous detection
The requirement of inspection rate selects corresponding numerical value as threshold value.
2. threshold selection method as described in claim 1, which is characterized in that described to be adopted according to metal-free product
Metal-free product sampled data is calculated in the regularity of distribution of sample data and the metalliferous product sampled data
Corresponding product false detection rate group and every group of metalliferous corresponding omission factor group of product sampled data include:
When metal-free product sampled data and the metalliferous product sampled data meet normal distribution law,
According to default normal state formula, preset bounds parameter, metal-free product sampled data and each metalliferous product
The corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous production is calculated in sampled data
The corresponding omission factor group of product sampled data.
3. threshold selection method as claimed in claim 2, which is characterized in that the basis presets normal state formula, default boundary
Limit parameter, metal-free product sampled data and each metalliferous product sampled data, be calculated it is described not
The metalliferous corresponding product false detection rate group of product sampled data and every group of metalliferous corresponding leakage of product sampled data
Inspection rate group includes:
According to metal-free product sampled data, calculates corresponding product average value and product standard is poor;
The metalliferous product sampled data according to every group, calculate every group described in metalliferous product sampled data respectively correspond to
Metal average value and metal master it is poor;
According to the product average value, the product standard is poor, the preset bounds parameter and default normal state formula, is calculated
The corresponding product false detection rate group of metal-free product sampled data;
It is corresponding according to metalliferous product sampled data described in the preset bounds parameter, default normal state formula and every group
Metal average value and metal master are poor, calculate separately to obtain the corresponding omission factor group of every group of metalliferous product sampled data.
4. threshold selection method as claimed in claim 2, which is characterized in that the preset bounds parameter include: minimum threshold,
Max-thresholds and threshold interval value.
5. threshold selection method as described in claim 1, which is characterized in that it is described according to the product false detection rate group, it is every
The requirement of group metalliferous product sampled data corresponding omission factor group and erroneous detection omission factor, selects corresponding numerical value to be set as
Threshold value includes:
According to every group of metalliferous corresponding omission factor group of product sampled data and the product false detection rate group, draws and correspond to
Curve graph;
According to the requirement of erroneous detection omission factor, select corresponding numerical value as threshold value from the curve graph.
6. threshold selection method a method as claimed in any one of claims 1 to 5, it is characterised in that:
The step S1 includes: to obtain one group of metal-free product sampled data and at least when reaching certain time interval
One group of metalliferous product sampled data;
The step S2 include: according to certain time interval obtain metal-free product sampled data and it is described containing gold
The corresponding product false detection rate of metal-free product sampled data is calculated in the regularity of distribution of the product sampled data of category
Group and every group of metalliferous corresponding omission factor group of product sampled data.
7. threshold selection method as claimed in claim 6, which is characterized in that the step S3 further comprises:
When the threshold value being arranged meets the requirement of the erroneous detection omission factor, still using this threshold value;
When the threshold value being arranged does not meet the requirement of the erroneous detection omission factor, then according to the product false detection rate group, every group containing gold
The requirement of the product sampled data of category corresponding omission factor group and erroneous detection omission factor, reselect corresponding numerical value update or
Prompt updates the threshold value.
8. a kind of threshold value selects equipment characterized by comprising
Module is obtained, for obtaining one group of metal-free product sampled data and at least one set of metalliferous product hits
According to;
Computing module, for point according to metal-free product sampled data and the metalliferous product sampled data
Cloth rule, is calculated the corresponding product false detection rate group of metal-free product sampled data and every group of metalliferous product
The corresponding omission factor group of sampled data;
Selecting module, for according to the product false detection rate group, every group of metalliferous corresponding missing inspection of product sampled data
The requirement of rate group and erroneous detection omission factor selects corresponding numerical value as threshold value.
9. threshold value as claimed in claim 8 selects equipment, which is characterized in that the computing module, for being free of according to
The regularity of distribution of the product sampled data of metal and the metalliferous product sampled data, is calculated described metal-free
The corresponding product false detection rate group of product sampled data and every group of metalliferous corresponding omission factor group packet of product sampled data
It includes:
The computing module, when metal-free product sampled data and the metalliferous product sampled data meet just
When the state regularity of distribution, according to default normal state formula, preset bounds parameter, metal-free product sampled data and each described
The corresponding product false detection rate group of metal-free product sampled data and every is calculated in metalliferous product sampled data
The metalliferous corresponding omission factor group of product sampled data of group.
