Multi-dimensional tricholoma matsutake quality grading detection device

Abstract

The invention belongs to the field of tricholoma matsutake multi-dimensional detection and quality classification detection equipment, and provides a multi-dimensional tricholoma matsutake quality classification detection device which comprises an impedance detection sub-device, a hyperspectral detection sub-device, an electronic nose detection sub-device and a data processing station, the impedance detection sub-device comprises a tricholoma matsutake information acquisition module; the tricholoma matsutake information acquisition module comprises a detection electrode plate, a spring and a pressure sensing patch. According to the invention, clamping and surface data acquisition of a detected target tricholoma matsutake are realized through the impedance detection sub-device, and the ductility and stability of the impedance detection sub-device are improved by using a spring and a rotatable detection electrode plate; the appearance and impedance of the tricholoma matsutake can be detected at the same time through the pressure sensing patch and the detection electrode plate, and the detection efficiency is improved; multidimensional nondestructive detection on the appearance, smell and components of the tricholoma matsutake can be realized by combining hyperspectrum and electronic nose detection; and the whole process of the tricholoma matsutake from freshness to corruption can be monitored in real time through the data processing station.

Description

Technical Field

[0001] The present invention relates to the field of multi-dimensional detection and quality classification detection equipment for Matsutake mushrooms, and in particular to a multi-dimensional Matsutake mushroom quality grading detection device. Background Art

[0002] Matsutake is a precious wild edible mushroom. Due to the growth environment, the quality of different individuals often varies greatly. Its quality directly affects the market value, so it will be graded before being sold. Traditional manual detection methods and conventional physical and chemical detection methods can only be used to screen based on the size of the Matsutake. However, the quality of Matsutake is not only reflected in the size, but also involves multiple factors such as appearance, chemical composition, odor, and insect pest residues. A single detection method can often only capture information in a certain dimension, and the comprehensive quality of Matsutake requires cross-validation of multiple indicators.

[0003] The core of adopting multi-dimensional detection technology is to integrate the advantages of multiple means to achieve comprehensive detection of multiple key qualities of matsutake. Although each single technology has its own advantages and disadvantages, its limitations are obvious: for example, impedance detection can effectively detect the internal structural characteristics by analyzing the changes in conductivity and resistance inside matsutake. However, impedance measurement is not effective for matsutake with complex shapes, and is easily affected by sample size and environmental factors. At the same time, it is unable to evaluate external qualities such as smell or appearance; hyperspectral imaging technology can analyze the appearance, color and some internal components of matsutake by capturing light reflections of different wavelengths. However, spectral technology has poor detection effects on samples with surface contamination, and cannot accurately evaluate problems such as insect pests or holes inside matsutake; electronic nose detection determines the freshness, presence of insect pests or decay by detecting the volatile organic compounds of matsutake. However, electronic noses can only detect odor-related indicators, and it is difficult to obtain appearance and internal structure information. They are also easily disturbed by environmental changes, such as temperature and humidity, which increases the possibility of misjudgment. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a multi-dimensional matsutake quality grading detection device, which realizes the clamping and surface data collection of the detection target matsutake through an impedance detection sub-device, and uses springs and rotatable detection electrode sheets to improve the ductility and stability of the impedance detection sub-device; realizes the simultaneous detection of the shape and impedance of matsutake through pressure sensing patches and detection electrodes, thereby improving the detection efficiency; realizes multi-dimensional non-destructive detection of the shape, smell and composition of matsutake through the combination of hyperspectral and electronic nose detection; and monitors the entire process of matsutake from freshness to corruption in real time through a data processing station.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A multi-dimensional matsutake quality grading detection device, comprising: an impedance detection sub-device, a hyperspectral detection sub-device, an electronic nose detection sub-device and a data processing station; the impedance detection sub-device comprises: a back-end circuit, a receiving wire, an excitation wire and a matsutake information acquisition module; the matsutake information acquisition module comprises: an outer ring, a rotatable detection electrode sheet, a spring, a pressure sensor output wire, a first electrode connector wire, a second electrode connector wire and a pressure sensing patch; the detection electrode sheet comprises: a rotating disk, a rectangular electrode sheet, a reserved hole and a fixed point;

[0007] The data processing station is connected to the impedance detection sub-device, the hyperspectral detection sub-device and the electronic nose detection sub-device respectively; the back-end circuit is connected to the second electrode connector wire through the excitation wire; the first electrode connector wire and the pressure sensor output wire are respectively connected to the back-end circuit through a receiving wire; the back-end circuit is connected to the data processing station; the six pressure sensing patches are evenly fixed on the outer wall of the outer ring; a spring is fixed on the inner side of each pressure sensing patch; a rotating disk is fixed on the end of each spring away from the outer ring; the rotating disk is attached and fixed to the surface of the rectangular electrode sheet; a reserved hole is respectively provided at both ends of the rectangular electrode sheet; each of the reserved holes is provided with a fixed point; adjacent detection electrode sheets are connected through the first electrode connector wire and the second electrode connector wire; the first electrode connector wire and the second electrode connector wire pass through the two reserved holes of each detection electrode sheet respectively, and are fixed at the fixed point corresponding to the reserved hole; the pressure sensor output wire is connected to each pressure sensing patch;

[0008] The second electrode connector wire is used to transmit the acquisition excitation output by the back-end circuit to the detection electrode sheet; the detection electrode sheet is used to apply the acquisition excitation to the surface of the target detection matsutake; the first electrode connector wire is used to transmit the electrical feedback information of the detection electrode sheet to the back-end circuit through the receiving wire; the pressure sensor output wire is used to transmit the mechanical feedback information output by the pressure sensor patch to the back-end circuit through the receiving wire; the back-end circuit is used to transmit the acquisition excitation to the second electrode connector wire through the excitation wire, receive and pre-process the electrical feedback information and the mechanical feedback information, obtain acquisition data, and transmit the acquisition data to the data processing station; the hyperspectral detection sub-device is used to collect the hyperspectral data of the target detection matsutake, and transmit the hyperspectral data to the data processing station; the electronic nose detection sub-device is used to collect the odor information of the target detection matsutake, and transmit the odor information to the data processing station; the data processing station is used to perform multi-dimensional freshness detection and real-time monitoring of the target detection matsutake according to the acquisition data, the hyperspectral data and the odor information.

[0009] Preferably, the back-end circuit is embedded with: a microprocessor unit, a constant current source signal generating unit, a front-end measuring unit, a response signal collecting and processing unit and an RS232 interface circuit unit;

[0010] The constant current source signal generating unit is used to generate the acquisition excitation; the front-end measurement unit is used to amplify, filter and suppress noise of the electrical feedback information and the mechanical feedback information to obtain the input signal; the response signal acquisition and processing unit is used to perform data analysis on the input signal to obtain the acquisition data; the RS232 interface circuit unit is used to send the acquisition data to the data processing station; the microprocessor unit is used to control the constant current source signal to generate the acquisition excitation and control the RS232 interface circuit unit to send the acquisition data to the data processing station.

[0011] Preferably, the data processing station has an embedded multi-dimensional comprehensive quality scoring system; the multi-dimensional comprehensive quality scoring system includes: a first quality evaluation subsystem, a second quality evaluation subsystem, a third quality evaluation subsystem and a multi-dimensional comprehensive quality scoring subsystem; the first quality evaluation subsystem includes: a Matsutake cross-section detection model, a Matsutake column uniformity detection model, a Matsutake freshness detection model and an impedance-based quality parameter calculation model;

[0012] The Matsutake cross-section detection model is used to calculate the cross-sectional area of ​​the target detected Matsutake using the pressure sensing patch; the Matsutake column uniformity detection model is used to calculate the Matsutake column uniformity data of the target detected Matsutake using the pressure sensing patch; the Matsutake freshness detection model is used to calculate the initial freshness judgment data of the target detected Matsutake using the detection electrode sheet; the quality parameter calculation model is used to perform data fusion on the Matsutake cross-sectional area, the Matsutake column uniformity data and the initial freshness judgment data to obtain the Matsutake quality impedance score; the second quality evaluation subsystem is used to obtain the Matsutake quality hyperspectral score using the hyperspectral detection sub-device; the third quality evaluation subsystem is used to obtain the Matsutake quality electronic nose score using the electronic nose detection sub-device; the multi-dimensional comprehensive quality scoring subsystem is used to perform data fusion on the Matsutake quality impedance score, the Matsutake quality hyperspectral score and the Matsutake quality electronic nose score to obtain the overall Matsutake quality score.

