CN101539517A - Method for evaluating high-gloss coating surfaces by fuzzy logic model - Google Patents
Method for evaluating high-gloss coating surfaces by fuzzy logic model Download PDFInfo
- Publication number
- CN101539517A CN101539517A CN200910061753A CN200910061753A CN101539517A CN 101539517 A CN101539517 A CN 101539517A CN 200910061753 A CN200910061753 A CN 200910061753A CN 200910061753 A CN200910061753 A CN 200910061753A CN 101539517 A CN101539517 A CN 101539517A
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- comprehensive evaluation
- value
- parameter
- glossiness
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention relates to a method for evaluating high-gloss coating surfaces by a fuzzy logic model. The method comprises the following steps: (1) selecting glossiness, haze and pebbling as analytical factors of comprehensive evaluation; (2) establishing a Mamdani fuzzy inference system model; (3) obtaining a fuzzy set of various parameters of the glossiness, the haze and the pebbling value; (4) obtaining a corresponding output value of the comprehensive evaluation and a corresponding fuzzy set; and (5) obtaining final comprehensive evaluation scores. The method has the following advantages: (1) flexible parameter selection and very good expandability of algorithm; (2) visual output results; and (3) being capable of easily comparing quality differences among different coatings according to a quantized coating surface comprehensive quality index, which has relatively positive effect in the coating field.
Description
Technical field
The invention belongs to the coating quality evaluation areas, relate to a kind of scoring of glossiness, mist shadow and three quality index of tangerine peel of the method synthesis high-gloss coating surfaces by fuzzy logic, high-gloss coating surfaces is estimated, made the evaluation result evaluation method consistent with the human eye vision effect.
Background technology
The quality of appearance of coat is one of index of estimating a product quality.At present the test event of evaluation table appearance quality has developed glossy, mist shadow, tangerine peel etc.Glossiness characterizes the light quantity of coating surface to certain orientation reflection incident light, and directly Fan She light quantity is big more, and the sensation of gloss is just obvious more.The mist shadow is meant that the regional area at coating surface mapping pattern presents the degree of vaporific haloing, and it has reflected the intensity of diffuse light.The coating surface outward appearance presents many semicircle shape projections, is called tangerine peel as the ripple of tangerine peel, and it has reflected that coating surface rises and falls highly to the influence degree of reflection ray.During its testing standard and detecting instrument are also constantly improving.Human eye can go it is estimated when observing an object from different aspects, also can form an overall impression to object, and this overall impression directly influences the standard that people judge an object quality.
In the comprehensive evaluation in early days, mainly be to come the coating surface quality is estimated by the mean value of asking for glossiness, mist shadow and the every quality index of tangerine peel, this method has been ignored the influence degree of every quality index to the human eye vision effect.
People came the coating surface quality is carried out comprehensive evaluation by the marking of comprehensive each quality index of coating of experimental formula afterwards.Be that a coefficient is set in glossiness, mist shadow and every coating quality index marking of tangerine peel in experimental formula, this coefficient determines the influence degree of every quality index to human eye subjective vision effect.Estimate the overall quality of coating surface by the linear combination of every quality index and coefficient product.But human eye when coating quality is carried out comprehensive evaluation, it is complicated many that the mutual contact of each coating quality index is wanted, because this process is the nonlinear system of importing more.Therefore the method that adopts experimental formula to make up every quality index formation comprehensive evaluation can not be simulated the visual effect of human eye accurately.
Classic method according to a series of " ... if ... so ... " empirical sentence make judgement, obtain this shape as " ", conclusion such as " better ", " good ".Fuzzy inference system utilizes fuzzy rule to carry out reasoning just, expresses transitional boundary or Qualitative Knowledge experience, and simulation human brain mode is carried out fuzzy synthesis and judged that reasoning solves the regular pattern composite fuzzy message problem that conventional method is difficult to tackle.Therefore use the method for fuzzy logic the assurance system to give a mark more near the human vision evaluation.