10. threshold value as claimed in claim 9 selects equipment, which is characterized in that the computing module, it is public according to default normal state
Formula, preset bounds parameter, metal-free product sampled data and each metalliferous product sampled data, calculate
Respectively to metal-free corresponding product false detection rate group of product sampled data and every group of metalliferous product sampled data
Corresponding omission factor group includes:
The computing module calculates corresponding product average value and product mark according to metal-free product sampled data
It is quasi- poor;And the metalliferous product sampled data according to every group, calculate every group described in metalliferous product sampled data it is each
Self-corresponding metal average value and metal master are poor;And it is poor, described pre- according to the product average value, the product standard
If limit parameter and default normal state formula, the corresponding product false detection rate of metal-free product sampled data is calculated
Group;And the metalliferous product sampled data according to the preset bounds parameter, default normal state formula and every group is respectively right
The metal average value and metal master answered are poor, calculate separately to obtain the corresponding omission factor of every group of metalliferous product sampled data
Group.
11. threshold value as claimed in claim 8 selects equipment, which is characterized in that the selecting module, for according to the product
The requirement of false detection rate group, every group of metalliferous product sampled data corresponding omission factor group and erroneous detection omission factor selects phase
The numerical value answered includes: as threshold value
Drawing submodule, for being missed according to every group of metalliferous corresponding omission factor group of product sampled data and the product
Inspection rate group, draws corresponding curve graph;
Submodule is selected, for being required according to erroneous detection omission factor, selects corresponding numerical value as threshold value from the curve graph.
12. threshold value as claimed in claim 8 selects equipment, which is characterized in that further include:
Photographing module, for obtaining the working condition of metal detector in real time;
The acquisition module is further used for identifying that one group is free of according to the working condition of the metal detector obtained in real time
The product sampled data of metal and at least one set of metalliferous product sampled data.
13. the threshold value as described in claim 8-12 is any selects equipment, which is characterized in that the acquisition module is further used
It is adopted in when reaching certain time interval, obtaining one group of metal-free product sampled data and at least one set of metalliferous product
Sample data.
14. threshold value as claimed in claim 13 selects equipment, which is characterized in that the selecting module, being further used for ought be
The threshold value of setting meets the requirement of the erroneous detection omission factor, still uses this threshold value;And described in not met when the threshold value being arranged
The requirement of erroneous detection omission factor, then according to the product false detection rate group, every group of metalliferous corresponding leakage of product sampled data
The requirement of inspection rate group and erroneous detection omission factor reselects corresponding numerical value and updates or prompt to update the threshold value.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811610109.1A CN109613108B (en) | 2018-12-27 | 2018-12-27 | Threshold selection method and equipment |
PCT/CN2019/091610 WO2020133954A1 (en) | 2018-12-27 | 2019-06-17 | Digital performance analysis method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811610109.1A CN109613108B (en) | 2018-12-27 | 2018-12-27 | Threshold selection method and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109613108A true CN109613108A (en) | 2019-04-12 |
CN109613108B CN109613108B (en) | 2022-09-20 |
Family
ID=66012734
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811610109.1A Active CN109613108B (en) | 2018-12-27 | 2018-12-27 | Threshold selection method and equipment |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN109613108B (en) |
WO (1) | WO2020133954A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110308491A (en) * | 2019-05-29 | 2019-10-08 | 浙江大华技术股份有限公司 | Article detection method, device, storage medium and electronic device |
WO2020133954A1 (en) * | 2018-12-27 | 2020-07-02 | 帝沃检测技术(上海)有限公司 | Digital performance analysis method and device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1488933A (en) * | 2002-08-22 | 2004-04-14 | 丰田自动车株式会社 | Good-or-not decision device, decision programme, method and multi-variable statistical analyzing device |
JP2007333562A (en) * | 2006-06-15 | 2007-12-27 | Anritsu Sanki System Co Ltd | X-ray foreign matter detector |
CN101840003A (en) * | 2010-03-05 | 2010-09-22 | 清华大学 | Green channel open vehicle radar detection method for metal contraband articles |
CN104700345A (en) * | 2015-01-16 | 2015-06-10 | 天津科技大学 | Method for improving detection rate of semi-fragile watermark authentication by establishing Benford's law threshold value library |