[0013] Preferably, the calculation process of the cross-sectional area of ​​the Tricholoma matsutake and the uniformity data of the Tricholoma matsutake column includes:

[0014] Using the pressure sensing patch to collect pressure data of the target detection matsutake mushroom to obtain six-position pressure values;

[0015] The radius is calculated by using the six-position pressure values ​​and the probe design parameters of the pressure sensor patch to obtain six-position matsutake radius data;

[0016] The six-position Matsutake radius data are fitted using polygon interpolation to obtain the cross-sectional area of ​​the Matsutake;

[0017] The standard deviation of the six-direction Matsutake radius data is calculated, and the standard deviation is normalized to obtain the Matsutake column uniformity data.

[0018] Preferably, the calculation process of the freshness initial judgment data includes:

[0019] Using the detection electrode sheet to collect electrical data of the target detection Tricholoma matsutake to obtain the electrical feedback information;

[0020] The real part orthogonal anisotropy characteristic parameter and the imaginary part orthogonal anisotropy characteristic parameter of the electrical feedback information are calculated by using a difference method to obtain an orthogonal anisotropy characteristic vector;

[0021] Extracting the angle value and weight value of the orthogonal anisotropic eigenvector;

[0022] Constructing a time series set, a characteristic signal sample matrix and a mapping function based on the angle value and the weight value;

[0023] Calculating morphological characteristic parameters and integral values ​​using the angle value and the weight value;

[0024] Based on the time series set, the characteristic signal sample matrix and the mapping function, freshness judgment is performed using the morphological characteristic parameters and the integral value to obtain the initial freshness judgment data.

[0025] Preferably, the weight value calculation process includes:

[0026] Adding the orthogonal anisotropic eigenvector to the vector diagram to obtain vector coordinates;

[0027] Calculate the angle between the vector coordinates and the 45° standard line to obtain angle data;

[0028] The weight value is obtained by using the angle data to calculate the weight; the calculation formula of the weight value is: Where, q = 1, 2, 3; is the weight value corresponding to the qth orthogonal anisotropic eigenvector; θ q is the angle data corresponding to the qth orthogonal anisotropic eigenvector; θ 1 ,θ 2 and θ 3 They are the angle data corresponding to the first, second and third orthogonal anisotropic eigenvectors respectively.

[0029] Preferably, the process of constructing the time series set includes:

[0030] Obtain impedance amplitude samples based on the orthogonal anisotropic eigenvector; the calculation formula for obtaining the impedance amplitude samples is:

[0031]

[0032] in, is the impedance amplitude sample; R d is the dimension; C 0 is the time series set; → represents mapping; is the weight value at the i-th time point; x * i and * i are the impedance value and phase value collected at the i-th time point respectively; X and Y are the impedance data set and phase data set respectively; t 0 and t 1 are the start time and end time of the time series respectively;

[0033] A phase sample is obtained based on the orthogonal anisotropic eigenvector; the calculation formula of the phase sample is:

[0034]

[0035] in, is the phase sample; x * ,y * are data points corresponding to the impedance value and the phase value respectively; and are the nth ZI and ZA phase points collected respectively; f(y) is the phase transformation function; and is the mth ZI and ZA phase point after f(y) transformation;

[0036] Data mapping is performed on the impedance amplitude samples and the phase samples to obtain the time series set; the calculation formula of the time series set is:

[0037] P: {k cj ,k c+1·j ,k c·c+1 ,…k c+n·j ,k c·j+n}∈P M d ;

[0038] Among them, P M d is the time series set in the feature space with dimension d greater than 10; P is the time series set; k cj is the element located at (c, j) in the matrix corresponding to the time series set.

[0039] Preferably, the calculation process of the Tricholoma matsutake quality impedance score includes:

[0040] The cross-sectional area of ​​the Tricholoma matsutake, the uniformity data of the Tricholoma matsutake column and the initial freshness judgment data are standardized to obtain standard cross-sectional area data, standard uniformity data and standard freshness data; the calculation formulas of the standard cross-sectional area data, the standard uniformity data and the standard freshness data are respectively: u norm =u and Among them, A norm 、u norm and D norm are the standard cross-sectional area data, the standard uniformity data and the standard freshness data respectively; A, u and D are the cross-sectional area of ​​the pine mushroom, the uniformity data of the pine mushroom column and the initial freshness judgment data respectively; A max and A min are the maximum and minimum cross-sectional areas of the Tricholoma matsutake, respectively; Dmax and D min are respectively the maximum and minimum values ​​of the freshness initial judgment data;

[0041] The importance weights of the standard cross-sectional area data, the standard uniformity data and the standard freshness data are allocated to obtain importance data; the importance data satisfies the condition: w A +w u +w D =1; where w A 、w u and w D are the importance weights of the standard cross-sectional area data, the standard uniformity data, and the standard freshness data respectively;

[0042] The importance data is used to perform weighted summation on the standard cross-sectional area data, the standard uniformity data, and the standard freshness data to obtain the Matsutake quality impedance score.

[0043] Preferably, the process of obtaining the high-spectral score of the Matsutake quality and the electronic nose score of the Matsutake quality includes:

[0044] Using the hyperspectral detection sub-device to collect optical data of the target detection Tricholoma matsutake to obtain the hyperspectral data;

[0045] Using the electronic nose detection sub-device to collect chemical data of the target detection matsutake mushroom to obtain the odor information;

[0046] Calculating the covariance of the hyperspectral data and the odor information respectively to obtain a first covariance matrix and a second covariance matrix;

[0047] Performing principal component extraction according to the first covariance matrix and the second covariance matrix to obtain first principal component data and second principal component data;

[0048] Establishing prediction models using the first principal component data and the second principal component data respectively to obtain a hyperspectral prediction model and an electronic nose prediction model;

[0049] The hyperspectral prediction model and the electronic nose prediction model are used to perform score prediction respectively to obtain the matsutake quality hyperspectral score and the matsutake quality electronic nose score.

[0050] Preferably, the process of obtaining the overall score of the Tricholoma matsutake quality includes:

[0051] The impedance score of the Matsutake quality, the hyperspectral score of the Matsutake quality and the electronic nose score of the Matsutake quality were standardized to obtain the impedance standardized score, the hyperspectral standardized score and the electronic nose standardized score; the calculation formulas of the impedance standardized score, the hyperspectral standardized score and the electronic nose standardized score are as follows: Where, e = 1, 2, 3; Q 1 Q is the impedance score of the Tricholoma matsutake quality; 2 The high-spectral score of the quality of the matsutake mushroom; Q 3 The electronic nose scores the quality of the matsutake mushroom; Q 1norm The impedance is standardized to score; Q 1max and Q 1min are the maximum and minimum values ​​of the Matsutake quality impedance score, respectively;

[0052] Determine the determination weights of the impedance standardization score, the hyperspectral standardization score, and the electronic nose standardization score to obtain determination weight data; the determination weight data satisfies the condition: μ 1 +μ 2 +μ 3 =1; where μ 1 , μ 2 and μ 3 The determination rights of the impedance normalized score, the hyperspectral normalized score and the electronic nose normalized score respectively;

[0053] The score is calculated according to the determined weight data to obtain the overall quality score of the Tricholoma matsutake. The calculation formula of the overall quality score of the Tricholoma matsutake is: Q = μ 1 Q 1 +μ 2 Q 2 +μ 3 Q 3 ; Wherein, Q is the overall quality score of the Matsutake.