The fuzzy reasoning method that is widely used now is the fuzzy inference system of Manda Buddhist nun (Mamdan i), therefore system adopts the model of Mamdani-type as overall evaluation system, model with input variable through obfuscation-＞reasoning-＞process of ambiguity solution gets comprehensive marking to the end.
Summary of the invention
The purpose of this invention is to provide a kind of method simulation human eye visual effect that adopts fuzzy logic, the quality marking of the glossiness of comprehensive coating surface, mist shadow, tangerine peel, based on some texture or the optical signature evaluation of body surface, obtain one of high reflectance body surface fully near the method for the comprehensive evaluation of human vision by analysis-by-synthesis to these evaluations.
In order to realize above purpose, the method that the present invention adopts is:
First step: the choosing of evaluation index: choose the analytical factor that glossiness, mist shadow and tangerine peel are used as analyzing the comprehensive evaluation of high reflectance body surface;
Second step:, set up Mamdani fuzzy inference system model according to using traditional evaluation method that on-gauge plate is tested the gained data as training set;
Third step: will measure the glossiness, mist shadow, the tangerine peel value input state parameter that obtain and carry out the index obfuscation, and obtain the fuzzy set of each parameter;
The 4th step: the occurrence according to the input state of each sampling instant is measured, utilize every rule in the fuzzy inference rule storehouse to carry out fuzzy reasoning, obtain the output valve and the corresponding fuzzy set of the comprehensive evaluation of this inference rule correspondence;
The 5th step: corresponding comprehensive evaluation output valve and fuzzy set are weighted merging, obtain final comprehensive evaluation output collection, at last the output collection is asked center of gravity, with the corresponding abscissa value of center of gravity as the comprehensive evaluation mark: the quantizing range of comprehensive evaluation output valve can specifically be set according to concrete applicable cases, the quantizing range of output valve is defined between the 0-99.9,0 presentation surface quality extreme difference or input measurement value are illegal, and 100 presentation surface quality are fabulous or be level crossing.
The invention has the advantages that: 1, flexible parameter selection, algorithm has favorable expansibility.The user can oneself need select the state parameter of investigation, even can be according to the measurement data custom parameter of oneself; 2, output visual result, the user need not to understand the practical significance of measurement parameter, as long as according to last output result, just can judge the coating quality on surface to be measured; 3, according to the coating surface overall quality index after quantizing, can be easy to the mass discrepancy between the comparison different coating, this has comparatively positive meaning in field of coating.
Description of drawings
Fig. 1 is the membership function curve of input state parameter tangerine peel of the present invention.
Fig. 2 is the membership function curve of input state parameter mist shadow of the present invention.
Fig. 3 is the membership function curve of input state parameter glossiness of the present invention.
Fig. 4 is comprehensive evaluation output membership function curve of the present invention.
The decision tree that Fig. 5 utilizes traditional decision-tree that experimental data is constructed for the present invention.
Fig. 6 is the employed rule base of the embodiment of the invention.
Embodiment
The present invention is described in further detail below in conjunction with drawings and Examples.
A kind of high reflectance body surface overall quality quantization method that the present invention proposes based on fuzzy logic, and according to quantized data the body surface overall quality is estimated, specific implementation method is as follows:
First step: at present the test event of evaluation table appearance quality has developed glossy, mist shadow, tangerine peel etc., thus at first evaluation index choose the analytical factor that glossiness, mist shadow and tangerine peel are analyzed the comprehensive evaluation of high reflectance body surface.
Second step:, set up Mamdani fuzzy inference system model according to using traditional evaluation method that on-gauge plate is tested the gained data as training set.What the present invention used is the method for knowledge excavation, extracts the rule that satisfies system accuracy from multidigit expert's marking and the measured result of same category of device, sets up decision tree.
Third step: will measure the glossiness, mist shadow, the tangerine peel value input state parameter that obtain and carry out the index obfuscation, and obtain the fuzzy set of a parameter.
