CN110411957B - Nondestructive rapid prediction method and device for shelf life and freshness of fruits - Google Patents
Nondestructive rapid prediction method and device for shelf life and freshness of fruits Download PDFInfo
- Publication number
- CN110411957B CN110411957B CN201910803691.1A CN201910803691A CN110411957B CN 110411957 B CN110411957 B CN 110411957B CN 201910803691 A CN201910803691 A CN 201910803691A CN 110411957 B CN110411957 B CN 110411957B
- Authority
- CN
- China
- Prior art keywords
- sample
- fruit
- samples
- apple
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 235000013399 edible fruits Nutrition 0.000 title claims abstract description 202
- 238000000034 method Methods 0.000 title claims abstract description 105
- 238000012937 correction Methods 0.000 claims abstract description 195
- 238000003860 storage Methods 0.000 claims abstract description 96
- 230000003595 spectral effect Effects 0.000 claims abstract description 58
- 238000012545 processing Methods 0.000 claims abstract description 26
- 238000013500 data storage Methods 0.000 claims abstract description 19
- 230000007613 environmental effect Effects 0.000 claims abstract description 13
- 235000010724 Wisteria floribunda Nutrition 0.000 claims description 119
- 238000012795 verification Methods 0.000 claims description 51
- 230000002159 abnormal effect Effects 0.000 claims description 33
- 238000013178 mathematical model Methods 0.000 claims description 26
- 238000001228 spectrum Methods 0.000 claims description 19
- 230000003287 optical effect Effects 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 14
- 230000001419 dependent effect Effects 0.000 claims description 11
- 238000010238 partial least squares regression Methods 0.000 claims description 11
- 238000010521 absorption reaction Methods 0.000 claims description 8
- 238000000862 absorption spectrum Methods 0.000 claims description 5
- 230000003203 everyday effect Effects 0.000 claims description 5
- 238000002474 experimental method Methods 0.000 claims description 5
- 230000010354 integration Effects 0.000 claims description 4
- -1 polytetrafluoroethylene Polymers 0.000 claims description 4
- 229920001343 polytetrafluoroethylene Polymers 0.000 claims description 4
- 239000004810 polytetrafluoroethylene Substances 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 238000012628 principal component regression Methods 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 238000000411 transmission spectrum Methods 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims description 3
- 230000002452 interceptive effect Effects 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 5
- 238000011161 development Methods 0.000 abstract description 4
- 241000220225 Malus Species 0.000 description 245
- 239000000523 sample Substances 0.000 description 184
- 235000021016 apples Nutrition 0.000 description 79
- 238000002329 infrared spectrum Methods 0.000 description 13
- 238000007781 pre-processing Methods 0.000 description 10
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 9
- 229910052744 lithium Inorganic materials 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 6
- 239000000463 material Substances 0.000 description 4
- 239000007787 solid Substances 0.000 description 4
- 230000001276 controlling effect Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 235000021022 fresh fruits Nutrition 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000002420 orchard Substances 0.000 description 2
- 239000010453 quartz Substances 0.000 description 2
- 229910052594 sapphire Inorganic materials 0.000 description 2
- 239000010980 sapphire Substances 0.000 description 2
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N silicon dioxide Inorganic materials O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 102000016550 Complement Factor H Human genes 0.000 description 1
- 108010053085 Complement Factor H Proteins 0.000 description 1
- 238000012271 agricultural production Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012468 concentrated sample Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012845 near infrared spectroscopy analysis Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/02—Food
- G01N33/025—Fruits or vegetables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
Abstract
The invention provides a method and a device for lossless and rapid prediction of shelf life and freshness of fruits. The method comprises the steps of obtaining spectral data of fruits with different storage times and recording the storage time and environmental data of the fruits; a regression algorithm is adopted to establish a fruit storage time model, a weighted correction predicted value is calculated based on the predicted value of the established model and the number of samples, a discrimination threshold value is calculated, and therefore the freshness degree of fruits is judged, and a method for formulating the shelf life of the fruits according to the key change period data of the fruits is provided. Based on the method, the fruit freshness analyzer is developed and comprises a light path system, a circuit system, a control system and a data storage and processing system; fruit spectral data can be collected and a model is called to predict the freshness of the fruit. The method obviously improves the prediction accuracy of the freshness and the shelf life of the fruits, can realize the lossless, rapid and accurate prediction of the freshness and the shelf life of the fruits, and provides technical reference for the development of a rapid analyzer of the freshness of the fruits.
Description
Technical Field
The invention relates to the field of spectral analysis, in particular to a method and a device for lossless and rapid prediction of shelf life and freshness of fruits.
Background
Fruit shelf life is an important measure of the freshness of fruit and is an important parameter in the marketing process. It is generally accepted that fruits that are in shelf life are fresh fruits, suitable for consumption; the fruits outside the shelf life are not fresh and are not suitable for eating. The loss and waste of fruits caused by the sudden drop of the freshness of the fruits due to improper storage conditions are often caused, and huge economic loss is caused. Therefore, it is necessary to predict the freshness of fruits and to study a method for establishing the shelf life of fruits based on the result of judging the freshness of fruits.
The change of different fruits in the storage process is different, even different individuals of the same fruit have different freshness degrees due to the storage environment, the storage position and the fruit individuals, and the prediction of the shelf life of the fruit is extremely difficult. These all make the prediction of the shelf life of the fruit and the judgment of the freshness of the fruit a technical problem in the field.
The spectral analysis has the technical characteristics of rapidness, high efficiency, no damage and environmental friendliness, and is always an important analysis technology in the field of rapid nondestructive analysis. In the field of industrial and agricultural production, multispectral analysis, typified by near-infrared spectroscopic analysis, has become a popular research field in recent years. The near infrared spectrum has the characteristics of no damage, rapidness, high efficiency, environmental friendliness and the like, and can record the spectral information of an experimental object in real time, so that the method is very suitable for the prediction and evaluation work of the freshness of the fruits.
The currently reported method for predicting the shelf life/quality guarantee period of fruits or other agricultural products/food based on near infrared spectrum and the like mainly has the following common problems: the method adopts samples at different time points to establish a discrimination model, and continuous changes of fruits or other agricultural products/food are not studied in detail, namely the prediction precision is not high; in addition, the storage time classification can only be carried out on the samples to be detected, namely the samples to be detected are judged to belong to a certain storage stage, and the prediction of the shelf life/freshness degree of the samples according to data such as near infrared spectrum and the like is not really realized. Therefore, it is urgently needed to develop a nondestructive, rapid and efficient analysis method to accurately predict the shelf life/freshness of fruits.
Disclosure of Invention
The invention aims to provide a method and a device for lossless and rapid prediction of shelf life and freshness of fruits.
In order to achieve the object, in a first aspect, the invention provides a method for lossless and rapid prediction of shelf life and freshness of fruit, comprising the following steps:
A. acquiring fruit sample spectral data and recording fruit storage time and environmental data; after removing abnormal values, preliminarily dividing the samples into three types of fresh, stale and critical periods according to the actual conditions and the actual storage time of the fruit samples, and performing diversity on the spectral data of the fresh samples and the stale samples by adopting an SPXY algorithm to divide the spectral data into a correction set and an external verification set, wherein the correction set is used for establishing a mathematical model, and the external verification set is used for performing blind sample verification on the established mathematical model;
B. the method comprises the steps of establishing a mathematical model of fruit storage time by taking spectral data as an independent variable and storage time as a dependent variable, respectively obtaining storage time predicted values of a fresh sample and an stale sample by utilizing the established model, calculating a weighted correction predicted value of the fresh sample according to the storage time predicted value of the fresh sample and the quantity of the fresh sample, calculating a weighted correction predicted value of the stale sample according to the storage time predicted value of the stale sample and the quantity of the stale sample, and calculating a discrimination threshold according to an average value of the weighted correction predicted values of the fresh sample, the quantity of the fresh sample, the average value of the weighted correction predicted values of the stale sample, the quantity of the stale sample and a correction coefficient, so as to judge the freshness degree of the fruit sample; when the weighted correction predicted value of the fruit sample correction set or the external verification set is smaller than the discrimination threshold, judging that the sample is a fresh sample; when the weighted correction predicted values of the fruit sample correction set and the external verification set are greater than or equal to the discrimination threshold, judging that the sample is an stale sample; evaluating the established model according to the fruit sample correction set and the judgment accuracy rate of the external verification set for concentrating the fresh samples and the stale samples, and judging the effectiveness of the model;
C. under the same experiment condition, acquiring the spectral data of the fruit sample to be tested, obtaining the storage time predicted value of the fruit sample to be tested by using the effective mathematical model in the step B, and comparing the predicted value with the corresponding actual storage time of the fruits in the fresh, stale and critical periods so as to judge the freshness degree of the fruit sample to be tested;
D. b, obtaining a storage time predicted value of the critical-period sample by using the effective mathematical model in the step B, and calculating a weighted correction predicted value of the critical-period sample according to the storage time predicted value of the critical-period sample, the number of fresh samples, the number of stale samples and the total number of the samples in the correction set; and when the weighted correction predicted value of the critical period sample shows that the prediction result of 'freshness' exceeds half of the number of observed samples on the same day, the predicted value is recorded as the original shelf life, and the original shelf life is multiplied by a correction coefficient to obtain the fruit shelf life.