[0054] The present invention discloses the following technical effects:

[0055] The present invention provides a multi-dimensional matsutake quality grading detection device. Through the special structure of an impedance detection sub-device, the problem of excessive collection of irrelevant information in the data collection process is solved, and the simultaneous detection of surface size and impedance is realized; through the impedance detection sub-device and the back-end circuit, the lossy defects of conventional technologies are solved, and lossless collection of matsutake freshness data is realized; through the impedance detection sub-device, the hyperspectral detection sub-device and the electronic nose detection sub-device, the shortcoming of the conventional technology that the detection direction is single is solved, and the multi-dimensional detection of matsutake is realized; through the data processing station, the defect that the conventional technology can only detect once is solved, and the monitoring of the whole process of matsutake from freshness to corruption is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0057] Figure 1 A schematic diagram of the structure of a multi-dimensional Matsutake quality grading detection device provided in an embodiment of the present invention;

[0058] Figure 2 A structural diagram of a matsutake detection sub-device provided in an embodiment of the present invention;

[0059] Figure 3 A diagram of a local device architecture of a detection electrode sheet provided in an embodiment of the present invention;

[0060] Figure 4 A comparison of the maximum and minimum detection dimensions of the impedance detection sub-device provided in an embodiment of the present invention and a schematic diagram of detection of irregular-shaped matsutake mushrooms;

[0061] Figure 5 A schematic diagram of the back-end circuit structure provided by an embodiment of the present invention;

[0062] Figure 6 This is a flowchart of the operation of the multi-dimensional comprehensive quality scoring system provided by an embodiment of the present invention.

[0063] Description of reference numerals:

[0064] 1-impedance detection sub-device, 11-outer ring, 12-detection electrode sheet, 121-rotating disk, 122-rectangular electrode sheet, 123-reserved hole, 124-fixed point, 13-spring, 14-pressure sensor output wire, 15-first electrode connector wire, 16-second electrode connector wire, 17-pressure sensor patch, 2-back-end circuit, 3-receiving wire, 4-excitation wire, 5-hyperspectral detection sub-device, 6-electronic nose detection sub-device, 7-data processing station. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] The purpose of the present invention is to provide a multi-dimensional matsutake quality grading detection device, which realizes the clamping and surface data collection of the detection target matsutake through an impedance detection sub-device, and uses a spring and a rotatable detection electrode to improve the ductility and stability of the impedance detection sub-device; realizes the simultaneous detection of the shape and impedance of matsutake through a pressure sensor patch and a detection electrode, thereby improving the detection efficiency; realizes multi-dimensional non-destructive detection of the shape, smell and composition of matsutake by combining hyperspectral and electronic nose detection; and monitors the entire process of matsutake from freshness to corruption in real time through a data processing station.

[0067] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0068] Figure 1 This is a schematic diagram of the structure of a multi-dimensional Matsutake quality grading detection device provided by an embodiment of the present invention. Figure 2 This is a structural diagram of a matsutake detection sub-device provided in an embodiment of the present invention. Figure 3 The local device architecture diagram of the detection electrode sheet 12 provided in the embodiment of the present invention is as follows: Figure 1 , Figure 2 as well as Figure 3 As shown, the present invention provides a multi-dimensional matsutake quality grading detection device, including: an impedance detection sub-device 1, a hyperspectral detection sub-device 5, an electronic nose detection sub-device 6 and a data processing station 7; the impedance detection sub-device 1 includes: a back-end circuit 2, a receiving wire 3, an excitation wire 4 and a matsutake information acquisition module; the matsutake information acquisition module includes: an outer ring 11, a rotatable detection electrode sheet 12, a spring 13, a pressure sensor output wire 14, a first electrode connector wire 15, a second electrode connector wire 16 and a pressure sensing patch 17; the detection electrode sheet 12 includes: a rotating disk 121, a rectangular electrode sheet 122, a reserved hole 123 and a fixed point 124;

[0069] The data processing station 7 is connected to the impedance detection sub-device 1, the hyperspectral detection sub-device 5 and the electronic nose detection sub-device 6 respectively; the back-end circuit 2 is connected to the second electrode connector wire 16 through the excitation wire 4; the first electrode connector wire 15 and the pressure sensor output wire 14 are connected to the back-end circuit 2 through a receiving wire 3 respectively; the back-end circuit 2 is connected to the data processing station 7; 6 pressure sensing patches 17 are evenly fixed on the outer wall of the outer ring 11; a spring 13 is fixed on the inner side of each pressure sensing patch 17; a rotating disk 12 is fixed on the end of each spring 13 away from the outer ring 11 1; the rotating disk 121 is attached and fixed on the surface of the rectangular electrode sheet 122; a reserved hole 123 is respectively provided at both ends of the rectangular electrode sheet 122; each reserved hole 123 is provided with a fixed point 124; adjacent detection electrode sheets 12 are connected through the first electrode connector wire 15 and the second electrode connector wire 16; the first electrode connector wire 15 and the second electrode connector wire 16 respectively pass through the two reserved holes 123 of each detection electrode sheet 12, and are fixed on the fixed points 124 corresponding to the reserved holes 123; the pressure sensor output wire 14 is connected to each pressure sensing patch 17;

[0070] The second electrode connector wire 16 is used to transmit the collection excitation output by the back-end circuit 2 to the detection electrode sheet 12; the detection electrode sheet 12 is used to apply the collection excitation to the surface of the target detection matsutake; the first electrode connector wire 15 is used to transmit the electrical feedback information of the detection electrode sheet 12 to the back-end circuit 2 through the receiving wire 3; the pressure sensor output wire 14 is used to transmit the mechanical feedback information output by the pressure sensor patch 17 to the back-end circuit 2 through the receiving wire 3; the back-end circuit 2 is used to transmit the collection excitation to the second electrode connector wire 16 through the excitation wire 4, receive and pre-process the electrical feedback information and the mechanical feedback information, obtain the collection data, and transmit the collection data to the data processing station 7; the hyperspectral detection sub-device 5 is used to collect the hyperspectral data of the target detection matsutake, and transmit the hyperspectral data to the data processing station 7; the electronic nose detection sub-device 6 is used to collect the odor information of the target detection matsutake, and transmit the odor information to the data processing station 7; the data processing station 7 is used to perform multi-dimensional freshness detection and real-time monitoring of the target detection matsutake according to the collection data, hyperspectral data and odor information.

[0071] Preferably, the back-end circuit 2 is embedded with: a microprocessor unit, a constant current source signal generating unit, a front-end measuring unit, a response signal acquisition and processing unit, and an RS232 interface circuit unit;

[0072] The constant current source signal generating unit is used to generate acquisition excitation; the front-end measurement unit is used to amplify, filter and suppress noise of electrical feedback information and mechanical feedback information to obtain input signals; the response signal acquisition and processing unit is used to perform data analysis on the input signals to obtain acquisition data; the RS232 interface circuit unit is used to send the acquisition data to the data processing station 7; the microprocessor unit is used to control the constant current source signal to generate acquisition excitation and control the RS232 interface circuit unit to send the acquisition data to the data processing station 7.