1, the present invention carries out quantification treatment to three input state amounts: present embodiment is divided into glossiness, mist shadow, tangerine peel degree basic, normal, high according to size, being the ambiguity in definition collection is { low, in, height }, be its definition membership function (Fig. 1-3), curve Low, Mid, High represent three states respectively, horizontal ordinate is the actual value of input parameter, ordinate is a degree of membership, and 0 expression does not belong to this state fully, and 1 expression belongs to this state fully.
2, the quantizing range and the obfuscation parameter (Fig. 4) of comprehensive evaluation output valve are set: present embodiment with comprehensive evaluation output valve scope definition between 0-99.9,0 presentation surface quality extreme difference or input measurement value are illegal, and 100 presentation surface quality are fabulous or be level crossing.And with comprehensive evaluation output valve poor location, general, good, fabulous four ranks.Horizontal ordinate is comprehensive marking value, and ordinate is a degree of membership.
The 4th step: the occurrence according to the input state of each sampling instant is measured, utilize every rule in the fuzzy inference rule storehouse to carry out fuzzy reasoning, obtain the output valve and the corresponding fuzzy set of the comprehensive evaluation of this inference rule correspondence.Fuzzy inference rule is at first determined the degree of membership of parameter as shown in Figure 6 according to membership function, 7 rules that each input value just needed Fig. 6 are carried out reasoning, all obtain an output valve O at every turn
i, and should rule and the matching degree α of actual conditions
i
The 5th step: corresponding comprehensive evaluation output valve and fuzzy set are weighted merging, obtain final comprehensive evaluation output collection, at last the output collection is asked center of gravity, with the corresponding abscissa value of center of gravity as the comprehensive evaluation mark.At first the membership function of each the input state parameter in the fuzzy inference rule is obtained the degree of membership of this parameter, value with the degree of membership minimum is as the criterion then, in conjunction with pairing output parameter state grade membership function in this fuzzy inference rule, obtain the comprehensive evaluation output valve of this fuzzy inference rule and the matching degree of this value.
The method of in the described third step parameter being carried out the concrete operations of obfuscation comprises:
The first step: the measurement numerical values recited according to each selected input parameter is divided into a plurality of state grades; For example, can be divided into three state grades according to the measured value size of this state parameter of size of glossiness: high, medium and low, the fine or not degree of corresponding glossiness respectively;
Second step: the measured value that each input state parameter of expression is set is to the membership function of subjection degree that should the parameter state grade; The real number representation of degree of membership between the 0-1.
Described obfuscation parameter can comprise: the size of described comprehensive evaluation output valve is divided into several ranks, defines corresponding membership function.
The fuzzy inference rule storehouse is to conclude according to expert's practical experience knowledge to obtain in described the 4th step, perhaps utilize the method for knowledge excavation to concentrate extraction to obtain from a large amount of empirical datas, wherein every fuzzy inference rule by the different conditions grade of described a plurality of input state parameters by certain and/or the condition that constitutes of logical relation, corresponding to the corresponding state grade of a described load output value, this inference rule can with " if ... then ... " conditional statement represent.What the present invention used is the method for knowledge excavation, extracts the rule that satisfies system accuracy from multidigit expert's marking and the measured result of same category of device, sets up decision tree, concrete grammar:
The first step: use the value that on-gauge plate is measured to train, obtain decision tree, as shown in Figure 5.
Second step: according to the membership function of output valve, each result is generally derived by 2-3 rule, so with original decision tree beta pruning, obtain the sub-tree of 10 leafy nodes.Decision tree is converted into the rule of correspondence to be represented.
Weighting merging disposal route is in described the 5th step: will its comprehensive evaluation output valve be computed weighted by the matching degree that every inference rule obtains, obtain a comprehensive output fuzzy set the heaviest, again this fuzzy set is asked the center of gravity computing, the abscissa value of corresponding center of gravity is as comprehensive evaluation value.This fuzzy system can be expressed as:
AI＝f(OP，DOI，GLS)，AI?∈(0，99.9)，
The AI presentation surface index of comprehensively giving a mark wherein, OP represents the tangerine peel input value, and DOI represents mist shadow input value, and GLS represents the glossiness input value.
Parameter fuzzy processing procedure of the present invention is exactly that original accurately input parameter is carried out Fuzzy processing according to predefined good fuzzy set and membership function, state with a state and a parameter of degree of membership relationship description, such as, be divided into low, medium the degree of tangerine peel and a Senior Three state, current tangerine peel input value is 30, through after the Fuzzy processing, right state degree of membership is 0.6 just can be described as current tangerine peel phenomenon according to membership function, is 0.08 to the degree of membership of medium state.
Fuzzy inference rule is the committed step of carrying out comprehensive evaluation, and it is the knowledge data base that controlled device is controlled, and the method for fuzzy reasoning is in conjunction with inference rule, carries out reasoning according to previously selected reasoning algorithm.Every inference rule all can act on mutually to the input parameter amount, provides an output valve O then
i, and should rule and the matching degree α of actual conditions
iFuzzy reasoning method of the present invention is exactly a kind of concrete application of Mamdari Fuzzy Logic Reasoning Algorithm.The implementation procedure of this algorithm mainly comprises: the degree of membership that at first membership function of each the input state parameter in the fuzzy inference rule is obtained this parameter, value with the degree of membership minimum is as the criterion then, in conjunction with pairing output parameter state grade membership function in this fuzzy inference rule, obtain the comprehensive evaluation output valve of this fuzzy inference rule and the matching degree of this value.
The content that is not described in detail in this instructions belongs to this area professional and technical personnel's known prior art.
Claims (5)
1, a kind of fuzzy logic model method that high-gloss coating surfaces is estimated, its method is:
First step: the choosing of evaluation index: choose the analytical factor that glossiness, mist shadow and tangerine peel are used as analyzing the comprehensive evaluation of high reflectance body surface;
Second step: use traditional evaluation method that on-gauge plate is tested the gained data as training set according to existing data, set up Mamdani fuzzy inference system model;
Third step: will measure the glossiness, mist shadow, the tangerine peel value input state parameter that obtain and carry out the index obfuscation, and obtain the fuzzy set of each parameter;
The 4th step: the occurrence according to the input state of each sampling instant is measured, utilize every rule in the fuzzy inference rule storehouse to carry out fuzzy reasoning, obtain the output valve and the corresponding fuzzy set of the comprehensive evaluation of this inference rule correspondence;
The 5th step: corresponding comprehensive evaluation output valve and fuzzy set are weighted merging, obtain final comprehensive evaluation output collection, at last the output collection is asked center of gravity, with the corresponding abscissa value of center of gravity as the comprehensive evaluation mark: the quantizing range of comprehensive evaluation output valve is specifically set according to concrete applicable cases, the quantizing range of output valve is defined between the 0-100,0 presentation surface quality extreme difference or input measurement value are illegal, and 100 presentation surface quality are fabulous or be level crossing.
2, fuzzy logic model as claimed in claim 1 method that high-gloss coating surfaces is estimated is characterized in that: the method for in the described third step parameter being carried out the concrete operations of obfuscation comprises:
The first step: the measurement numerical values recited according to selected each input parameter is divided into a plurality of state grades: the measured value size according to this state parameter of size of glossiness is divided into three state grades: high, medium and low, and the fine or not degree of corresponding glossiness respectively;
Second step: the measured value that each input state parameter of expression is set is to the membership function of subjection degree that should the parameter state grade; The real number representation of degree of membership between the 0-1.
3, fuzzy logic model as claimed in claim 1 or 2 method that high-gloss coating surfaces is estimated, it is characterized in that: described obfuscation parameter comprises: the size of described comprehensive evaluation output valve is divided into several ranks, defines corresponding membership function.
4, the method that fuzzy logic model as claimed in claim 1 is estimated high-gloss coating surfaces, it is characterized in that: the fuzzy inference rule storehouse is to conclude according to expert's practical experience knowledge to obtain in described the 4th step, perhaps utilize the method for knowledge excavation to concentrate extraction to obtain from a large amount of empirical datas, wherein every fuzzy inference rule by the different conditions grade of described a plurality of input state parameters by certain and/or the condition that constitutes of logical relation, corresponding state grade corresponding to a described load output value, this inference rule can with " if ... then ... " conditional statement represent, wherein from multidigit expert's marking and the measured result of same category of device, extract the rule that satisfies system accuracy, set up decision tree, concrete grammar is:
The first step: use the value that on-gauge plate is measured to train, obtain decision tree;
Second step: according to the membership function of output valve, each result is derived by 2-3 rule, with original decision tree beta pruning, obtains the sub-tree of 10 leafy nodes, decision tree is converted into the rule of correspondence represents.
5, fuzzy logic model as claimed in claim 1 method that high-gloss coating surfaces is estimated, it is characterized in that: weighting merging disposal route is in described the 5th step: will its comprehensive evaluation output valve be computed weighted by the matching degree that every inference rule obtains, obtain a comprehensive output fuzzy set the heaviest, again this fuzzy set is asked the center of gravity computing, the abscissa value of corresponding center of gravity is as comprehensive evaluation value, and this fuzzy system is expressed as:
AI＝f(OP，DOI，GLS)，AI∈(0，99.9)，
The AI presentation surface index of comprehensively giving a mark wherein, OP represents the tangerine peel input value, and DOI represents mist shadow input value, and GLS represents the glossiness input value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910061753A CN101539517A (en) | 2009-04-27 | 2009-04-27 | Method for evaluating high-gloss coating surfaces by fuzzy logic model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910061753A CN101539517A (en) | 2009-04-27 | 2009-04-27 | Method for evaluating high-gloss coating surfaces by fuzzy logic model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101539517A true CN101539517A (en) | 2009-09-23 |
Family
ID=41122823
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910061753A Pending CN101539517A (en) | 2009-04-27 | 2009-04-27 | Method for evaluating high-gloss coating surfaces by fuzzy logic model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101539517A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103439398A (en) * | 2013-09-06 | 2013-12-11 | 北华航天工业学院 | Radon detection system and method based on fuzzy logic |
| CN104040320A (en) * | 2011-11-16 | 2014-09-10 | 涂层国外知识产权有限公司 | Process for predicting degree of mottling in coating compositions by wet color measurement |
| CN104136913A (en) * | 2011-11-01 | 2014-11-05 | 涂层国外知识产权有限公司 | Process for predicting metallic gloss of coating resulting from coating compositions by wet color measurement |
| CN104819831A (en) * | 2015-05-05 | 2015-08-05 | 上海机动车检测中心 | Method for quantitatively evaluating light distribution performance of automotive headlamp |
| CN111882495A (en) * | 2020-07-05 | 2020-11-03 | 东北林业大学 | Image highlight processing method based on user-defined fuzzy logic and GAN |
| CN112235566A (en) * | 2020-10-09 | 2021-01-15 | 华侨大学 | Network video quality assessment method and system combining decision tree and fuzzy inference |
| CN113537718A (en) * | 2021-06-18 | 2021-10-22 | 宝钢日铁汽车板有限公司 | Automatic evaluation method for cold rolling oiling machine steel strip missing coating quality |
-
2009
- 2009-04-27 CN CN200910061753A patent/CN101539517A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104136913A (en) * | 2011-11-01 | 2014-11-05 | 涂层国外知识产权有限公司 | Process for predicting metallic gloss of coating resulting from coating compositions by wet color measurement |