Wherein, the fruit sample to be tested in the step C and the fruit sample in the steps A, B and D are the same kind of fruit and have the same storage environment.
The spectral data is selected from ultraviolet, visible, near-infrared, mid-infrared, fluorescence or terahertz spectral data and the like, and the expression form of the spectral data is absorption spectrum, absorption coefficient spectrum, transmission spectrum or reflection spectrum and the like; preferably, the spectral data is near infrared absorption spectral data.
The mathematical model is established by combining a regression algorithm with an interactive verification algorithm, wherein the regression algorithm is selected from multiple linear regression, principal component regression, partial least square regression, neural network regression, support vector machine regression and the like; partial least squares regression is preferred.
The environmental data includes temperature and relative humidity data of the fruit storage environment. When the fruit is apple, the environmental temperature is in the range of 19.7-23.2 ℃, and the optimal temperature is 21.8 ℃; the relative humidity of the environment is in the range of 10% to 28%, preferably 16%.
The fruit includes, but is not limited to, apples, preferably Fuji apples or Wanglin apples.
In the method, the sample volume ratio of the correction set to the external verification set in the step A is 7: 1-2.5: 1.
Performing data preprocessing on the correction set by adopting a baseline correction and data centralization data preprocessing mode; and establishing a correction model by adopting a partial least squares regression algorithm and a full-interactive verification algorithm, taking Fuji apple near infrared spectrum data as independent variables and Fuji apple storage days as dependent variables.
The calibration is concentrated, the number of fresh samples nFAnd number of stale samples nRThe ratio satisfies the following conditions: n is more than or equal to 0.892F/nRN is not more than 1.125 and not more than 0.892R/nF≤1.125。
When the fruits are apples, marking the picking day as day 0, preliminarily dividing samples from day 0 to day 14 into fresh samples, preliminarily dividing samples from day 15 to day 21 into critical period samples, and preliminarily dividing samples from day 22 and above into stale samples according to the actual conditions and the actual storage time of the apples.
The method for acquiring the spectral data of the fruit sample in step a comprises the following steps: the time is calculated from the day of picking fruits, the data acquisition times are not less than 25 times in a period of time (which can be 25-32 days), and the spectrum data are acquired at most once every day.
When the spectral data is near-infrared absorption spectral data, the data acquisition method comprises the following steps: a near-infrared spectrometer is utilized, and a polytetrafluoroethylene white board is taken as a background; single integration time 50 ms; accumulating for 50 times and taking an average value; the wavelength range is 901.841 nm-1700.930 nm, the spectrum center resolution is 8.00 nm-12.00 nm, and the optimal wavelength is 9.36 nm; the number of the spectral variables is 128, and the spectral variables are separated by 4.882 nm-7.883 nm, preferably 6.292 nm.
In the foregoing method, the removing the abnormal value in step a includes: the method comprises the steps of performing pre-modeling by adopting data centering preprocessing and combining a partial least square regression algorithm, taking spectral data as independent variables and storage time as dependent variables, calculating pre-modeling residual errors, namely the difference between pre-modeling predicted values and actual values, and determining a pre-modeling residual error judgment threshold value of an abnormal value according to the pre-modeling residual errors.
Preferably, the pre-modeling residual decision threshold of the abnormal value is set to 47.0 to 53.0, and more preferably 49.5 to 50.4. The pre-modeled residual decision threshold for outliers is most preferably 50.0 when the fruit is Fuji apple and 50.2 when the fruit is Wanglin apple.
And the samples of which the pre-modeling residual errors are greater than or equal to the pre-modeling residual error judgment threshold value of the set abnormal value are regarded as abnormal values and are to be eliminated.
In the foregoing method, the method for calculating the weighted correction predicted value of the fresh sample in step B is shown in formula (1):
in the formula (1), yFRCiWeighted correction of the predicted value, y, for the ith fresh sampleFiPrediction of storage time for the model for the ith fresh sample, nFIs to correct the number of fresh samples in the set, nTThe total number of samples was the calibration set.
The calculation method of the weighted correction predicted value of the stale sample is shown as the formula (2):
in the formula (2), yRRCjWeighted correction of the predicted value, y, for the jth stale sampleRjStorage time prediction for the model for the jth stale sample, nRIs to correct the number of stale samples in the set, nTThe total number of samples was the calibration set.
The method for calculating the discrimination threshold is shown as formula (3):
in the formula (3), Eps is a discrimination threshold, yFRCmWeighted correction of the mean value of the predicted values for fresh samples, yRRCmThe weighted average of the predicted correction values of the stale samples is shown, and a is a correction coefficient.
Preferably, when the fruit is an apple, the value range of the correction coefficient a is 1.0-2.0;
when the fruit is Fuji apple, the value range of the correction coefficient a is 1.4-1.7, more preferably 1.5-1.6, and most preferably 1.55;
when the fruit is Wanglin apple, the value range of the correction coefficient a is 1.0-1.3, more preferably 1.1-1.2, and most preferably 1.16.
In a preferred embodiment of the invention, when the fruit is Fuji apple, Eps-7.8517.
In a preferred embodiment of the invention, when the fruit is royal apple, Eps ═ 8.8297.
In the foregoing method, the method for calculating the weighted correction prediction value of the critical-period sample in step C is shown in formula (4):
in the formula (4), yRCCkWeighted correction predictor, y, for the kth critical phase samplekPrediction of the storage time of the kth critical phase sample for the model, nFIs to correct the number of fresh samples in the set, nRIs to correct the number of stale samples in the set, nTThe total number of samples was the calibration set.
The calculation method of the fruit shelf life is shown as the formula (5):
T＝floor(T0x H) formula (5)
In the formula (5), T is the shelf life of the fruit0The original shelf life is represented by H as a correction coefficient and floor as a downward takeAnd (5) integer operator.
Preferably, when the fruit is an apple, the value range of H is 0.70-1.00;
when the fruit is Fuji apple, the value range of H is 0.78-0.92, preferably 0.80-0.90, and more preferably 0.82;
when the fruit is Wanglin apple, the value range of H is 0.75-0.95, preferably 0.79-0.89, and more preferably 0.81.
In a second aspect, the present invention provides an apparatus (analyser) for carrying out the above method, the apparatus comprising an optical system, a control system, circuitry and a data storage and processing system;
the optical system is used for collecting the spectral data of the sample;
the circuit system is used for stably supplying power to the device;
the control system is used for controlling the working process of the device;
the data storage and processing system is used for data storage, mathematical model calling, result prediction, storage and output.
In a third aspect, the present invention provides the use of a method or apparatus as described above in the prediction of the freshness and/or shelf life of fruit.
By the technical scheme, the invention at least has the following advantages and beneficial effects:
the invention provides a method and a device for lossless and rapid prediction of shelf life and freshness of fruits. The method comprises the steps of acquiring spectral data of fruits with different storage time and recording the storage time and environmental data (temperature and relative humidity data) of the fruits; a regression algorithm is adopted to establish a fruit storage time model, a weighted correction predicted value is calculated for a fruit storage time predicted value and a sample number based on the established model, a discrimination threshold value is calculated, and therefore the freshness degree of fruits is judged, and a method for formulating the shelf life of the fruits according to the key change period data of the fruits is provided. Based on the method, the fruit freshness analyzer is developed and comprises a light path system, a circuit system, a control system and a data storage and processing system; the instrument can collect the spectral data of the fruit and call a mathematical model to predict the freshness of the fruit.
The method for predicting the shelf life and the freshness of the fruits provided by the invention obviously improves the freshness and the accuracy of the prediction of the shelf life of the fruits, and can realize the lossless, rapid and accurate prediction of the freshness and the shelf life of the fruits; the method effectively improves the efficiency of judging the freshness of the fruits and setting the shelf life, provides an effective method and technical support for reducing the loss and waste of the fruits in the storage process, and provides technical reference for the development of a portable device for rapidly analyzing the freshness of the fruits.
Drawings
FIG. 1 is a flow chart of a method for lossless and rapid prediction of freshness and shelf life of fruits according to a preferred embodiment of the present invention.
Fig. 2 is a graph showing the correlation between the predicted value and the actual value of the storage time of the Fuji apple correction set in embodiment 1 of the present invention.
Fig. 3 is a weighted correction prediction value of the fuji apple correction set in embodiment 1 of the present invention.
FIG. 4 is a weighted modified prediction value of the Fuji apple external verification set in example 1 of the present invention.
Fig. 5 is a diagram showing a correlation between the predicted value and the actual value of the storage time of the Wanglin apple correction set in embodiment 2 of the present invention.
Fig. 6 shows the weighted correction prediction value of the Wanglin apple correction set in embodiment 2 of the present invention.
Fig. 7 shows the weighted modified prediction values of the external verification set of the wang apple in embodiment 2 of the present invention.
FIG. 8 is a schematic sectional side view of a nondestructive rapid analyzer for fruit freshness in example 3 of the present invention; wherein, 1-1: probe section, 1-2: handle portion, 1-3: base portion, 2: spectrometer, 2-1: spectrometer radiator fan, 3: window piece, 3-1: window setter, 4: touch screen display, 5: engineering main board, 6: data interface, 7: lithium battery, 8: annular automatic trigger switch, 9: regulated power supply and charging protection module, 10: switch, 11: charging jack, 11-1: power indicator, 11-2: and a working indicator light.
FIG. 9 is a schematic front view of a nondestructive rapid analyzer for fruit freshness in example 3 of the present invention; wherein, 1-1: probe section, 1-2: handle portion, 1-3: base portion, 2: spectrometer, 3: window piece, 3-1: window setter, 8: annular automatic trigger switch.
FIG. 10 is a schematic rear view of a nondestructive rapid analyzer for fruit freshness in example 3 of the present invention; wherein, 1-1: probe section, 1-2: handle portion, 1-3: base portion, 4: touch screen display, 6: data interface, 10: switch, 11: charging jack, 11-1 power indicator light, 11-2 work indicator light.
Detailed Description
The invention provides a nondestructive rapid prediction method of fruit shelf life and freshness based on near infrared spectrum technology, aiming at the difficult problems of low prediction precision of fruit freshness and establishment of shelf life, and aiming at nondestructive rapid prediction of fruit freshness and efficient establishment of shelf life: acquiring spectral data of fruits with different storage times and recording the storage time, temperature and relative humidity data of the fruits; and establishing a fruit storage time model by adopting a regression algorithm, calculating a weighted correction predicted value for a fruit storage time predicted value and a sample number based on the established model, calculating a discrimination threshold value, judging the freshness degree of the fruit, and formulating the shelf life of the fruit according to the key change period data of the fruit. A nondestructive rapid analyzer for the freshness of fruits is developed based on the method: the system comprises an optical path system, a circuit system, a control system and a data storage and processing system; the instrument collects the spectral data of the fruits and calls the mathematical model to predict the freshness of the fruits, so that the problems that the prediction precision of the freshness of the fruits is not high and the shelf life is difficult to accurately set in the current actual work are solved.
In order to solve the problems of low prediction precision of the freshness of fruits and difficulty in accurately establishing the shelf life in the current practical work, the invention provides a nondestructive rapid prediction method of the shelf life and the freshness of fruits, which comprises the following steps:
(1) acquiring sample spectrum data and recording environmental data; after removing the abnormal value, dividing the sample into a fresh sample, an stale sample and a critical period sample according to the actual condition and the actual storage time of the sample.
(2) And establishing a fruit storage time model by adopting a regression algorithm, calculating a weighted correction predicted value based on the storage time predicted value and the sample number of the fruit samples by the established model, calculating a discrimination threshold value, judging the freshness degree of the fruit, and formulating the shelf life of the fruit according to the sample data of the critical period.
(3) Based on the method, the fruit freshness degree analyzer is developed and comprises an optical path system, a circuit system, a control system and a data storage and processing system; the analyzer stores the data acquired by the optical system and calls a mathematical model to predict the freshness of the fruits.
Further, the spectral data can be ultraviolet, visible, near infrared, mid-infrared, fluorescence, terahertz spectral data, and the representation form of the spectral data can be absorption spectrum, absorption coefficient spectrum, transmission spectrum, and reflection spectrum; further preferred is a near infrared absorption spectrum. More preferably, the near infrared absorption spectrum of the sample is measured using a near infrared absorption spectrometer of a digital light processing technology core.
The fruit is an apple, preferably a Fuji apple or a Wanglin apple.
Collecting spectral data of the sample: the calculation time of the apple picking day from the fruit tree is 0 day, the storage time is increased by 1 day every 1 day, the spectral data is collected within a period of time (which can be 25 to 32 days), the spectral data is collected at most once every day, and the observation/data collection frequency is not less than 25 times; the environmental data comprises temperature and relative humidity data of the environment where the apples are located in the experiment and storage process.
According to the actual condition and the actual storage time of the apples in the experimental process, the apples are preliminarily divided into three types of freshness, criticality and freshness: preliminarily dividing the samples from the 0 th day to the 14 th day into fresh samples, preliminarily dividing the samples from the 15 th day to the 21 st day into critical period samples, and preliminarily dividing the samples from the 22 nd day and above into stale samples. The pre-modeling is to establish a correction model by only adopting data centralization preprocessing and combining a partial least square regression algorithm, taking the near infrared spectrum data of the apples as independent variables and taking the storage days of the apples as dependent variables, and calculating a pre-modeling residual error, namely the difference between a pre-modeling predicted value and an actual value, through the model. Removing abnormal values, and adopting a pre-modeling residual error judgment threshold value as an abnormal value judgment standard, wherein the pre-modeling residual error judgment threshold value of the abnormal values is set to be 47.0-53.0, and preferably, the pre-modeling residual error judgment threshold value of the abnormal values is set to be 49.5-50.4; the pre-modeling residual error judgment threshold of the abnormal value of the Fuji apple is further preferably 50.0, and Fuji apples with the pre-modeling residual error value being more than or equal to 50.0 are regarded as the abnormal value and should be removed; for the Wanglin apple, the pre-modeling residual judgment threshold value of the abnormal value is further preferably 50.2, and the Wanglin apple sample with the pre-modeling residual value being greater than or equal to 50.2 is considered as the abnormal value and should be removed.
Data of two types of apples, namely 'fresh' and 'stale', which are divided according to the actual situation and the actual storage time of the apples in the experimental process are used for data diversity. The correction set is used for establishing a mathematical model; the external verification set is used as an external blind sample to carry out external blind sample verification on the established mathematical model; preferably, the sample volume ratio of the correction set to the external verification set is 7: 1-2.5: 1, and the correction set and the external verification set are selected from two types of apples, namely 'fresh' and 'stale', after abnormal values are removed by using an SPXY algorithm. Further, the calibration is concentrated, fresh sample number (n)F) And the number of stale samples (n)R) The ratio of n to n is required to satisfy 0.892 ≦ nF/nRN is not more than 1.125 and not more than 0.892R/nF≤1.125。
Further, the regression algorithm may be multiple linear regression, principal component regression, partial least squares regression, neural network regression, support vector machine regression; partial least squares regression is preferred.
Data preprocessing is carried out on the correction set data by adopting a data preprocessing method; and establishing a correction model by adopting the regression algorithm and combining with a full-interactive verification algorithm and taking the spectral data as independent variables and the storage days as dependent variables.
The method for judging the freshness of the fruit sample correction set and the external verification set comprises the following steps: and calculating a fresh sample weighted correction predicted value according to the fresh sample predicted value and the number of the fresh samples, calculating an stale sample weighted correction predicted value according to the stale sample predicted value and the number of the stale samples, and calculating a discrimination threshold according to the fresh sample weighted correction predicted value average value, the number of the fresh samples, the stale sample weighted correction predicted value average value, the number of the stale samples and the correction coefficient. When the weighted correction predicted values of the fruit sample correction set and the external verification set are smaller than the discrimination threshold, judging that the sample is a fresh sample; and when the weighted correction predicted values of the fruit sample correction set and the external verification set are greater than or equal to the discrimination threshold, judging that the sample is not fresh.
The method for making the shelf life of the fruits comprises the following steps: and calculating the weighted correction predicted value of the critical-period samples according to the storage time predicted value, the number of fresh samples, the number of stale samples and the total number of the correction set samples of the critical-period samples. When the predicted value of the critical period sample is stale and the predicted result exceeds half of the number of the samples in the current day, the predicted value is determined as the original shelf life T0The original shelf life T0Multiplying by a correction factor H, and rounding the product to obtain the shelf life T, i.e. T-floor (T)0X H), wherein, T0For the original shelf life, "floor" is the round-down operator. When the fruit is an apple, the value range of the correction coefficient H is 0.70-1.00; when the fruit is Fuji apple, the value range of H is 0.78-0.92, preferably 0.80-0.90, and more preferably 0.82; when the fruit is Wanglin apple, the value range of H is 0.75-0.95, preferably 0.79-0.89, and more preferably 0.81.
The calculation method of the weighted correction predicted value of the fresh sample is shown as the formula (1).
In the formula (1), yFRCiWeighted correction of predicted value, y, for ith fresh fruit sampleFiPrediction of storage time for the model for the ith fresh sample, nFIs to correct the number of fresh samples in the set, nTThe total number of samples was the calibration set.
The calculation method of the weighted correction predicted value of the stale sample is shown as the formula (2).
In the formula (2), yRRCjWeighted correction of predicted value, y, for jth fresh fruit sampleRjStorage time prediction for the model for the jth stale sample, nRIs to correct the number of stale samples in the set, nTThe total number of samples was the calibration set.
The calculation method of the discrimination threshold is shown as formula (3).
In the formula (3), Eps is a discrimination threshold, yFRCmWeighted correction of the predicted mean value, y, for fresh samplesRRCmThe average value of the predicted values is weighted and corrected for the stale sample, and a is a correction coefficient.
Preferably, when the fruit is an apple, the value range of the correction coefficient a is 1.0-2.0; when the fruit is Fuji apple, the value range of the correction coefficient a is 1.4-1.7, more preferably 1.5-1.6, and most preferably 1.55; when the fruit is Wanglin apple, the value range of the correction coefficient a is 1.0-1.3, more preferably 1.1-1.2, and most preferably 1.16.
Preferably, when the fruit is Fuji apple, Eps ═ 7.8517.
Preferably, when the fruit is Wanglin apple, Eps ═ 8.8297.
The calculation method of the weighted correction predicted value of the critical-period sample is shown as the formula (4).
In the formula (4), yRCCkWeighted correction predictor, y, for the kth critical phase samplejFor model to the kth criticalPrediction of the storage time of the phase samples, nFIs the number of fresh samples, nRIs the number of stale samples, nTThe total number of samples was the calibration set.
Further, the nondestructive fruit freshness fast analyzer is designed and developed based on the fruit sample correction set and external verification set freshness distinguishing method, and comprises an optical system, a circuit system, a control system and a data storage and processing system, wherein the analyzer comprises a probe part, a handle part and a base part. The analyzer stores data acquired by the optical system, obtains sample prediction data according to the established model, calculates a weighted correction prediction value, and judges the freshness degree of the fruit sample through a discrimination threshold value.
Further, the optical system comprises a spectrometer, a spectrometer cooling fan, a window and a window positioner; the spectrometer can be an ultraviolet-visible spectrometer, a near-infrared spectrometer, a mid-infrared spectrometer, a fluorescence spectrometer and a terahertz spectrometer, and preferably a near-infrared spectrometer; the near-infrared spectrometer preferably adopts a digital light processing technology as a near-infrared spectrometer with an inner core; the window sheet is made of a material allowing working light to penetrate, and preferably made of near-infrared quartz or sapphire material; the middle of the objective table is provided with a light through hole, and the diameter of the light through hole is 2 mm-10 mm, preferably 3 mm-5 mm. The optical system is used for collecting the spectral data of the sample in the working process of the analyzer.
Furthermore, the circuit system comprises a lithium battery, a stabilized voltage power supply, a charging protection module, a switch, a power indicator, a work indicator and a charging jack. The analyzer adopts two power supply modes of a lithium battery and a stabilized voltage supply, and can adapt to different application scenes such as laboratory operation, field/field operation and the like, wherein the stabilized voltage supply and the charging protection module can convert alternating current into an available power supply of an instrument for the instrument to work and use, and simultaneously play a role in charging the lithium battery; the switch adopts a key switch or a ship-shaped switch, and the ship-shaped switch is preferably selected; the power indicator lights are turned on after the instrument is powered on and turned off after the instrument is powered off, red is displayed in a non-charging state powered by a lithium battery, yellow is displayed in a charging state of an external power supply, and green is displayed in a charging completion state; the work indicator lights up when the spectrum appearance gathers the spectral data, shows blue, extinguishes after the spectrum appearance gathers the spectral data. The circuit system is used for stably supplying power to the analyzer in the working process of the analyzer.
Further, the control system comprises an engineering mainboard, a touch screen display and an annular automatic trigger switch. The engineering mainboard is provided with a central processing unit, a display card and a mainboard radiator. The control system is used for controlling the working process of the analyzer.
Furthermore, the data storage and processing system comprises a random access memory, a read-only memory, a solid state disk and a data interface. The random access memory, the read-only memory and the solid state disk are fixed on the engineering mainboard. The data interface can be one or more of USB, OTG, Type-C interface. The data is stored in a processing system and used for data storage, mathematical model calling, result prediction and output.
The spectrometer is electrically connected with the circuit system, the control system and the data storage and processing system. The analyzer stores data acquired by the optical system, obtains sample prediction data according to the established model, calculates a weighted correction prediction value, and judges the freshness degree of the fruit sample through a discrimination threshold value.
The method can accurately predict the freshness of the fruits and provide a method for making the shelf life of the fruits, and meanwhile, the analyzer can realize the rapid nondestructive analysis of the freshness of the fruits and output a judgment result, so that technical references can be provided for the rapid judgment of the shelf life of the fruits and the research and development of portable instruments.
The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention. Unless otherwise specified, the technical means used in the examples are conventional means well known to those skilled in the art, and the raw materials used are commercially available products.
Example 1 Fuji apple shelf life establishment and freshness discrimination analysis
With reference to fig. 1 to 4, the formulation of the shelf life of Fuji apples and the judgment of the freshness are analyzed. The agricultural product is apple, preferably Fuji apple produced in Changping district of Beijing city. Ripe Fuji apples are picked in the orchard, the picking day is recorded as '0 th day', and the storage days of the apples are increased by 1 every 1 day thereafter. The method comprises the steps of collecting near infrared spectrum data of 8 Fuji apples each time, collecting the data at most 1 time every day, and simultaneously recording environmental temperature and relative humidity data. After the experiment is finished, the duration is 32 days, and the times of observing/collecting the spectral data are 25 times. Carrying out data statistics on the ambient temperature and the relative humidity, wherein the minimum value of the ambient temperature is 19.7 ℃, the maximum value is 23.2 ℃, and the average value is 21.8 ℃; the ambient relative humidity is 10% minimum, 28% maximum and 16% average.
The method comprises the following steps of (1) acquiring experimental data by adopting a near infrared spectrometer based on Digital Light Processing (DLP), wherein the instrument parameters are as follows: a polytetrafluoroethylene white board is used as a spectrum background reference; single integration time 50 ms; accumulating for 50 times and averaging; the wavelength range is 901.841 nm-1700.930 nm, the spectrum center resolution is 8.00 nm-12.00 nm, preferably 9.36nm, the number of the spectrum variables is 128, and the spectrum variables are separated by 4.882 nm-7.883 nm, preferably 6.292 nm.
According to the actual situation and the actual storage time of Fuji apples, Fuji apples are divided into three categories of 'fresh', 'critical' and 'stale'. For Fuji apples in the experimental process, apples stored within 14 days (including day 14) are initially used as 'fresh' samples, apples stored for more than 22 days (including day 22) are initially used as 'stale' samples, and apples stored for 15-21 days (including day 15 and day 21) are initially used as 'critical' samples. The method comprises the steps of taking Fuji apple near infrared spectrum data as independent variables, taking apple storage days as dependent variables, establishing a correction model by adopting a Partial Least Squares Regression (PLSR) algorithm on the basis of centralization processing of the data, and calculating a pre-modeling residual error according to model data, namely the difference between a pre-modeling predicted value and an actual value. The method adopts a pre-modeling residual as a judgment standard of an abnormal value, the pre-modeling residual judgment threshold of the abnormal value is set to 47.0-53.0, preferably, the pre-modeling residual judgment threshold of the abnormal value is set to 49.5-50.4, and for Fuji apples, the pre-modeling residual judgment threshold of the abnormal value is further preferably 50.0. Fuji apple samples with a pre-modeling residual value greater than or equal to 50.0 are considered outliers and should be rejected.
And dividing the 'fresh' and 'stale' Fuji apple data after the abnormal values are removed into a correction set and an external verification set by adopting an SPXY algorithm. The correction set is used for establishing a mathematical model; and the external verification set is used as an external blind sample to carry out external blind sample verification on the established mathematical model. Further, the calibration is concentrated, fresh sample number (n)F) And the number of stale samples (n)R) The ratio of n to n is required to satisfy 0.892 ≦ nF/nRN is not more than 1.125 and not more than 0.892R/nFLess than or equal to 1.125. Specifically, after removing the outliers, the Fuji apple correction set contains 99 observation samples, wherein 52 samples are "fresh" and 47 samples are "not fresh"; the Fuji apple external validation set contained 33 observed samples, of which 18 were "fresh" samples and 15 were "stale". In addition, after outliers were removed, the Fuji apple was "borderline" for 41 samples.
Performing data preprocessing on the correction set by adopting a baseline correction and data centralization data preprocessing mode; and establishing a correction model by adopting a partial least squares regression algorithm and a full-interactive verification algorithm, taking Fuji apple near infrared spectrum data as independent variables and Fuji apple storage days as dependent variables. The correlation diagram of predicted values and actual values of the Fuji apple correction set is shown in figure 2.
And calculating the weighted correction predicted value of the fresh Fuji apple sample according to the predicted value of the fresh Fuji apple sample and the number of the fresh Fuji apple samples. The method for calculating the weighted correction predicted value of the fresh Fuji apple sample is shown as the formula (1).
In the formula (1), yFRCiWeighted correction of predicted value, y, for the ith fresh Fuji apple sampleFiStorage time prediction for the model on the ith fresh Fuji apple sample, nFIs the corrected and concentrated number of fresh Fuji apple samples, nTTotal number of samples in the correction set for Fuji apple.
And calculating the weighted correction predicted value of the fresh-off Fuji apple sample according to the predicted value of the fresh-off Fuji apple sample and the number of the fresh-off Fuji apple samples. The method for calculating the weighted correction predicted value of the stale Fuji apple sample is shown as the formula (2).
In the formula (2), yRRCjWeighted correction of predicted value, y, for jth fresh Fuji apple sampleRjStorage time prediction value, n, for the model on the jth fresh Fuji apple sampleRThe number of samples of the non-fresh Fuji apples in the Fuji apple correction set, nTTotal number of samples in the correction set for Fuji apple.
The method includes that a model is adopted to predict an external verification set sample to obtain a predicted value, and then a calculator is used for weighting and correcting the predicted value according to an equation (1) or an equation (2) according to the actual situation of the external verification set sample. The weighting correction prediction value of the Fuji apple correction set is shown in figure 3, and the weighting correction prediction value of the Fuji apple external verification set is shown in figure 4.