US20150234075A1 (en) * | 2012-02-10 | 2015-08-20 | Illinois Tool Works Inc. | Metal detector |
CN107361763A (en) * | 2017-08-09 | 2017-11-21 | 广东虹勤通讯技术有限公司 | A kind of ECG data R wave detecting methods and device |
CN108364017A (en) * | 2018-01-24 | 2018-08-03 | 华讯方舟科技有限公司 | A kind of picture quality sorting technique, system and terminal device |
CN108918427A (en) * | 2018-06-06 | 2018-11-30 | 北京云端光科技术有限公司 | Method, apparatus, storage medium and the electronic equipment of substance detection |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103576206B (en) * | 2012-07-31 | 2017-10-31 | 上海太易检测技术有限公司 | A kind of metal detection machine |
JP6697302B2 (en) * | 2016-03-25 | 2020-05-20 | マークテック株式会社 | Flaw detection device and defect detection method using flaw detection device |
CN109613108B (en) * | 2018-12-27 | 2022-09-20 | 帝沃检测技术(上海)有限公司 | Threshold selection method and equipment |
-
2018
- 2018-12-27 CN CN201811610109.1A patent/CN109613108B/en active Active
-
2019
- 2019-06-17 WO PCT/CN2019/091610 patent/WO2020133954A1/en active Application Filing
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1488933A (en) * | 2002-08-22 | 2004-04-14 | 丰田自动车株式会社 | Good-or-not decision device, decision programme, method and multi-variable statistical analyzing device |
JP2007333562A (en) * | 2006-06-15 | 2007-12-27 | Anritsu Sanki System Co Ltd | X-ray foreign matter detector |
CN101840003A (en) * | 2010-03-05 | 2010-09-22 | 清华大学 | Green channel open vehicle radar detection method for metal contraband articles |
US20150234075A1 (en) * | 2012-02-10 | 2015-08-20 | Illinois Tool Works Inc. | Metal detector |
CN104700345A (en) * | 2015-01-16 | 2015-06-10 | 天津科技大学 | Method for improving detection rate of semi-fragile watermark authentication by establishing Benford's law threshold value library |
CN107361763A (en) * | 2017-08-09 | 2017-11-21 | 广东虹勤通讯技术有限公司 | A kind of ECG data R wave detecting methods and device |
CN108364017A (en) * | 2018-01-24 | 2018-08-03 | 华讯方舟科技有限公司 | A kind of picture quality sorting technique, system and terminal device |
CN108918427A (en) * | 2018-06-06 | 2018-11-30 | 北京云端光科技术有限公司 | Method, apparatus, storage medium and the electronic equipment of substance detection |
Non-Patent Citations (1)
Title |
---|
徐东 等: "滚动轴承加速寿命试验技术研究", 《国防科技大学学报》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020133954A1 (en) * | 2018-12-27 | 2020-07-02 | 帝沃检测技术(上海)有限公司 | Digital performance analysis method and device |
CN110308491A (en) * | 2019-05-29 | 2019-10-08 | 浙江大华技术股份有限公司 | Article detection method, device, storage medium and electronic device |
Also Published As
Publication number | Publication date |
---|---|
CN109613108B (en) | 2022-09-20 |
WO2020133954A1 (en) | 2020-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106412573B (en) | A kind of method and apparatus of detector lens stain | |
CN109613108A (en) | Threshold selection method and equipment | |
CN106851341B (en) | For measuring the method and mobile terminal of network video delivery quality | |
US20220206575A1 (en) | Method and apparatus for identifying gaze behavior in three-dimensional space, and storage medium | |
JP6288249B2 (en) | Visual inspection support device, visual inspection support method, and visual inspection support program | |
CN101539533A (en) | Automatic detecting device and method for battery internal defect | |
CN102901473A (en) | Correction system and method for measuring coordinate | |
CN105681124B (en) | A kind of wire-speed detection method and device | |
CN108268200A (en) | Image processing method and device, electronic equipment, computer program and storage medium | |
CN108896186A (en) | The method and device of equipment intelligent diagnosis | |
CN110261392A (en) | Quality determining method, device, electronic equipment and system | |
CN106595496A (en) | Man-machine interaction part size flexibility vision measurement method | |
CN108875709A (en) | One kind flocks together behavioral value method, apparatus, electronic equipment and storage medium | |
JP2007159934A (en) | Comparative diagnostic reading supporting apparatus | |
CN106326672B (en) | Sleep detection method and system | |
CN106502756B (en) | Application control method and device | |
CN109106999A (en) | A kind of infusion remaining time acquisition methods and infusion electronic claim | |
CN108268903A (en) | Article control method, device, readable storage medium storing program for executing and control terminal | |
CN105180821B (en) | Fiber winds the system and method for pitch in on-line measurement multiple tube | |
CN110852998B (en) | Taylor cone detection method and detection system based on deep learning | |
CN116369906A (en) | Automatic testing method, device, equipment and medium for arm bending suspension | |
WO2021206226A1 (en) | Image artificial intelligence-based method for estimating skin loss and moisture level | |
ITFI20120072A1 (en) | METHOD AND SYSTEM FOR IMPROVING THE VISUAL EXPLORATION OF AN IMAGE DURING THE SEARCH FOR A TARGET. | |
CN107767461A (en) | A kind of panoramic picture jump method | |
CN113095461A (en) | Pull-up counter based on machine vision |
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 | ||
CB03 | Change of inventor or designer information |
Inventor after: Xu Dong Inventor after: Wu Zheng Inventor before: Xu Dong |
|
CB03 | Change of inventor or designer information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |