System and method for selecting optimal detection wave band of cotton foreign fiber

Abstract

The invention relates to the related field of multispectral detection and multi-attribute decision, discloses a cotton foreign fiber optimal detection wave band optimization method based on an interval intuitionistic fuzzy set and a fusion weighting method, and aims to solve the problems that a traditional wave band division method is single in index and does not consider the correlation degree between different indexes and the like. The system comprises a data processing module M1 used for collecting and preprocessing foreign fiber spectrum data; the information feature module M2 is used for spectral information feature construction based on the preprocessed data; the multi-attribute decision-making module M3 is used for establishing a decision-making mechanism of the optimal detection wave band of the spectral data; and the result analysis module M4 is used for sensitivity analysis and method comparison. According to the method, the separability, the correlation and the information amount of the foreign fiber spectrum data are comprehensively considered, the weight information of different indexes is determined by using the fusion weighting method, the foreign fiber spectrum data are optimized based on the interval intuitionistic fuzzy aggregation operator, and a scientific method is provided for predicting the optimal detection spectrum of the foreign fibers in the cotton.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of multi-spectral detection technology and multi-attribute group decision-making method, and particularly relates to a system and method for selecting optimal detection wave bands of foreign fibers in cotton. BACKGROUND

[0002] Foreign fibers (commonly known as "three threads") in cotton refer to non-cotton fibers and colored fibers mixed into cotton during the process of cotton picking, spreading, purchasing, storage, transportation and processing, which have a serious impact on the quality of cotton and its products, such as hair, ropes, mulch, sugar paper, woven bag silk, etc. At present, image processing technology is mainly used to detect and identify foreign fibers in cotton. However, the detection effect of image processing technology on transparent, white and covered foreign fibers by cotton layer is not good, and multi-spectral detection technology can make up for this defect. Different wave bands of spectrum are used to identify different types of foreign fibers, which can effectively eliminate foreign fibers and improve the detection rate. However, the traditional wave band division method, such as inter-class separability wave band selection method and adaptive wave band selection method, cannot comprehensively consider the multi-index attributes of foreign fiber spectral information. How to establish a multi-index system of foreign fiber spectral information is a basic problem to be solved. The weight information of different index attributes will directly affect the final decision result. How to obtain suitable attribute index weight information is a key problem to be solved. Through the research on the change rule of frequency domain and amplitude in spectral image, combined with the multi-index system of foreign fiber spectral information, how to establish a multi-attribute index foreign fiber optimal wave band decision method is the core problem to be considered. SUMMARY

[0003] Therefore, the present application provides a system and method for selecting optimal detection wave bands of foreign fibers in cotton to solve the problems of single index, not considering the mutual relationship between multiple attribute indexes and ignoring index weight information in the traditional wave band division method. The interval intuitionistic fuzzy set and the characteristics and advantages of the fusion weighting method are used to construct a foreign fiber optimal detection wave band decision method.

[0004] The present application can be implemented by the following technical solutions:

[0005] Step 1, foreign fiber spectral information data acquisition and preprocessing;

[0006] Step 2, analyzing spectral characteristic attribute index to construct a foreign fiber spectral information characteristic index system, and extracting features from the preprocessed spectral data;

[0007] Step 3, constructing a feature matrix, determining a decision matrix and standardizing, and using an interval intuitionistic fuzzy multi-attribute decision-making method to construct a foreign fiber spectral optimal detection wave band decision mechanism;

[0008] Step 4: Based on the decision matrix obtained in Step 3, consider the correlation between the objectives and use the entropy weight-Critic fusion weighting method to determine the corresponding weight information of the decision indicators and attribute indicators.

[0009] Step 5: Calculate the support metric between decision sets based on the decision matrix obtained in Step 3. Solve for the support matrix of the decision set Sum of power weight matrices

[0010] Step 6: Use the WIVIFPMSM operator to aggregate different decision information from the data matrix obtained in Step 5. To a set matrix (α) ij ) m×n ;

[0011] Step 7: Calculate the support Sup(α) between attribute sets on the aggregation matrix obtained in Step 6. ij ,α ik ), to obtain the support matrix T(α) between attribute sets. ij ) and the power weight matrix η ij ;

[0012] Step 8: Submit solution X i All attributes are aggregated into the total value α. i =(α i1 ,α i2 ,...,α in ), (i=1,2,...,m), calculate the evaluation function S(α) i ),H(α i The ranking results are derived based on the evaluation criteria.