[0073] Preferably, the data processing station 7 has a built-in multi-dimensional comprehensive quality scoring system; the multi-dimensional comprehensive quality scoring system includes: a first quality evaluation subsystem, a second quality evaluation subsystem, a third quality evaluation subsystem and a multi-dimensional comprehensive quality scoring subsystem; the first quality evaluation subsystem includes: a Matsutake cross-section detection model, a Matsutake column uniformity detection model, a Matsutake freshness detection model and an impedance-based quality parameter calculation model;

[0074] The Matsutake cross-section detection model is used to calculate the cross-sectional area of ​​the target Matsutake using the pressure sensor patch 17; the Matsutake column uniformity detection model is used to calculate the Matsutake column uniformity data of the target Matsutake using the pressure sensor patch 17; the Matsutake freshness detection model is used to calculate the initial judgment data of the freshness of the target Matsutake using the detection electrode sheet 12; the quality parameter calculation model is used to fuse the Matsutake cross-sectional area, the Matsutake column uniformity data and the initial freshness judgment data to obtain the Matsutake quality impedance score; the second quality evaluation subsystem is used to obtain the Matsutake quality hyperspectral score using the hyperspectral detection sub-device 5; the third quality evaluation subsystem is used to obtain the Matsutake quality electronic nose score using the electronic nose detection sub-device 6; the multi-dimensional comprehensive quality scoring subsystem is used to fuse the Matsutake quality impedance score, the Matsutake quality hyperspectral score and the Matsutake quality electronic nose score to obtain the overall Matsutake quality score.

[0075] Preferably, the calculation process of the cross-sectional area of ​​​​Matsutake and the uniformity data of Matsutake columns includes:

[0076] The pressure sensing patch 17 is used to collect pressure data of the target detection matsutake mushroom to obtain six-position pressure values;

[0077] The radius is calculated by using the six-position pressure values ​​and the probe design parameters of the pressure sensing patch 17 to obtain the six-position Matsutake radius data;

[0078] The polygon interpolation method was used to fit the six-dimensional matsutake radius data to obtain the cross-sectional area of ​​matsutake.

[0079] The standard deviation of the six-dimensional Matsutake radius data was calculated and normalized to obtain the Matsutake column uniformity data.

[0080] Preferably, the calculation process of the initial freshness judgment data includes:

[0081] The detection electrode sheet 12 is used to collect electrical data of the target detected matsutake mushroom to obtain electrical feedback information;

[0082] The real part orthogonal anisotropy characteristic parameter and the imaginary part orthogonal anisotropy characteristic parameter of the electrical feedback information are calculated by using the difference method to obtain the orthogonal anisotropy characteristic vector;

[0083] Extract the angle value and weight value of the orthogonal anisotropic eigenvector;

[0084] Construct a time series set, a characteristic signal sample matrix and a mapping function based on the angle value and the weight value;

[0085] The morphological characteristic parameters and integral values ​​are calculated using the angle value and the weight value;

[0086] Based on the time series set, characteristic signal sample matrix and mapping function, freshness judgment is performed using morphological characteristic parameters and integral values ​​to obtain initial freshness judgment data.

[0087] Preferably, the weight value calculation process includes:

[0088] Put the orthogonal anisotropic eigenvector into the vector diagram to obtain the vector coordinates;

[0089] Calculate the angle between the vector coordinates and the 45° standard line to obtain the angle data;

[0090] The angle data is used to calculate the weight to obtain the weight value; the calculation formula of the weight value is: Where, q = 1, 2, 3; is the weight value corresponding to the qth orthogonal anisotropic eigenvector; θ q is the angle data corresponding to the qth orthogonal anisotropic eigenvector; θ 1 ,θ 2 and θ 3 They are the angle data corresponding to the 1st, 2nd and 3rd orthogonal anisotropic eigenvectors respectively.

[0091] Preferably, the process of constructing the time series set includes:

[0092] The impedance amplitude samples are obtained based on the orthogonal anisotropic eigenvectors. The calculation formula for obtaining the impedance amplitude samples is:

[0093]

[0094] in, is the impedance amplitude sample; R d is the dimension; C0 is a set of time series; → represents a mapping; is the weight value at the i-th time point; x * i and * i are the impedance value and phase value collected at the i-th time point respectively; X and Y are the impedance data set and phase data set respectively; t 0 and t 1 are the start time and end time of the time series respectively;

[0095] Phase samples are obtained based on orthogonal anisotropic eigenvectors; the calculation formula for phase samples is:

[0096]

[0097] in, is the phase sample; x * ,y * They are the data points corresponding to the impedance value and phase value respectively; and are the nth ZI and ZA phase points collected respectively; f(y) is the phase transformation function; and is the mth ZI and ZA phase point after f(y) transformation;

[0098] Data mapping is performed on the impedance amplitude samples and phase samples to obtain a time series set; the calculation formula for the time series set is:

[0099] P: k cj ,k c+1·j ,k c·c+1 ,…k c+n·j ,k c·j+n}∈P M d ;

[0100] Among them, P M d is a set of time series in a feature space with dimension d greater than 10; P is a set of time series; k cj is the element at (c, j) in the matrix corresponding to the time series set.

[0101] Preferably, the calculation process of the Matsutake quality impedance score includes:

[0102] The cross-sectional area of ​​pine mushrooms, the uniformity data of pine mushroom cylinders and the initial freshness judgment data were standardized to obtain standard cross-sectional area data, standard uniformity data and standard freshness data; the calculation formulas for the standard cross-sectional area data, standard uniformity data and standard freshness data were: u norm =u and Among them, A norm 、u norm and D norm are the standard cross-sectional area data, standard uniformity data and standard freshness data respectively; A, u and D are the cross-sectional area of ​​​​Matsutake mushroom, the uniformity data of ​​Matsutake mushroom column and the initial judgment data of freshness respectively; A max and A min are the maximum and minimum cross-sectional areas of Matsutake, respectively; D max and D min They are the maximum and minimum values ​​of the initial freshness judgment data respectively;

[0103] The importance weights of the standard cross-sectional area data, the standard uniformity data, and the standard freshness data are assigned to obtain the importance data; the importance data satisfies the condition: w A +w u +w D =1; where w A 、w u and w D are the importance weights of standard cross-sectional area data, standard uniformity data, and standard freshness data respectively;

[0104] Importance data were used to perform weighted summation on standard cross-sectional area data, standard uniformity data, and standard freshness data to obtain the Matsutake quality impedance score.

[0105] Preferably, the process of obtaining the high-spectral score of Matsutake quality and the electronic nose score of Matsutake quality includes:

[0106] The hyperspectral detection sub-device 5 is used to collect optical data of the target detection matsutake mushroom to obtain hyperspectral data;

[0107] The electronic nose detection sub-device 6 is used to collect chemical data of the target detection matsutake mushroom to obtain odor information;

[0108] Calculate the covariance of the hyperspectral data and the odor information respectively to obtain a first covariance matrix and a second covariance matrix;

[0109] Perform principal component extraction according to the first covariance matrix and the second covariance matrix to obtain first principal component data and second principal component data;

[0110] The prediction models are established by using the first principal component data and the second principal component data respectively, so as to obtain a hyperspectral prediction model and an electronic nose prediction model;

[0111] The hyperspectral prediction model and electronic nose prediction model were used for score prediction, respectively, to obtain the hyperspectral score of Matsutake quality and the electronic nose score of Matsutake quality.