And calculating a discrimination threshold value according to the average value of the weighted correction predicted values of the fresh Fuji apple samples, the number of the fresh Fuji apple samples, the average value of the weighted correction predicted values of the stale Fuji apple samples, the number of the stale Fuji apple samples and the Fuji apple correction coefficient. The calculation method of the discrimination threshold is shown as formula (3).
In the formula (3), Eps is a discrimination threshold, yFRCmWeighted correction of the predicted mean value, y, for fresh Fuji apple samplesRRCmThe average value of the weighted correction predicted values of the samples of the fresh Fuji apples is shown, and a is a Fuji apple correction coefficient. For Fuji apples, the value range of the correction coefficient a is 1.4-1.7, preferably 1.5-1.6, and further preferablyAnd is selected to be 1.55. And (4) calculating a Fuji apple discrimination threshold Eps-7.8517 according to Fuji apple experimental data.
The method for judging the freshness of Fuji apples comprises the following steps: comparing the weighted correction predicted values of the Fuji apple sample correction set and the external verification set with a discrimination threshold Eps, and judging that the sample is fresh when the weighted correction predicted values of the Fuji apple sample correction set and the external verification set are smaller than the discrimination threshold Eps; and when the weighted correction predicted values of the Fuji apple sample correction set and the external verification set are greater than or equal to the judgment threshold Eps, judging that the sample is not fresh.
The accuracy of the established model can be evaluated by the judgment accuracy of the fresh Fuji apple sample and the fresh Fuji apple sample in the Fuji apple correction set and the external verification set, as shown in Table 1. As can be seen from table 1, for the fuji apples, the discrimination accuracy rates of the fresh fuji apple sample, the stale fuji apple sample and the total sample in the calibration set are all 100%, and the discrimination accuracy rates of the fresh fuji apple sample, the stale fuji apple sample and the total sample in the external verification set are 94.4%, 100% and 97.0%, respectively. The data in table 1 show that the weighted average prediction value calculated based on the prediction result of the established model is compared with the judgment threshold value of the freshness degree of the Fuji apples, so that the judgment accuracy of the freshness degree of the Fuji apples is high, and the actual detection requirement can be met.
TABLE 1 Fuji apple model accuracy statistics
The method for calculating the weighted correction predicted value of the Fuji apple sample in the critical period is shown as a formula (4).
In the formula (4), yRCCkWeighted correction of predicted value, y, for the kth critical-period Fuji apple samplejFor model to the kth criticalPredicted storage time, n, for Fuji apple samplesFNumber of samples of fresh Fuji apples, nRNumber of samples of fresh Fuji apples, nTThe total number of Fuji apple samples was corrected.
The method for making the shelf life of Fuji apples comprises the following steps: and calculating the weighted correction predicted value of the Fuji apple sample in the critical period according to the storage time predicted value of the Fuji apple sample in the critical period, the number of the fresh Fuji apple samples, the number of the non-fresh Fuji apple samples and the total number of the Fuji apple samples in the correction set. When the predicted value of the Fuji apple sample in the critical period is 'stale', the predicted result exceeds half of the number of the samples in the same day, the predicted value is determined as the original shelf life T0The original shelf life T0Multiplying by Fuji apple correction coefficient H, and rounding the product to obtain shelf life T, i.e. T-floor (T)0H) Wherein "floor" is the floor operator. Preferably, samples within 14 days (including 14 days) are determined as fresh Fuji apple samples, samples within 15-21 days (including 15 days and 21 days) are determined as critical-period Fuji apple samples, and samples above 22 days (including 22 days) are determined as non-fresh Fuji apple samples; the value range of apple H is 0.70-1.00, the value range of Fuji apple H is 0.78-0.92, preferably 0.80-0.90, and further preferably H is 0.82.
The actual storage days, the weighted correction predicted values and the discrimination results of the Fuji apple critical period samples are shown in Table 2. As can be seen from Table 2, when the storage time of Fuji apple samples in the critical period is 15-18 days, the number of samples which are judged to be 'stale' does not exceed half of the number of observed samples on the day; when the storage time of Fuji apple samples in the critical period reaches 20 days, the number of samples judged to be 'stale' exceeds half of the number of samples observed on the same day. Thus, the original shelf life T of Fuji apples 020; combining with the Fuji apple correction coefficient H, calculating to obtain the shelf life T ═ floor (T) of Fuji apples0Xh floor (20 × 0.82) ═ floor (16.4) ═ 16 days.
TABLE 2 actual days of storage, weighted correction prediction and discrimination results for Fuji apple critical period samples
Example 2 establishment of shelf life and freshness discrimination of Wanglin apples
And (3) combining the figure 1 and the figures 5-7 to analyze the establishment of the shelf life of the Wanglin apples and the judgment of the freshness. The agricultural product is apple, preferably Wanglin apple produced in Changping district of Beijing city. Mature Wanglin apples are picked in an orchard, the picking day is recorded as '0 th day', and the storage days of the apples are increased by 1 every 1 day thereafter. The near infrared spectrum data of 8 Wanglin apples are collected each time, the data are collected at most 1 time every day, and meanwhile, the environmental temperature and the relative humidity data are recorded. After the experiment is finished, the duration is 32 days, and the times of observing/collecting the spectral data are 25 times. Carrying out data statistics on the ambient temperature and the relative humidity, wherein the minimum value of the ambient temperature is 19.7 ℃, the maximum value is 23.2 ℃, and the average value is 21.8 ℃; the ambient relative humidity is 10% minimum, 28% maximum and 16% average.
The method comprises the following steps of (1) acquiring experimental data by adopting a near infrared spectrometer based on Digital Light Processing (DLP), wherein the instrument parameters are as follows: a polytetrafluoroethylene white board is used as a spectrum background reference; single integration time 50 ms; accumulating for 50 times and averaging; the wavelength range is 901.841 nm-1700.930 nm, the spectrum center resolution is 8.00 nm-12.00 nm, preferably 9.36nm, the number of the spectrum variables is 128, and the spectrum variables are separated by 4.882 nm-7.883 nm, preferably 6.292 nm.
According to the actual situation and the actual storage time of the Wanglin apples, the Wanglin apples are divided into three types of 'fresh', 'critical' and 'stale'. Aiming at Wanglin apples in the experimental process, apples stored for 14 days (including 14 days) are taken as 'fresh' samples preliminarily, apples stored for more than 22 days (including 22 days) are taken as 'stale' samples preliminarily, and apples stored for 15-21 days (including 15 days and 21 days) are taken as 'critical' samples preliminarily. The method comprises the steps of taking Wanglin apple near infrared spectrum data as independent variables, taking apple storage days as dependent variables, establishing a correction model by adopting a Partial Least Squares Regression (PLSR) algorithm on the basis of centralization processing of the data, and calculating a pre-modeling residual error according to model data, namely the difference between a pre-modeling predicted value and an actual value. The method adopts a pre-modeling residual as a judgment standard of an abnormal value, the pre-modeling residual judgment threshold of the abnormal value is set to 47.0-53.0, preferably, the pre-modeling residual judgment threshold of the abnormal value is set to 49.5-50.4, and for Wanglin apples, the pre-modeling residual judgment threshold of the abnormal value is further preferably 50.2. Samples of Wanglin apples with a pre-modeling residual value of 50.2 or more are considered as abnormal values and should be removed.
And dividing the 'fresh' and 'stale' Wanglin apple data after the abnormal values are removed into a correction set and an external verification set by adopting an SPXY algorithm. The correction set is used for establishing a mathematical model; and the external verification set is used as an external blind sample to carry out external blind sample verification on the established mathematical model. Further, the calibration is concentrated, fresh sample number (n)F) And the number of stale samples (n)R) The ratio of n to n is required to satisfy 0.892 ≦ nF/nRN is not more than 1.125 and not more than 0.892R/nFLess than or equal to 1.125. Specifically, after removing abnormal values, the Wanglin apple correction set comprises 92 observation samples, wherein 45 samples are fresh, and 47 samples are not fresh; the Wanglin apple external validation set contained 31 observation samples, 15 "fresh" samples, and 16 "stale" samples. In addition, 38 "critical" samples of Fuji apples were obtained after outliers were removed.
Performing data preprocessing on the correction set by adopting a baseline correction and data centralization data preprocessing mode; and establishing a correction model by adopting a partial least squares regression algorithm and a full-interactive verification algorithm, taking the near infrared spectrum data of the Wanglin apples as independent variables and taking the storage days of the Wanglin apples as dependent variables. The correlation diagram of the predicted value and the actual value of the Wanglin apple correction set is shown in figure 5.
And calculating the weighted correction predicted value of the fresh Wanglin apple sample according to the predicted value of the fresh Wanglin apple sample and the number of the fresh Wanglin apple samples. The method for calculating the weighted correction predicted value of the fresh Wanglin apple sample is shown as the formula (1).
In the formula (1), yFRCiWeighted correction of predicted value, y, for the ith fresh Wanglin apple sampleFiStorage time prediction value, n, for the model on the ith fresh Wanglin apple sampleFIs the number of samples, n, of Wanglin apples in corrected and concentrated fresh Wanglin applesTThe total number of samples in the Wanglin apple calibration set.