[0013] Step 9: Perform sensitivity analysis on the results and compare them with other multi-attribute methods.

[0014] The beneficial technical effects of this invention are as follows:

[0015] Spectrum acquisition device is used to collect spectrum data information of various hetero fibers, data preprocessing is carried out on the spectrum image, noise points are removed, the spectrum image is smoothed, and the reliability Cpk and the in-group and inter-group capacity Ppk of experimental data quality are analyzed. In view of the fact that the traditional wave band division method cannot comprehensively consider the spectrum information characteristic index, the frequency domain and the assignment of the spectrum image are analyzed, the analyzability, the correlation and the information amount between the wave bands are researched, the spectrum multi-attribute index of the hetero fiber is established, and the characteristic extraction is carried out. The spectrum wave band optimization problem of various attributes of various hetero fibers is divided into the category of multi-attribute decision, the data characteristics and the decision environment are analyzed, the interval intuitionistic fuzzy multi-attribute decision mechanism is used for selecting the best spectrum wave band, the interval intuitionistic fuzzy set theory and the selection function operator are deduced, the fusion weighting method is used for determining the multi-index attribute weight information, the decision mechanism of the best detection wave band of the spectrum data of the hetero fiber is constructed, the sensitivity index is used for analyzing the stability of the decision result, and the feasibility of the method is verified by comparing the decision results of other multi-attribute decision methods. In this way, a reasonable decision method is provided for the best spectrum selection of the multi-attribute index hetero fiber. The method not only considers the mutual relationship between different attribute indexes, removes the negative influence of extreme data values, and comprehensively considers the mutual relationship between data, scientifically determines the attribute index weight information, and simultaneously provides theoretical guidance for the best spectrum wave band selection and the expansion of the multi-attribute decision method. The present application belongs to the cross-research content of spectrum detection technology and multi-attribute decision method, has certain novelty and innovation, and is very powerful and convenient to popularize and use. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is the overall system function block diagram of the present application.

[0017] Figure 2 It is the theory and application correlation diagram of the present application.

[0018] Figure 3 It is the system running flow chart of the present application.

[0019] Figure 4 It is the spectrum characteristic curve of common hetero fiber.

[0020] Among them, Figure 1 It is the hetero fiber multi-spectrum information optimization system, and it is divided into four modules and corresponding execution functions. Figure 2 Through theoretical and application analysis, the best spectrum detection wave band decision mechanism of the hetero fiber under the interval direct fuzzy environment is researched. Figure 3 It is the specific execution process of the best spectrum detection wave band of the hetero fiber based on the WIVIFPMSM operator. Figure 4 It is the spectrum characteristic data of the collected various hetero fibers, and T1-T12 are common hetero fiber types. DETAILED DESCRIPTION

[0021] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings and preferred embodiments.

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative effort are all within the scope of protection of the present invention.

[0023] This invention provides a system and method for selecting the optimal detection band for foreign-type cotton fibers. In the multi-attribute decision problem of the optimal detection band for foreign fiber spectra, let X = {X1, X2, ..., X...} m As a set of m band schemes, C = {C1, C2, ..., C} n Let} be a set of n heterogeneous fiber attributes, D = {D1, D2, ..., D} s Let} be a set of s heterogeneous fiber decision sets, then any element D k (k = 1, 2, ..., s) represents each scheme X i (i = 1, 2, ..., m) For each attribute set C j The performance of (j = 1, 2, ..., n) using decision data. It means that w = {w1, w2, ..., w} n Let w' be the attribute set weight vector, where w' = {w'1, w'2, ..., w'}. s} represents the weight vector of the decision set, and then the interval intuitionistic fuzzy multi-attribute decision method is used to construct the decision mechanism for the optimal detection band of heterogeneous fiber spectroscopy.

[0024] The experimental materials for this invention were obtained from cotton processed by 10 cotton textile enterprises across China, including Hubei Gucheng Yinfang Company, after removing foreign fibers. Feathers, plastic film, polypropylene filament, and chemical fibers were found to be present in large quantities. These four types of foreign fibers were used as a decision set. Combining the general method of spectral band information division, the spectral information of foreign fibers was divided into 6 sub-bands as a scheme set. Addressing the deficiency of traditional band division methods with single indicators, inter-wavelength separability, correlation, and adaptive indicators were used as the attribute set of foreign fibers. Step 1: Spectral data of foreign fibers were collected using specialized spectroscopic equipment. For different foreign fibers, a Hitachi UH4150 UV-Vis-NIR spectrophotometer was used to collect the changes in reflectance and absorptivity in each band.

[0025] The data preprocessing function can perform smoothing denoising, sharpening and edge processing, etc. on the spectral data, study the data quality reliability and group capacity, remove the influence of extreme data, smooth the spectral curve, consider the errors in the data acquisition and processing process, and use interval value; Step 2: using adaptive waveband division method and inter-class separable waveband division method, determine the main attribute index of anisotropic fiber spectral data, construct anisotropic fiber spectral information characteristic index system, and extract features from the preprocessed spectral data;

[0026] Step 3: evaluate the three attributes, and express the evaluation information of the four types of anisotropic fibers on the six waveband sets. The non-membership degree is determined according to the critical edge condition of the intuitionistic fuzzy set as shown in formula (1). Combining the measurement error and the calculation error, the decision matrix is standardized according to formula (2). Since all the attributes in this example are benefit type, we do not need to standardize the decision matrix.

[0027]

[0028] Step 4: taking the decision set and the attribute set as the research objects respectively, the weight value is calculated by using the single weighting method, and then the fusion weight value is obtained. Formula (3) is the entropy value, formula (4) is the Critic value, and formula (5) is the fusion weight.

[0029]

[0030] Step 5: calculate the support of the anisotropic fiber decision set

[0031]

[0032] Calculate the decision set support matrix And the power weight matrix

[0033]

[0034] Step 6: using WIVIFPMSM aggregation operator to aggregate the decision information matrix of the anisotropic fiber decision set To the set matrix (α ij ) m×n ;

[0035] Step 7: calculate the support Sup(α ij ,α ik ),

[0036] Sup(α ij ,α ik )=1-D(α ij ,α ik), (j, k = 1, 2,..., n) (9)

[0037] The calculation attribute set supports matrix T'(a ij ) and power weight matrix η ij ;

[0038]

[0039] Step 8: aggregate all attributes of the wave band scheme set to the overall value a i =(a i1 , a i2 ,..., a in ), (i = 1, 2,..., m), calculate evaluation functions S(a i ), H(a i ), and obtain the ranking result according to the evaluation standard, to obtain the best detection wave band of the heterogeneous fiber considering multiple attribute indexes;

[0040]

[0041]

[0042] Step 9: sensitivity analysis is performed on the decision result to determine the stability of the result; and the experimental results of other multi-attribute decision methods are compared and analyzed to illustrate the effectiveness and practicality of the present application.

[0043] Although the specific embodiments of the present application are described above, those skilled in the art should understand that these are only illustrative, and various changes or modifications can be made to these embodiments without departing from the principles and essence of the present application, therefore, the protection scope of the present application is defined by the appended claims.

Claims

1. A system and method for selecting the optimal detection wavelength for cotton foreign fibers, characterized in that: The spectral data processing module M1 is used to preprocess the spectral data of heterogeneous fibers acquired using specialized spectral equipment, removing noise and smoothing the spectral curves. This module is the foundation for obtaining standardized data and has a significant impact on the accuracy of experimental results. The spectral information feature system construction module M2 is based on module M1. It studies the changes of multiple indicators based on the frequency domain and amplitude changes of spectral data, making up for the shortcomings of the single indicator in traditional band division, constructing a multi-indicator system of heterogeneous fiber spectral information, and extracting features. The multi-attribute decision mechanism module M3 is used for the derivation of interval intuitionistic fuzzy sets and the selection of functional operators. Based on the multi-index system obtained by module M2, it uses the multi-attribute decision aggregation matrix to fuse information with the normalized decision matrix obtained by module M1, and then evaluates the aggregated information according to the evaluation criteria to obtain the optimal detection spectral band, thus constructing a decision mechanism for the optimal detection band of heterogeneous fiber spectral data. The result analysis module M4 is used for sensitivity analysis of the decision results and comparison of the results obtained by various decision methods, thereby demonstrating the effectiveness and superiority of the fuzzy aggregation operator in the multi-attribute decision mechanism module M3.