[0112] Preferably, the process of obtaining the overall score of the quality of Matsutake mushrooms includes:

[0113] The impedance score, hyperspectral score and electronic nose score of Matsutake quality were standardized to obtain the impedance standardized score, hyperspectral standardized score and electronic nose standardized score. The calculation formulas for the impedance standardized score, hyperspectral standardized score and electronic nose standardized score are as follows: Where, e = 1, 2, 3; Q 1 Q is the impedance score of Matsutake quality; 2 High-spectrum scoring of Matsutake quality; Q 3 Electronic nose scoring of Matsutake quality; Q 1norm The impedance normalization score is Q 1max and Q 1min are the maximum and minimum values ​​of the Matsutake quality impedance score, respectively;

[0114] Determine the determination weights of impedance standardization score, hyperspectral standardization score and electronic nose standardization score to obtain determination weight data; the determination weight data satisfies the condition: μ 1 +μ 2 +μ 3 =1; where μ 1 , μ 2 and μ 3 They are the determination weights of impedance normalized score, hyperspectral normalized score, and electronic nose normalized score;

[0115] The score is calculated based on the determined weight data to obtain the overall quality score of Matsutake. The calculation formula for the overall quality score of Matsutake is: Q = μ 1 Q 1 +μ 2 Q 2 +μ 3 Q 3 ; Among them, Q is the overall quality score of Matsutake.

[0116] refer to Figure 2The detection electrode head is connected to each adjacent electrode head through the first electrode connector wire 15 and the second electrode connector wire 16 to form an electrode array; the first electrode connector wire 15 and the second electrode connector wire 16 are both arranged at the junction of the head and the tail of the electrode array. The detection electrode sheet 12 and the spring 13 are connected to form a detection electrode head through a rotating disk 121. The uniformly surrounded group of detection electrode heads constitutes an electrode array. The spring 13 is used to provide elasticity for the electrode array. The left end of the rotating disk 121 is fixedly connected to the spring 13, and the right end is connected to the detection electrode sheet 12 so that it can rotate relatively. The rear end of the spring 13 contacts the pressure sensing patch 17, so that the pressure generated by the spring 13 each time it contracts can be detected by the pressure sensing patch 17.

[0117] Furthermore, the multi-dimensional comprehensive quality scoring system includes: an impedance-based quality evaluation subsystem (first quality evaluation subsystem), a hyperspectral-based quality evaluation subsystem (second quality evaluation subsystem), an electronic nose-based quality evaluation subsystem (third quality evaluation subsystem) and a multi-dimensional comprehensive quality scoring subsystem. The impedance-based quality evaluation subsystem includes: a matsutake cross-section detection model, a matsutake column uniformity detection model, a matsutake freshness detection model and an impedance-based quality parameter calculation model; the input information of the matsutake cross-section detection model and the matsutake uniformity detection model is obtained by the pressure sensor patch 17; the matsutake freshness detection model is obtained by the detection electrode sheet 12; the hyperspectral-based quality evaluation subsystem and the electronic nose-based quality evaluation subsystem are obtained by the hyperspectral detection device and the electronic nose detection device, respectively.

[0118] refer to Figure 2, detection electrode heads are evenly attached to the inside of the outer ring 11, the upper end of the detection electrode sheet 12 is connected by the first electrode head wire, and the lower end is connected by the second electrode connecting head wire 16, and adjacent detection electrode sheets 12 are connected in series. Pressure sensing patches 17 are evenly attached to the outside of the outer ring 11, and adjacent pressure sensing patches 17 are connected in series by pressure sensor output wires 14, and the series wires have sufficient margins to meet the length changes caused by the deformation of the spring 13. The spring 13 is connected between the outer ring 11 and the detection electrode sheet 12. When detecting a product, the spring 13 contracts, and the detection electrode sheet 12 is attached to the detected pine mushroom. The second electrode connecting head wire 16 is passed through After the current applied by the excitation wire 4 releases the excitation signal, it is transmitted back to the back-end circuit 2 through the first electrode head wire through the receiving wire 3, and the impedance information of the detected matsutake is obtained after processing, and then input to the data processing station 7. According to the preset model, the freshness of the matsutake can be detected. At the same time, the pressure generated by the contraction of the spring 13 is equally applied to the pressure sensing patch 17. After processing, the pressure information of the detected matsutake is obtained. The pressure sensing patch 17 is transmitted back to the back-end circuit 2 through the pressure sensor output wire 14. After processing, the pressure information of the detected matsutake is obtained, and then input to the data processing station 7. According to the preset model, the area of ​​the cross-section of the matsutake and the uniformity of the matsutake column can be detected. By moving the detected object, the detection information of multiple cross sections can be obtained, so as to more specifically judge the state information of the detected matsutake. It can ensure that it is in the best working condition in real time. This unique design combines high efficiency and convenience, so that the detection sub-device has better performance in practical applications.

[0119] refer to Figure 3 The left end of the rotating disk 121 is fixedly connected to the spring 13, and the right end is connected to the electrode sheet so that it can rotate relatively. The electrode sheet is provided with reserved holes 123 at upper and lower symmetrical positions so that the wires can connect the electrode sheets in series. A fixed point 124 is set next to the reserved hole 123, and the wire is fixed to the point by insulating glue, which is conducive to the uniform distribution of current in the object to be measured to avoid unnecessary disturbance.

[0120] refer to Figure 4, the left side is the minimum size that the detection sub-device can measure when the spring 13 is in a relaxed state, the right side is the maximum size that the detection sub-device can measure when the spring 13 is in a contracted state, and the lower side is a schematic diagram of the detection of irregularly shaped crops (Matsutake). From the top views under the two extreme states, it can be obtained that the detection sub-device can be applied to target detection crops of different sizes. The difference in the measurable sizes of the detection sub-device under the two extreme states is large, that is, the detection sub-device has a large measurement range. In order to ensure the stability of the installation position of the measuring device and maximize the contact area of ​​the detection, the detection sub-device can be installed at the cross-sectional position where the target detection Matsutake has the largest size. Such an installation method can not only improve the measurement accuracy of the detection sub-device, but also enhance the overall stability, so as to ensure the accuracy and reliability of the measurement results. The probe of the detection sub-device includes 6 detection electrode sheets 12 to form an electrode group. Due to the symmetry of the detection sub-device and the distribution angle of the electrode sheets is 60°, the complex impedance value obtained can be defined as Zvm, and m is 1 to 6; since the detection heads of the designed group of detection electrodes are independent of each other, they can show different compression amounts for the unevenness on the surface of the detected object. The detection electrode sheets 12 are installed through a rotating disk 121, so the detection electrode sheets 12 can fix the detected object at different angles, which can meet the detection of irregularly shaped matsutake mushrooms.

[0121] refer to Figure 5 , the back-end circuit 2 includes: a microprocessor unit, a constant current source signal generating unit, a front-end measurement unit, a response signal acquisition and processing unit and an RS232 interface circuit unit. In the constant current source signal generating unit, the two differential excitation signals output by the constant current source enter the internal measured resistance network or the external measured object through the signal switching module. The collected response voltage signal processing flow is that after being amplified by the PGA (program global area), it enters the ADC (analog-to-electric converter) and is converted into a digital signal and stored in the SRAM (static random access memory). After the acquisition is completed, the microcontroller uniformly reads the voltage and current signal data stored in the SRAM for preliminary processing, and the processing results are transmitted to the host computer (data processing station 7) through the RS232 communication module. Thus, a complete and high-performance bioimpedance measurement process is carried out.

[0122] refer to Figure 6 The calculation process of the cross-sectional area of ​​​​Matsutake mushrooms and the uniformity of the ​​Matsutake mushroom column includes:

[0123] 1) Pressure distribution measurement: The pressure value P recorded by each detection electrode sheet 12 1 , P 2 ,…,P 6 ,The pressure value reflects the shape information of Matsutake at different positions.

[0124] 2) Determination of electrode position: Since the electrodes of the probe are evenly distributed on a circular outer ring, the electrode position can be expressed as an angle θ in the polar coordinate system. α and distance r (i.e., the radius of the outer ring), where α=1, 2,…, 6.