And calculating the weighted correction predicted value of the samples of the fresh Wanglin apples according to the predicted value of the samples of the fresh Wanglin apples and the number of the samples of the fresh Wanglin apples. The method for calculating the weighted correction predicted value of the fresh Wanglin apple sample is shown as the formula (2).
In the formula (2), yRRCjWeighted correction of predicted value, y, for jth King apple sampleRjStorage time prediction value, n, for the model on the jth cowry apple sampleRIs the corrected and concentrated sample number of the fresh Wanlin apples, nTThe total number of samples in the Wanglin apple calibration set.
The method includes that a model is adopted to predict an external verification set sample to obtain a predicted value, and then a calculator is used for weighting and correcting the predicted value according to an equation (1) or an equation (2) according to the actual situation of the external verification set sample. The correction set weighting correction prediction value of the Wanglin apple is shown in figure 6, and the external verification set weighting correction prediction value of the Wanglin apple is shown in figure 7.
And calculating a discrimination threshold value according to the average value of the weighted correction predicted values of the fresh Wanglin apple samples, the number of the fresh Wanglin apple samples, the average value of the weighted correction predicted values of the fresh Wanglin apple samples, the number of the fresh Wanglin apple samples and the correction coefficient. The calculation method of the discrimination threshold is shown as formula (3).
In the formula (3), Eps is a discrimination threshold, yFRCmWeighted correction of the predicted value average, y, for fresh Wanglin apple samplesRRCmThe average value of the weighted correction predicted values of the samples of the fresh Wanglin apples is obtained, and a is the correction coefficient of the Wanglin apples. For Wanglin apples, the value range of the correction coefficient a is 1.0-1.3, preferably 1.1-1.2, and further preferably 1.16. And (4) calculating to obtain a Wanlin apple discrimination threshold Eps-8.8297 according to the Wanlin apple experimental data.
The method for judging the freshness of Wanglin apples comprises the following steps: comparing the weighted correction predicted values of the Wanglin apple sample correction set and the external verification set with a discrimination threshold Eps, and judging that the sample is fresh when the weighted correction predicted values of the Wanglin apple sample correction set and the external verification set are smaller than the discrimination threshold Eps; and when the weighted correction predicted values of the Wanglin apple sample correction set and the external verification set are greater than or equal to the judgment threshold Eps, judging that the sample is not fresh.
The accuracy of the established model can be evaluated by the judgment accuracy conditions of the fresh Wang Lin apple sample and the fresh Anhui Wang Lin apple sample concentrated through the Wang Lin apple correction set and the external verification, as shown in Table 3. As can be seen from table 3, for the royal apples, the discrimination accuracy rates of the fresh royal apple sample, the fresh royal apple sample and the total sample collected in the calibration are all 100%, and the discrimination accuracy rates of the fresh royal apple sample, the fresh royal apple sample and the total sample collected in the external verification are 93.3%, 93.8% and 93.5%, respectively. The data in table 3 show that the weighted average prediction value calculated based on the prediction result of the established model is compared with the judgment threshold value of the freshness degree of the Wanglin apples, so that the judgment accuracy of the freshness degree of the Wanglin apples is high, and the actual detection requirement can be met.
TABLE 3 Wanglin apple model accuracy statistics
The calculation method of the weighted correction predicted value of the Wanglin apple sample in the critical period is shown as the formula (4).
In the formula (4), yRCCkWeighted correction of predicted value, y, for the kth Critical stage Wanglin apple samplejStorage time prediction value, n, of the model on the k-th critical-period Wanglin apple sampleFIs the number of samples of fresh Wanglin apples, nRNumber of samples of Anuman apple, nTThe total number of samples of Wanglin apples was corrected.
The method for establishing the shelf life of Wanglin apples comprises the following steps: and calculating the weighted correction predicted value of the Wanglin apple sample in the critical period according to the storage time predicted value of the Wanglin apple sample in the critical period, the number of the fresh Wanglin apple samples and the total number of the Wanglin correction apple samples. The predicted value of the Wanglin apple sample in the critical period is judged to be the original shelf life T when the predicted result that the sample is not fresh is more than half of the sample number on the same day0The original shelf life T0Multiplying by Wanglin apple correction coefficient H, and rounding the product to obtain the shelf life T, i.e. T ═ floor (T)0H) Wherein "floor" is the floor operator. Preferably, a sample within 14 days (including 14 days) is determined as a fresh Wanglin apple sample, a sample within 15-21 days is determined as a Wanglin apple sample in a critical period, and a sample more than 22 days (including 22 days) is determined as an unrivaled Wanglin apple sample; the value range of apple H is 0.70-1.00, the value range of Wanglin apple H is 0.75-0.95, preferably 0.79-0.89, and further preferably H is 0.81.
The actual storage days, the weighted correction predicted values and the discrimination results of the Wanglin apple critical period samples are shown in Table 4. As can be seen from Table 4, when the samples of Wanglin apples in the critical period are stored for 15-18 days, the number of samples judged to be 'stale' does not exceed half of the number of samples observed on the same day; judging the samples to be not fresh when the storage time of the Wanglin apple samples in the critical period reaches 20 daysThe number of samples exceeds half of the number of samples observed on the day. Thus, the original shelf life T of Wanglin apples 020; combining with the correction coefficient H of Wanglin apple, the shelf life T of Wanglin apple is floor (T)0Xh floor (20 × 0.81) floor (16.2) 16 days.
TABLE 4 actual days of storage, weighted correction prediction and discrimination results for Wanglin apple critical period samples
Example 3 development and application of nondestructive fast analyzer for fresh degree of fruit
In order to realize the prediction of the freshness and shelf life of the apples, the invention provides a design scheme of a nondestructive rapid fruit freshness analyzer. The design scheme of the nondestructive fast fruit freshness analyzer is described with reference to fig. 8-10.
The nondestructive fast analyzer for the freshness of fruits comprises an optical system, a circuit system, a control system and a data storage and processing system, and the analyzer comprises a probe part 1-1, a handle part 1-2 and a base part 1-3. The analyzer stores data acquired by the optical system, obtains sample prediction data according to the established model, calculates a weighted correction prediction value, and judges the freshness degree of the fruit sample through a discrimination threshold value.
The optical system comprises a spectrometer 2, a spectrometer cooling fan 2-1, a window 3 and a window positioner 3-1; the spectrometer can be an ultraviolet-visible spectrometer, a near-infrared spectrometer, a mid-infrared spectrometer, a fluorescence spectrometer and a terahertz spectrometer, and preferably a near-infrared spectrometer; the near-infrared spectrometer is preferably a near-infrared spectrometer with a kernel based on Digital Light Processing (DLP) technology; the window piece 3 is made of a material allowing working light to penetrate through, and preferably made of near-infrared quartz or sapphire material; the middle of the objective table is provided with a light through hole, and the diameter of the light through hole is 2 mm-10 mm, preferably 3 mm-5 mm. The optical system is used for collecting the spectral data of the sample in the working process of the analyzer.
The circuit system comprises a lithium battery 7, a stabilized voltage power supply and charging protection module 9, a switch 10, a power indicator lamp, a work indicator lamp and a charging jack 11. The analyzer adopts two power supply modes of a lithium battery 7 and a stabilized voltage supply, and can adapt to different application scenes such as laboratory operation, field/field operation and the like, wherein the stabilized voltage supply and charging protection module 9 can convert alternating current into an available power supply for the instrument to work and use, and simultaneously plays a role in charging the lithium battery 7; the switch 10 adopts a key switch or a boat-shaped switch, preferably a boat-shaped switch; the power indicator lights are turned on after the instrument is powered on and turned off after the instrument is powered off, red is displayed in a non-charging state powered by a lithium battery 7, yellow is displayed in a charging state of an external power supply, and green is displayed in a charging completion state; the work indicator lights up when the spectrometer collects the spectral data, displays blue, and lights out after the spectrometer 2 finishes collecting the spectral data. The circuit system is used for stably supplying power to the analyzer in the working process of the analyzer.
The control system comprises an engineering mainboard 5, a touch screen display 4 and an annular automatic trigger switch 8. Wherein, be equipped with central processing unit, display card, mainboard radiator on the engineering mainboard 5. The touch screen display 4 and the annular automatic trigger switch 8 are electrically connected with the engineering mainboard 5. The annular automatic trigger switch 8 is internally provided with a pressure sensor, can automatically trigger the spectrometer to collect the tested spectrum when the instrument is in contact with the tested sample and generates a certain pressure, and has higher sensitivity compared with the traditional trigger switch. The design adopts the annular automatic trigger switch with the built-in pressure sensor instead of the traditional hardware, namely the key-type trigger switch, so that the hardware number is reduced, the hardware manufacturing cost can be further reduced, and the compiling difficulty of related control software can be reduced. The control system is used for controlling the working process of the analyzer.
The data storage and processing system comprises a random access memory, a read-only memory, a solid state disk and a data interface. The random access memory, the read-only memory and the solid state disk are fixed on the engineering mainboard. The data interface 6 can be one or more of USB, OTG and Type-C interface. The data is stored in a processing system and used for data storage, mathematical model calling, result prediction and output.
The spectrometer is electrically connected with the circuit system, the control system and the data storage and processing system. The analyzer stores data acquired by the optical system, obtains sample prediction data according to the established model, calculates a weighted correction prediction value, and judges the freshness degree of the fruit sample through a discrimination threshold value.