2. The system and method for selecting the optimal detection band for cotton heterogeneous fibers according to claim 1, characterized in that: The spectral data processing module M1 uses specialized spectral equipment to collect spectral data of heterogeneous fibers and performs data preprocessing. The use of specialized spectroscopic equipment to collect spectral data of heterogeneous fibers involves using a Hitachi UH4150 UV-Vis-NIR spectrophotometer to collect the changes in reflectance and absorptivity of different heterogeneous fibers in various wavelength bands. The data preprocessing function can perform operations such as smoothing, denoising, sharpening, and edge processing on spectral data, and perform Weibull fitting curve analysis on spectral data to study the reliability of experimental data quality Cpk and the intra- and inter-group capabilities Ppk.

3. The system and method for selecting the optimal detection band for cotton heterogeneous fibers according to claim 1, characterized in that: The spectral information feature system construction module M2 includes the establishment of multi-attribute indicators of heterogeneous fiber spectral information and the extraction of heterogeneous fiber spectral data features. The establishment of the multi-attribute index for heterogeneous fiber spectral information is mainly aimed at addressing the shortcomings of the single nature of traditional band division indexes. It integrates multiple index attributes to construct a multi-attribute index for heterogeneous fiber spectral information. Formula (1) is the gradability index, formula (2) is the adaptive information content index, and formula (3) is the correlation index. Among them, the gradability index introduces parameters α and β to increase adjustability, and the adaptive information content index adds the mean R. i Influence; The feature extraction of heterogeneous fiber spectral data is achieved under the premise of a determined multi-index system.

4. The system and method for selecting the optimal detection band for cotton heterogeneous fibers according to claim 1, characterized in that: The problem of multi-index attribute information of heterogeneous fiber spectral information is mapped to the category of multi-attribute decision-making problem. The multi-attribute decision-making mechanism module M3 is applied to the multi-attribute decision-making problem of heterogeneous fiber with multiple types and multiple indices. It includes the derivation of interval intuition fuzzy operators, the combination of multiple functional operators, the determination of the weight information of indicators using the fusion weighting method, the information fusion of the normalized decision matrix obtained by module M1 using the multi-attribute decision aggregation matrix, and the evaluation of the aggregated information according to the evaluation criteria to obtain the optimal detection spectral band, thus constructing a decision-making mechanism for the optimal detection band of heterogeneous fiber spectral data. The derivation of the interval intuitionistic fuzzy operator is used to describe the information representation of the data decision matrix. The interval intuitionistic fuzzy operator, expressing membership and non-membership degrees in the form of interval values, can contain more fuzzy information, and has low operator complexity and low computational cost. Its expression form is as follows: This is the interval membership information. For non-interval membership information, satisfying The various functional operators include idempotent operators and Maclaurin symmetric average operators, which are used to eliminate the influence of unreasonable input parameters and measure the interrelationship between input parameters, respectively. The organic coupling of functional operators with interval direct fuzzy sets generates fuzzy aggregation operators with specific functions. The Critic method is used to compensate for the shortcomings of the entropy weight method. The two weighting methods are organically combined to construct a more comprehensive combined weight method, as shown in formula (4). This method considers both the degree of change within the attribute (entropy weight method) and the conflict and correlation between attributes (Critic method). It is suitable for complex decision-making scenarios, and the fused form can effectively avoid the subjectivity of linear weights. The entropy weight method is based on the weight of data dispersion. The original entropy weight formula may not be able to fully reflect the influence of hesitation. It is improved and the corresponding weight information is calculated, as shown in formula (6). The Critic method considers the correlation and contrast strength between attributes, and its weight is determined as shown in formula (10). Its advantage lies in the in-depth mining of the attribute data of the indicators, effectively eliminating information redundancy, and better meeting the actual evaluation needs. The weighting method determines the weight information of the indicators by calculating the corresponding entropy value and Critic value based on the normalized decision matrix obtained by module M1, and then performing algorithmic fusion using formula (4) to obtain the final weight after fusion.

5. The system and method for selecting the optimal detection band for cotton heterogeneous fibers according to claim 1, characterized in that: The results analysis module M4 uses sensitivity to analyze the impact of parameters and compares the results with those of other multi-attribute decision-making methods. The sensitivity analysis explores the impact of aggregation operator parameters in module M3 on the stability of decision results. The comparison of various multi-attribute decision-making methods is a way to illustrate the effectiveness and superiority of the multi-attribute decision-making methods described in module M3.

Images