[0125] 3) Calculate the relationship between pressure and radius: The pressure P measured by each electrode m It can be compared with the "radius" R of the matsutake in this direction. m The compression amount d of the spring 13 is related to m With the known probe design parameters, the radius R of the matsutake mushroom at each electrode position can be calculated. i Usually, the relationship between pressure and radius can be obtained by the spring force formula:

[0126] Where r is the initial radius of the probe, k is the spring constant, and d m is the compression amount of spring 13.

[0127] 4) Cross-sectional area calculation: With the "radius" R of each electrode position 1 , R 2 , …, R 6 After that, these points can be used to reconstruct the cross-sectional profile of the matsutake. Using polygon interpolation, a smooth polygon is fitted to represent the cross-sectional boundary. The area A of the polygon is calculated as the cross-sectional area of ​​the matsutake:

[0128]

[0129] Here (x i ,y i ) is the point converted to Cartesian coordinates, n is the number of electrodes, (x n+1 ,y n+1 )=(x i ,y i ).

[0130] 5) Uniformity calculation: Based on the calculated radius R 1 , R 2 , …, R 6 The average Then calculate the standard deviation σ R , standard deviation σ R The smaller the size, the more uniform the Matsutake column. The calculation formula is as follows:

[0131]

[0132] The standard deviation σ RAfter normalization, it is converted into a score between 0 and 1 to indicate uniformity, where 0 indicates completely uneven and 1 indicates completely uniform.

[0133] The calculation formula for uniformity is as follows:

[0134]

[0135] Among them, σ max is the maximum allowed standard deviation set in the experiment.

[0136] Furthermore, the method for detecting the freshness of Tricholoma matsutake comprises:

[0137] 1) Since the object being tested is non-uniform, it is necessary to calculate the weights of the three orthogonal electrode groups to make the measurement results more reliable. The difference method is used to calculate the real part orthogonal anisotropy characteristic parameters ROACPS at a frequency of 2000 Hz. o and imaginary orthogonal anisotropy characteristic parameter IOACPS p :

[0138] ROACPSo: ROACPSo=2|(Zvm-Zvs) / (Zvm+Zvs)|

[0139] IOACPSp: IOACPSp=2|(Zvm,-Zvs) / (Zvm+Zvs)|

[0140] Zvm and Zvs are the impedance values ​​of two different electrodes. To ensure orthogonal characteristics, select opposite electrodes, such as Z1 and Z4, Z2 and Z5, and Z3 and Z6;

[0141] The three sets of orthogonal anisotropic characteristic parameters are input into the vector diagram to obtain their vector R q , and set the initial position of its vector at the dot, we can get the angle θ between the vector and the 45° standard line q , the weights of the orthotropic characteristic parameters can be calculated: is the weight value of the orthotropic characteristic parameter; q=1, 2, 3.

[0142] 2) Construct R d The time series set C of the impedance characteristic data 0 , which represents the impedance value of the fruit at different frequencies or time points;

[0143] Construct the characteristic signal sample matrix S and its mapping functions f(x) and f(y), which can represent the relationship between the fruit impedance characteristic data and its freshness;

[0144] Impedance magnitude sample: These samples represent the impedance magnitude data collected from time t0 to t1. Each sample is d In the dimensional space, the amplitude characteristic C 0 composition;

[0145] Phase Sample: These samples contain the time from t0 to t 1 The impedance phase data for the period is shown in Figure 1. Each sample is a tuple (x, y). Each tuple (x, y) represents an impedance phase data point in a given time period, where x is the input feature (time, frequency, etc.) and y is the corresponding measured output (phase value). The set of these tuples is obtained from t 0 to 1 All impedance phase changes are covered. x represents the input measurement data point. In impedance phase detection, it is a 0 to 1 The physical quantity related to the impedance data points between y and x. y represents the output value corresponding to the input data x, is the instance phase value, and represents t 0 to 1 The phase change characteristics between them.

[0146] Data Processing: {k cj ,k c+1·j ,k c·c+1 ,…k c+n·j ,k c·j+n}∈P M d ; Through data processing P, the impedance amplitude and phase data are mapped to and In the d-dimensional space, the morphological feature parameter k is obtained, and the parameter k is a d×d dimensional matrix.

[0147] 3) Based on the amplitude and phase angle features extracted from the impedance amplitude and phase data, the extracted amplitude and phase angle features are standardized so that the data are on the same scale for subsequent analysis; the amplitude features and phase angle features are fused to form a comprehensive feature vector, and each statistical feature value is combined into a vector to obtain a morphological feature deformation parameter, which is recorded as "K": L n Indicates the nth freshness level; n is the number of freshness levels; The impedance amplitude sample and the phase sample are then integrated to further represent the freshness level of the detected object, which is recorded as "L": L = {L 1 ,L 2 ,…L n}, In the time period t 0 The integration result of the impedance amplitude samples within T; In the time period t 0 The integration result of the impedance phase sample in T; 1 and λ 2 are weight coefficients, which are used to adjust the influence of impedance amplitude and phase in calculating freshness level. Based on the above steps, the initial detection model is obtained.

[0148] 4) Input the overall integrated signal into the initial detection model to verify the feasibility of the model, and operate according to the freshness feature recognition and division rules. This process can realize the calculation of the freshness level of the detected object, where "L n " indicates the key stress monitoring parameter signal, and "D" indicates the final freshness level of the detected object:

[0149] Specifically, the step of fusing the data of the cross-sectional area of ​​pine mushroom, the uniformity of the pine mushroom column and the freshness of pine mushroom includes:

[0150] 1) The three indicators of cross-sectional area, column uniformity and freshness of Matsutake are standardized so that they can be easily compared and integrated on the same scale. The standardization formula is as follows:

[0151]

[0152] u norm =u

[0153]

[0154] 2) Assign weights to each indicator based on its importance in practical applications. Assume that the weights of cross-sectional area, uniformity, and freshness are w A 、w u 、w D These weights can be set based on experimental experience or the advice of domain experts and meet the following requirements:

[0155] w A +w u +w D =1

[0156] The standardized indicators are weighted and summed according to the weights to obtain the comprehensive score Q based on impedance. 1 :

[0157] Q 1 =w A ×A norm +w u ×u norm +w D ×D norm

[0158] Furthermore, the calculation steps of the quality assessment subsystem based on hyperspectral detection include:

[0159] 1) First, calculate the covariance matrix C of the hyperspectral data H , used to analyze the correlation between each band. The covariance matrix formula is as follows:

[0160]

[0161] Among them, X hi is the spectral reflectance of the ith sample, is the mean vector of the samples, and N is the number of samples. H Perform eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors. H ·v hi =λ hi ·v hi , where v hi is the eigenvector of the covariance matrix, λ hi The eigenvalue eigenvector represents the main direction of the data, and the eigenvalue represents the variance information contained in these directions. Several principal components with larger eigenvalues ​​are selected (usually the principal components that can explain more than 90% of the variance are selected), and the high-dimensional original data is projected into the new space constructed by these principal components, thereby achieving dimensionality reduction: Y h =X h ·P h , where Y h is the data matrix after dimensionality reduction, X h is the original hyperspectral data matrix, P h is the eigenvector matrix.

[0162] 2) Using the extracted principal component data, a prediction model (such as a regression model or a classification model) can be further established to predict the specific quality indicators of Matsutake. For example, the moisture content and degree of decay of Matsutake can be predicted by a linear regression model. 2 :

[0163] Q 2 =β 0 +β 1 ·P h C h1 +β 2 ·P h C h2 +…+β n ·P h C hn

[0164] Among them, Q 2 is a comprehensive score based on the hyperspectral spectrum, Ph C h1 , P h C h2 …, P h C hn is the principal component, β 0 , β 1 …,β n is the regression coefficient.