The analysis and the predicted value, the weighted correction predicted value, the actual situation and the judgment result of the Fuji apples with different storage time are shown in a table 5. According to example 1, the shelf life of Fuji apples is 16 days. The sample with the judged result of 'fresh' is correct when being in the shelf life of Fuji apples, otherwise, the sample is wrong; the sample with the judgment result of 'stale' is correct when the sample is out of the shelf life of Fuji apples, otherwise, the sample is wrong. As can be seen from table 5, of the 41 observation samples, 23 samples in the shelf life of fuji apples were correctly distinguished by 22 samples, and the accuracy was 95.7%; 18 samples out of shelf life are correctly distinguished, wherein the accuracy is 100%; the total accuracy rate is 97.6%.
TABLE 5 Fuji apple prediction results based on fruit freshness nondestructive rapid analyzer
The analysis and the predicted values, the weighted correction predicted values, the actual conditions and the judgment results of the Wanglin apples in different storage times are shown in a table 6. According to example 2, the shelf life of Wanglin apples is 16 days. Judging whether the sample with the judged result of 'fresh' is in the shelf life of the Wanglin apples, if so, judging that the sample is correct, otherwise, judging that the sample is wrong; and judging that the sample with the 'stale' result is out of the shelf life of the Wanglin apples, if the sample is correct, otherwise, the sample is wrong. As can be seen from Table 6, 22 samples in the shelf life of Wanglin apples are obtained from 40 observation samples, 21 samples are correctly distinguished, and the accuracy is 95.5%; 18 samples out of shelf life are correctly distinguished, wherein the accuracy is 100%; the total accuracy rate is 97.5%.
TABLE 6 prediction of Wanglin apples based on fruit freshness nondestructive rapid analyzer
The invention provides a nondestructive rapid prediction method and an analyzer for shelf life and freshness of fruits. Based on the near infrared spectrum technology, acquiring the spectrum data of the fruits at different storage times and recording the storage time, the temperature and the relative humidity of the fruits; and establishing a fruit storage time model by adopting a regression algorithm, calculating a weighted correction predicted value for a fruit storage time predicted value and a sample number based on the established model, calculating a discrimination threshold value, judging the freshness degree of the fruit, and formulating the shelf life of the fruit according to the key change period data of the fruit. Based on the method, the fruit freshness analyzer is developed and comprises a light path system, a circuit system, a control system and a data storage and processing system; the instrument collects the spectral data of the fruit and calls a mathematical model to predict the freshness of the fruit. The invention can accurately predict the freshness of the fruits, provides a method for making the shelf life of the fruits and provides a technical reference for the development of a rapid analyzer for the freshness of the fruits.
Although the invention has been described in detail hereinabove with respect to a general description and specific embodiments thereof, it will be apparent to those skilled in the art that modifications or improvements may be made thereto based on the invention. Accordingly, such modifications and improvements are intended to be within the scope of the invention as claimed.
Claims (19)
1. The nondestructive rapid prediction method for the shelf life and the freshness of the fruits is characterized by comprising the following steps of:
A. acquiring fruit sample spectral data and recording fruit storage time and environmental data; after removing abnormal values, preliminarily dividing the samples into three types of fresh, stale and critical periods according to the actual conditions and the actual storage time of the fruit samples, and performing diversity on the spectral data of the fresh samples and the stale samples by adopting an SPXY algorithm to divide the spectral data into a correction set and an external verification set, wherein the correction set is used for establishing a mathematical model, and the external verification set is used for performing blind sample verification on the established mathematical model;
B. the method comprises the steps of establishing a mathematical model of fruit storage time by taking spectral data as an independent variable and storage time as a dependent variable, respectively obtaining storage time predicted values of a fresh sample and an stale sample by utilizing the established model, calculating a weighted correction predicted value of the fresh sample according to the storage time predicted value of the fresh sample and the quantity of the fresh sample, calculating a weighted correction predicted value of the stale sample according to the storage time predicted value of the stale sample and the quantity of the stale sample, and calculating a discrimination threshold according to an average value of the weighted correction predicted values of the fresh sample, the quantity of the fresh sample, the average value of the weighted correction predicted values of the stale sample, the quantity of the stale sample and a correction coefficient, so as to judge the freshness degree of the fruit sample; when the weighted correction predicted value of the fruit sample correction set or the external verification set is smaller than the discrimination threshold, judging that the sample is a fresh sample; when the weighted correction predicted values of the fruit sample correction set and the external verification set are greater than or equal to the discrimination threshold, judging that the sample is an stale sample; evaluating the established model according to the fruit sample correction set and the judgment accuracy rate of the external verification set for concentrating the fresh samples and the stale samples, and judging the effectiveness of the model;
C. under the same experiment condition, acquiring the spectral data of the fruit sample to be tested, obtaining the storage time predicted value of the fruit sample to be tested by using the effective mathematical model in the step B, and comparing the predicted value with the corresponding actual storage time of the fruits in the fresh, stale and critical periods so as to judge the freshness degree of the fruit sample to be tested;
D. b, obtaining a storage time predicted value of the critical-period sample by using the effective mathematical model in the step B, and calculating a weighted correction predicted value of the critical-period sample according to the storage time predicted value of the critical-period sample, the number of fresh samples, the number of stale samples and the total number of the samples in the correction set; when the predicted value of the critical period sample is stale and the predicted result exceeds half of the number of the samples in the current day, the predicted value is determined as the original shelf life T0The original shelf life T0Multiplying by a correction coefficient H, and then rounding the obtained product downwards to obtain the shelf life T of the fruit;
the fruit sample to be detected in the step C and the fruit samples in the steps A, B and D are the same type of fruit and have the same storage environment;
the calculation method of the weighted correction predicted value of the fresh sample in the step B is shown as the formula (1):
in the formula (1), yFRCiWeighted correction of the predicted value, y, for the ith fresh sampleFiPrediction of storage time for the model for the ith fresh sample, nFIs to correct the number of fresh samples in the set, nTThe total number of samples in the calibration set;
the calculation method of the weighted correction predicted value of the stale sample is shown as the formula (2):
in the formula (2), yRRCjWeighted correction of the predicted value, y, for the jth stale sampleRjStorage time prediction for the model for the jth stale sample, nRIs to correct the number of stale samples in the set, nTThe total number of samples in the calibration set;
the method for calculating the discrimination threshold is shown as formula (3):
in the formula (3), Eps is a discrimination threshold, yFRCmWeighted correction of the mean value of the predicted values for fresh samples, yRRCmThe average value of the weighted correction predicted values of the stale samples is used, and a is a correction coefficient;
when the fruit is an apple, the value range of the correction coefficient a is 1.0-2.0;
when the fruit is Fuji apple, Eps is 7.8517;
when the fruit is Wanglin apple, Eps is 8.8297;
the calculation method of the weighted correction predicted value of the critical period sample in the step D is shown as the formula (4):
in the formula (4), yRCCkWeighted correction predictor, y, for the kth critical phase samplekPrediction of the storage time of the kth critical phase sample for the model, nFIs to correct the number of fresh samples in the set, nRIs to correct the number of stale samples in the set, nTThe total number of samples in the calibration set;
the calculation method of the fruit shelf life is shown as the formula (5):
T＝floor(T0x H) formula (5)
In the formula (5), T is the shelf life of the fruit0Taking the shelf life as the original, H as a correction coefficient, floor as a downward rounding operator;
when the fruit is apple, the value range of H is 0.70-1.00.
2. The method according to claim 1, wherein when the fruit is Fuji apple, the correction coefficient a is in a range of 1.4-1.7; when the fruit is Wanglin apple, the value range of the correction coefficient a is 1.0-1.3.
3. The method according to claim 2, wherein when the fruit is Fuji apple, the correction coefficient a is in a range of 1.5-1.6; when the fruit is Wanglin apple, the value range of the correction coefficient a is 1.1-1.2.
4. The method according to claim 3, wherein when the fruit is Fuji apple, the correction factor a has a value in the range of 1.55; when the fruit is Wanglin apple, the value range of the correction coefficient a is 1.16.
5. The method according to claim 1, wherein when the fruit is Fuji apple, H has a value in the range of 0.78-0.92;
and when the fruit is a Wanglin apple, the value range of H is 0.75-0.95.
6. The method according to claim 5, wherein when the fruit is Fuji apple, H has a value in the range of 0.80 to 0.90;
and when the fruit is a Wanglin apple, the value range of H is 0.79-0.89.
7. The method according to claim 6, wherein when the fruit is Fuji apple, H has a value in the range of 0.82;
when the fruit is Wanglin apple, the value range of H is 0.81.
8. The method of claim 1, wherein the spectral data is selected from the group consisting of ultraviolet, visible, near infrared, mid infrared, fluorescent, and terahertz spectral data represented in the form of an absorption spectrum, an absorption coefficient spectrum, a transmission spectrum, and a reflection spectrum;
the mathematical model is established by combining a regression algorithm with an interactive verification algorithm, wherein the regression algorithm is selected from multiple linear regression, principal component regression, partial least square regression, neural network regression and support vector machine regression;
the environmental data comprises temperature and relative humidity data of a fruit storage environment;
the fruit comprises apple;
when the fruit is apple, the temperature range of the storage environment of the apple is 19.7-23.2 ℃; the relative humidity of the environment ranges from 10% to 28%.