| CN104136913B (en) * | 2011-11-01 | 2016-10-05 | 涂层国外知识产权有限公司 | Utilize the method that the paint metal gloss being obtained by paint ingredient is predicted in the measurement of wet look |
| CN104040320A (en) * | 2011-11-16 | 2014-09-10 | 涂层国外知识产权有限公司 | Process for predicting degree of mottling in coating compositions by wet color measurement |
| CN104040320B (en) * | 2011-11-16 | 2016-08-17 | 涂层国外知识产权有限公司 | For the method utilizing wet color to measure the mottle degree in prediction paint ingredient |
| CN103439398A (en) * | 2013-09-06 | 2013-12-11 | 北华航天工业学院 | Radon detection system and method based on fuzzy logic |
| CN104819831A (en) * | 2015-05-05 | 2015-08-05 | 上海机动车检测中心 | Method for quantitatively evaluating light distribution performance of automotive headlamp |
| CN104819831B (en) * | 2015-05-05 | 2017-07-07 | 上海机动车检测认证技术研究中心有限公司 | Car headlamp luminous intensity distribution performance quantitative evaluating method |
| CN111882495A (en) * | 2020-07-05 | 2020-11-03 | 东北林业大学 | Image highlight processing method based on user-defined fuzzy logic and GAN |
| CN112235566A (en) * | 2020-10-09 | 2021-01-15 | 华侨大学 | Network video quality assessment method and system combining decision tree and fuzzy inference |
| CN113537718A (en) * | 2021-06-18 | 2021-10-22 | 宝钢日铁汽车板有限公司 | Automatic evaluation method for cold rolling oiling machine steel strip missing coating quality |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101539517A (en) | Method for evaluating high-gloss coating surfaces by fuzzy logic model | |
| CN103730130B (en) | A kind of detection system of pathological voice | |
| CN110009605A (en) | A kind of stone age prediction technique and system based on deep learning | |
| CN105426889A (en) | PCA mixed feature fusion based gas-liquid two-phase flow type identification method | |
| CN111797364B (en) | Landslide multilayer safety evaluation method based on composite cloud model | |
| CN107122548B (en) | Product modeling style-oriented modeling method for user cognition and quantification model | |
| CN108334838A (en) | Based on light remotely-sensed data National Macroscopic economic indicator estimating and measuring method | |
| CN104713835A (en) | Online numerical recognition method for colors of tobacco leaves | |
| Suwartono et al. | Validation of the Emotion Regulation Questionnaire (ERQ): network analysis as an alternative of confirmatory factor analysis (CFA) | |
| CN110852181A (en) | Piano music score difficulty identification method based on attention mechanism convolutional neural network | |
| CN108229592A (en) | Outlier detection method and device based on GMDH neuroids | |
| CN107274065A (en) | The subjective assessment of food taste and flavor and spectroscopic data modeling method and system | |
| Wang et al. | Scientific creativity research based on generalizability theory and BP_Adaboost RT | |
| CN109567746B (en) | Texture roughness quantitative evaluation method based on signal detection theory | |
| CN100454290C (en) | Cigarette organoleptic quality qualitative index estimating method | |
| Richards et al. | Optimisation of competition indices using simulated annealing and artificial neural networks | |
| CN107578785A (en) | The continuous emotional feature analysis evaluation method of music based on Gamma distributional analysis | |
| CN110517111A (en) | A kind of product customization method | |
| CN112685562B (en) | XGboost model-based multidimensional index integration technical evaluation method | |
| CN115153549A (en) | BP neural network-based man-machine interaction interface cognitive load prediction method | |
| Zhang | Benefit evaluation of human resource management in agricultural enterprises based on convolutional neural network. | |
| CN113657726A (en) | Personnel risk analysis method based on random forest | |
| CN113361653A (en) | Deep learning model depolarization method and device based on data sample enhancement | |
| CN113205274A (en) | Quantitative ranking method for construction quality | |
| Matisoff et al. | Mathematical & Statistical Analysis of BLOODSTAIN PATTERN EVIDENCE Part II |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20090923 |