[0165] Specifically, the calculation steps of the quality assessment subsystem based on electronic nose detection are the same as those of the quality assessment subsystem based on hyperspectral detection, including:

[0166] 1) Use the electronic nose device to collect the smell data of Matsutake to form a data matrix X. Use the principal component analysis (PAC) method to establish a regression model between the electronic nose data and the quality of Matsutake. Calculate the covariance matrix C of the collected data E :

[0167]

[0168] Perform eigenvalue decomposition on the covariance matrix and obtain the eigenvalue λ e and the eigenvector V e .

[0169] 2) According to the size of the eigenvalue, select the first k principal components P e C e1 , P e C e2 ,…,P e C ek , these principal components can retain most of the variance information. Based on the scores and weights of the principal components, the comprehensive score Q based on the electronic nose is calculated 3 :

[0170] Q 3 =ω 0 +ω 1 ·P e C e1 +ω 2 ·P e C e2 +…+ω n ·P e C ek

[0171] Among them, Q 3 is a comprehensive score based on the electronic nose, ω i It is the weight of the corresponding principal component, which can be determined according to the variance contribution rate of each principal component.

[0172] Furthermore, the calculation steps of the multi-dimensional comprehensive quality scoring subsystem include:

[0173] 1) First, standardize the comprehensive score Q based on impedance obtained above 1 The comprehensive score Q based on hyperspectral 2 The comprehensive score Q based on electronic nose 3 as follows: where e = 1, 2, 3, representing the comprehensive scores based on impedance, hyperspectral, and electronic nose respectively.

[0174] 2) Determine the weight μ for each score according to the importance or relevance of different detection methods 1 , μ 2 , μ 3 . These weights can be determined through expert evaluation, experimental results, or accuracy indicators of each method, ensuring that their sum is 1: μ 1 + μ 2 + μ 3 = 1. Calculate the comprehensive quality score Q according to the set weights:

[0175] Q = μ 1 ·Q 1 + μ 2 ·Q 2 + μ 3 ·Q 3

[0176] Through this comprehensive evaluation method, the overall quality of matsutake can be effectively evaluated, providing a scientific basis for subsequent storage, sales, or processing.

[0177] Preferably, according to the value of the comprehensive score Q, the quality of matsutake can be divided into different grades. Thresholds can be set based on experimental data or industry standards to divide the quality grades. For example: excellent quality: Q > 0.8; good quality: 0.6 < Q ≤ 0.8; average quality: 0.4 < Q ≤ 0.6; poor quality: Q ≤ 0.4. Through the data fusion method, the three key indicators of the cross-sectional area, column uniformity, and freshness of matsutake are integrated into a comprehensive quality score Q. This score can provide a scientific basis for grading the quality of matsutake and can reflect the overall quality of matsutake. The larger the cross-sectional area, the higher the uniformity, and the fresher the matsutake, the higher the comprehensive score Q and the better the quality.

[0178] The beneficial effects of the present invention are as follows:

[0179] The present invention realizes the clamping and surface data collection of the detection target matsutake through an impedance detection sub-device, and improves the ductility and stability of the impedance detection sub-device by using a spring and a rotatable detection electrode sheet; the pressure sensing patch and the detection electrode sheet can realize the simultaneous detection of the shape and impedance of matsutake, thereby improving the detection efficiency; the combination of hyperspectral and electronic nose detection can realize multi-dimensional non-destructive detection of the shape, smell and composition of matsutake; and the data processing station can monitor the whole process of matsutake from freshness to corruption in real time.

[0180] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0181] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A multi-dimensional matsutake quality grading detection device, characterized in that: include: Impedance detection sub-device, hyperspectral detection sub-device, electronic nose detection sub-device and data processing station; The impedance detection sub-device includes: a back-end circuit, a receiving wire, an excitation wire and a matsutake information acquisition module; the matsutake information acquisition module includes: an outer ring, a rotatable detection electrode sheet, a spring, a pressure sensor output wire, a first electrode connector wire, a second electrode connector wire and a pressure sensor patch; the detection electrode sheet includes: a rotating disk, a rectangular electrode sheet, a reserved hole and a fixed point; The data processing station is connected to the impedance detection sub-device, the hyperspectral detection sub-device and the electronic nose detection sub-device respectively; the back-end circuit is connected to the second electrode connector wire through the excitation wire; the first electrode connector wire and the pressure sensor output wire are respectively connected to the back-end circuit through a receiving wire; the back-end circuit is connected to the data processing station; the six pressure sensing patches are evenly fixed on the outer wall of the outer ring; a spring is fixed on the inner side of each pressure sensing patch; a rotating disk is fixed on the end of each spring away from the outer ring; the rotating disk is attached and fixed to the surface of the rectangular electrode sheet; a reserved hole is respectively provided at both ends of the rectangular electrode sheet; each of the reserved holes is provided with a fixed point; adjacent detection electrode sheets are connected through the first electrode connector wire and the second electrode connector wire; the first electrode connector wire and the second electrode connector wire pass through the two reserved holes of each detection electrode sheet respectively, and are fixed at the fixed point corresponding to the reserved hole; the pressure sensor output wire is connected to each pressure sensing patch; The second electrode connector wire is used to transmit the acquisition excitation output by the back-end circuit to the detection electrode sheet; the detection electrode sheet is used to apply the acquisition excitation to the surface of the target detection matsutake; the first electrode connector wire is used to transmit the electrical feedback information of the detection electrode sheet to the back-end circuit through the receiving wire; the pressure sensor output wire is used to transmit the mechanical feedback information output by the pressure sensor patch to the back-end circuit through the receiving wire; the back-end circuit is used to transmit the acquisition excitation to the second electrode connector wire through the excitation wire, receive and pre-process the electrical feedback information and the mechanical feedback information, obtain acquisition data, and transmit the acquisition data to the data processing station; the hyperspectral detection sub-device is used to collect the hyperspectral data of the target detection matsutake, and transmit the hyperspectral data to the data processing station; the electronic nose detection sub-device is used to collect the odor information of the target detection matsutake, and transmit the odor information to the data processing station; the data processing station is used to perform multi-dimensional freshness detection and real-time monitoring of the target detection matsutake according to the acquisition data, the hyperspectral data and the odor information.

2. A multi-dimensional matsutake quality grading detection device according to claim 1, characterized in that: The back-end circuit is embedded with: a microprocessor unit, a constant current source signal generating unit, a front-end measuring unit, a response signal collecting and processing unit and an RS232 interface circuit unit; The constant current source signal generating unit is used to generate the acquisition excitation; The front-end measurement unit is used to amplify, filter and suppress noise of the electrical feedback information and the mechanical feedback information to obtain an input signal; The response signal acquisition and processing unit is used to perform data analysis on the input signal to obtain the acquired data; the RS232 interface circuit unit is used to send the acquired data to the data processing station; the microprocessor unit is used to control the constant current source signal to generate the acquisition excitation and control the RS232 interface circuit unit to send the acquired data to the data processing station.

3. A multi-dimensional matsutake quality grading detection device according to claim 1, characterized in that: The data processing station is embedded with a multi-dimensional comprehensive quality scoring system; the multi-dimensional comprehensive quality scoring system includes: a first quality evaluation subsystem, a second quality evaluation subsystem, a third quality evaluation subsystem and a multi-dimensional comprehensive quality scoring subsystem; the first quality evaluation subsystem includes: a matsutake cross-section detection model, a matsutake column uniformity detection model, a matsutake freshness detection model and an impedance-based quality parameter calculation model; The Matsutake cross-section detection model is used to calculate the cross-sectional area of ​​the target detected Matsutake using the pressure sensing patch; the Matsutake column uniformity detection model is used to calculate the Matsutake column uniformity data of the target detected Matsutake using the pressure sensing patch; the Matsutake freshness detection model is used to calculate the initial freshness judgment data of the target detected Matsutake using the detection electrode sheet; the quality parameter calculation model is used to perform data fusion on the Matsutake cross-sectional area, the Matsutake column uniformity data and the initial freshness judgment data to obtain the Matsutake quality impedance score; the second quality evaluation subsystem is used to obtain the Matsutake quality hyperspectral score using the hyperspectral detection sub-device; the third quality evaluation subsystem is used to obtain the Matsutake quality electronic nose score using the electronic nose detection sub-device; the multi-dimensional comprehensive quality scoring subsystem is used to perform data fusion on the Matsutake quality impedance score, the Matsutake quality hyperspectral score and the Matsutake quality electronic nose score to obtain the overall Matsutake quality score.

4. A multi-dimensional matsutake quality grading detection device according to claim 3, characterized in that: The calculation process of the cross-sectional area of ​​the Tricholoma matsutake and the uniformity data of the Tricholoma matsutake column includes: Using the pressure sensing patch to collect pressure data of the target detection matsutake mushroom to obtain six-position pressure values; The radius is calculated by using the six-position pressure values ​​and the probe design parameters of the pressure sensor patch to obtain six-position matsutake radius data; The six-position Matsutake radius data are fitted using polygon interpolation to obtain the cross-sectional area of ​​the Matsutake; The standard deviation of the six-direction Matsutake radius data is calculated, and the standard deviation is normalized to obtain the Matsutake column uniformity data.

5. A multi-dimensional matsutake quality grading detection device according to claim 3, characterized in that: The calculation process of the initial freshness judgment data includes: Using the detection electrode sheet to collect electrical data of the target detection Tricholoma matsutake to obtain the electrical feedback information; The real part orthogonal anisotropy characteristic parameter and the imaginary part orthogonal anisotropy characteristic parameter of the electrical feedback information are calculated by using a difference method to obtain an orthogonal anisotropy characteristic vector; Extracting the angle value and weight value of the orthogonal anisotropic eigenvector; Constructing a time series set, a characteristic signal sample matrix and a mapping function based on the angle value and the weight value; Calculating morphological characteristic parameters and integral values ​​using the angle value and the weight value; Based on the time series set, the characteristic signal sample matrix and the mapping function, freshness judgment is performed using the morphological characteristic parameters and the integral value to obtain the initial freshness judgment data.

6. A multi-dimensional matsutake quality grading detection device according to claim 5, characterized in that: The calculation process of the weight value includes: Adding the orthogonal anisotropic eigenvector to the vector diagram to obtain vector coordinates; Calculate the angle between the vector coordinates and the 45° standard line to obtain angle data; The weight value is obtained by using the angle data to calculate the weight; the calculation formula of the weight value is: Where, q = 1, 2, 3; is the weight value corresponding to the qth orthogonal anisotropic eigenvector; θ q is the angle data corresponding to the qth orthogonal anisotropic eigenvector; θ1, θ2 and θ3 are the angle data corresponding to the 1st, 2nd and 3rd orthogonal anisotropic eigenvectors respectively.

7. A multi-dimensional matsutake quality grading detection device according to claim 5, characterized in that: The construction process of the time series set includes: Obtain impedance amplitude samples based on the orthogonal anisotropic eigenvector; the calculation formula for obtaining the impedance amplitude samples is: in, is the impedance amplitude sample; R d is the dimension; C0 is the time series set; → represents mapping; is the weight value at the i-th time point; x * i and * i are the impedance value and phase value collected at the i-th time point respectively; X and Y are the impedance data set and phase data set respectively; t0 and t1 are the start time and end time of the time series respectively; A phase sample is obtained based on the orthogonal anisotropic eigenvector; the calculation formula of the phase sample is: in, is the phase sample; x * ,y * are data points corresponding to the impedance value and the phase value respectively; and are the nth ZI and ZA phase points collected respectively; f(y) is the phase transformation function; and is the mth ZI and ZA phase point after f(y) transformation; Data mapping is performed on the impedance amplitude samples and the phase samples to obtain the time series set; the calculation formula of the time series set is: P： Among them, P M d is the time series set in the feature space with dimension d greater than 10; P is the time series set; k cj is the element located at (c, j) in the matrix corresponding to the time series set.

8. A multi-dimensional matsutake quality grading detection device according to claim 3, characterized in that: The calculation process of the Matsutake quality impedance score includes: The cross-sectional area of ​​the Tricholoma matsutake, the uniformity data of the Tricholoma matsutake column and the initial freshness judgment data are standardized to obtain standard cross-sectional area data, standard uniformity data and standard freshness data; the calculation formulas of the standard cross-sectional area data, the standard uniformity data and the standard freshness data are respectively: u norm =u and Among them, A norm 、u norm and D norm are the standard cross-sectional area data, the standard uniformity data and the standard freshness data respectively; A, u and D are the cross-sectional area of ​​the pine mushroom, the uniformity data of the pine mushroom column and the initial freshness judgment data respectively; A max and A min are the maximum and minimum cross-sectional areas of the Tricholoma matsutake, respectively; D max and D min are respectively the maximum and minimum values ​​of the freshness initial judgment data; The importance weights of the standard cross-sectional area data, the standard uniformity data and the standard freshness data are allocated to obtain importance data; the importance data satisfies the condition: w A +w u +w D =1; where w A 、w u and w D are the importance weights of the standard cross-sectional area data, the standard uniformity data, and the standard freshness data respectively; The importance data is used to perform weighted summation on the standard cross-sectional area data, the standard uniformity data, and the standard freshness data to obtain the Matsutake quality impedance score.

9. A multi-dimensional matsutake quality grading detection device according to claim 3, characterized in that: The process of obtaining the high-spectrum score of the Matsutake quality and the electronic nose score of the Matsutake quality includes: Using the hyperspectral detection sub-device to collect optical data of the target detection Tricholoma matsutake to obtain the hyperspectral data; Using the electronic nose detection sub-device to collect chemical data of the target detection matsutake mushroom to obtain the odor information; Calculating the covariance of the hyperspectral data and the odor information respectively to obtain a first covariance matrix and a second covariance matrix; Performing principal component extraction according to the first covariance matrix and the second covariance matrix to obtain first principal component data and second principal component data; Establishing prediction models using the first principal component data and the second principal component data respectively to obtain a hyperspectral prediction model and an electronic nose prediction model; The hyperspectral prediction model and the electronic nose prediction model are used to perform score prediction respectively to obtain the matsutake quality hyperspectral score and the matsutake quality electronic nose score.

10. A multi-dimensional matsutake quality grading detection device according to claim 3, characterized in that: The process of obtaining the overall score of the Matsutake quality includes: The impedance score of the Matsutake quality, the hyperspectral score of the Matsutake quality and the electronic nose score of the Matsutake quality were standardized to obtain the impedance standardized score, the hyperspectral standardized score and the electronic nose standardized score; the calculation formulas of the impedance standardized score, the hyperspectral standardized score and the electronic nose standardized score are as follows: Wherein, e=1, 2, 3; Q1 is the impedance score of the Matsutake quality; Q2 is the high-spectrum score of the Matsutake quality; Q3 is the electronic nose score of the Matsutake quality; Q 1norm The impedance is normalized to score; Q 1max and Q 1min are the maximum and minimum values ​​of the Matsutake quality impedance score, respectively; Determine the determination weights of the impedance standardized score, the hyperspectral standardized score, and the electronic nose standardized score to obtain determination weight data; the determination weight data satisfies the condition: μ1+μ2+μ3=1; wherein μ1, μ2, and μ3 are the determination weights of the impedance standardized score, the hyperspectral standardized score, and the electronic nose standardized score, respectively; The score is calculated according to the determined weight data to obtain the overall score of the Matsutake quality; the calculation formula for the overall score of the Matsutake quality is: Q=μ1·Q1+μ2·Q2+μ3·Q3; wherein Q is the overall score of the Matsutake quality.

Images