9. The method of claim 8, wherein the spectral data is near infrared absorption spectral data;
the regression algorithm is partial least squares regression;
the fruit is Fuji apple or Wanglin apple.
10. The method of claim 8, wherein the temperature of the apple storage environment is in the range of 21.8 ℃; the relative humidity range of the environment is 16%.
11. The method according to claim 1, wherein the sample volume ratio of the calibration set to the external validation set in step a is 7: 1-2.5: 1; and/or
The calibration is concentrated, the number of fresh samples nFAnd number of stale samples nRThe ratio satisfies the following conditions: n is more than or equal to 0.892F/nRN is not more than 1.125 and not more than 0.892R/nF≤1.125。
12. The method according to claim 1, wherein when the fruit is apple, taking the day of picking as day 0, preliminarily classifying the samples from day 0 to day 14 as fresh samples, preliminarily classifying the samples from day 15 to day 21 as critical period samples, and preliminarily classifying the samples from day 22 and above as stale samples according to actual conditions and actual storage time of the apple.
13. The method according to claim 1, wherein the method of obtaining spectral data of a fruit sample in step a comprises: calculating time from the day of picking fruits, wherein the data acquisition times are not less than 25 times within 25-32 days after picking, and the spectrum data are acquired at most once every day;
when the spectral data is near-infrared absorption spectral data, the data acquisition method comprises the following steps: a near-infrared spectrometer is utilized, and a polytetrafluoroethylene white board is taken as a background; single integration time 50 ms; accumulating for 50 times and taking an average value; the wavelength range is 901.841 nm-1700.930 nm, and the spectrum center resolution is 8.00 nm-12.00 nm; the number of the spectral variables is 128, and the interval between the spectral variables is 4.882 nm-7.883 nm.
14. The method of claim 13, wherein the spectral center resolution is 9.36 nm; the spectral variables are spaced 6.292nm apart.
15. The method of claim 1, wherein the removing outliers of step a comprises: adopting data centralization pretreatment combined with partial least square regression algorithm, performing pre-modeling by taking spectral data as independent variable and storage time as dependent variable, calculating pre-modeling residual error, namely the difference between a pre-modeling predicted value and an actual value, and determining a pre-modeling residual error judgment threshold value of an abnormal value according to the pre-modeling residual error;
wherein the pre-modeling residual error judgment threshold value of the abnormal value is set to 47.0-53.0;
samples with pre-modeling residuals greater than or equal to the pre-modeling residual determination threshold of the set abnormal value are regarded as abnormal values and should be rejected.
16. The method of claim 15, wherein the pre-modeled residual decision threshold for the outlier is set to 49.5-50.4.
17. The method of claim 16, wherein the pre-modeled residual decision threshold for outliers is 50.0 when the fruit is Fuji apple and 50.2 when the fruit is Wanglin apple.
18. An apparatus for implementing the method of any one of claims 1-17, wherein the apparatus comprises an optical system, a control system, circuitry, and a data storage and processing system;
the optical system is used for collecting the spectral data of the sample;
the circuit system is used for stably supplying power to the device;
the control system is used for controlling the working process of the device;
the data storage and processing system is used for data storage, mathematical model calling, result prediction, storage and output.
19. Use of the method of any one of claims 1 to 17 or the apparatus of claim 18 for predicting the freshness and/or shelf life of fruit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910803691.1A CN110411957B (en) | 2019-08-28 | 2019-08-28 | Nondestructive rapid prediction method and device for shelf life and freshness of fruits |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910803691.1A CN110411957B (en) | 2019-08-28 | 2019-08-28 | Nondestructive rapid prediction method and device for shelf life and freshness of fruits |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110411957A CN110411957A (en) | 2019-11-05 |
| CN110411957B true CN110411957B (en) | 2021-11-19 |
Family
ID=68368990
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910803691.1A Active CN110411957B (en) | 2019-08-28 | 2019-08-28 | Nondestructive rapid prediction method and device for shelf life and freshness of fruits |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110411957B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9069725B2 (en) | 2011-08-19 | 2015-06-30 | Hartford Steam Boiler Inspection & Insurance Company | Dynamic outlier bias reduction system and method |
| CA3116974A1 (en) | 2014-04-11 | 2015-10-15 | Hartford Steam Boiler Inspection And Insurance Company | Improving future reliability prediction based on system operational and performance data modelling |
| US11636292B2 (en) | 2018-09-28 | 2023-04-25 | Hartford Steam Boiler Inspection And Insurance Company | Dynamic outlier bias reduction system and method |
| US11328177B2 (en) | 2019-09-18 | 2022-05-10 | Hartford Steam Boiler Inspection And Insurance Company | Computer-based systems, computing components and computing objects configured to implement dynamic outlier bias reduction in machine learning models |
| US11288602B2 (en) | 2019-09-18 | 2022-03-29 | Hartford Steam Boiler Inspection And Insurance Company | Computer-based systems, computing components and computing objects configured to implement dynamic outlier bias reduction in machine learning models |
| US11615348B2 (en) | 2019-09-18 | 2023-03-28 | Hartford Steam Boiler Inspection And Insurance Company | Computer-based systems, computing components and computing objects configured to implement dynamic outlier bias reduction in machine learning models |
| CN111060473B (en) * | 2020-01-15 | 2021-06-25 | 王丽娟 | Food quality analysis detection device |
| CN111624317A (en) * | 2020-06-22 | 2020-09-04 | 南京农业大学 | Nondestructive testing method for judging freshness of baby cabbage |
| CN111932190B (en) * | 2020-09-30 | 2021-02-02 | 北京每日优鲜电子商务有限公司 | Article information display method, apparatus, device and computer readable medium |
| CN112906939B (en) * | 2021-01-18 | 2022-08-09 | 齐鲁工业大学 | Method for predicting fig harvesting time point |
| CN115271602B (en) * | 2022-08-01 | 2023-06-20 | 深圳市昂捷信息技术股份有限公司 | Intelligent fruit and vegetable warehouse management method and device and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103353445A (en) * | 2013-07-22 | 2013-10-16 | 洛阳农林科学院 | Technical method for quickly identifying drought resistance of wheat by using near-infrared spectroscopy |
| CN106324011A (en) * | 2016-08-25 | 2017-01-11 | 江南大学 | United detection method for determinming freshness of prepared aquatic product at low temperature shelf life |
| WO2018023280A1 (en) * | 2016-07-31 | 2018-02-08 | 赵晓丽 | Information push method for food expiration reminding technology and refrigerator |
| CN108663367A (en) * | 2018-03-30 | 2018-10-16 | 中国农业大学 | A kind of egg quality lossless detection method based on egg unit weight |
| CN109839358A (en) * | 2019-01-22 | 2019-06-04 | 北京农业质量标准与检测技术研究中心 | Analyzing The Quality of Agricultural Products method and device |
-
2019
- 2019-08-28 CN CN201910803691.1A patent/CN110411957B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103353445A (en) * | 2013-07-22 | 2013-10-16 | 洛阳农林科学院 | Technical method for quickly identifying drought resistance of wheat by using near-infrared spectroscopy |
| WO2018023280A1 (en) * | 2016-07-31 | 2018-02-08 | 赵晓丽 | Information push method for food expiration reminding technology and refrigerator |
| CN106324011A (en) * | 2016-08-25 | 2017-01-11 | 江南大学 | United detection method for determinming freshness of prepared aquatic product at low temperature shelf life |
| CN108663367A (en) * | 2018-03-30 | 2018-10-16 | 中国农业大学 | A kind of egg quality lossless detection method based on egg unit weight |
| CN109839358A (en) * | 2019-01-22 | 2019-06-04 | 北京农业质量标准与检测技术研究中心 | Analyzing The Quality of Agricultural Products method and device |
Non-Patent Citations (2)
| Title |
|---|
| 《酥梨货架期的高光谱成像无损检测模型研究》;李雄 等;《光谱学与光谱分析》;20190815;第39卷(第8期);第2578-2583页 * |
| Non-destructive tests on the prediction of apple fruit flesh firmness and soluble solids content on tree and in shelf life;M. Zude et al.;《Journal of Food Engineering》;20050809;第254-260页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110411957A (en) | 2019-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110411957B (en) | Nondestructive rapid prediction method and device for shelf life and freshness of fruits | |
| CN102818777B (en) | Fruit maturity degree evaluation method based on spectrum and color measurement | |
| CN108663339B (en) | On-line detection method for mildewed corn based on spectrum and image information fusion | |
| CN102564993B (en) | Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method | |
| CN106841103A (en) | Near infrared spectrum detects fruit internal quality method and dedicated test system | |
| CN109839358B (en) | Agricultural product quality analysis method and device | |
| CN103543123A (en) | Infrared spectrum recognition method for adulterated milk | |
| CN109409350B (en) | PCA modeling feedback type load weighting-based wavelength selection method | |
| CN107515203A (en) | The research of near infrared technology quantitative analysis rice single grain amylose content | |
| Srivastava et al. | Data processing approaches and strategies for non-destructive fruits quality inspection and authentication: a review | |
| CN107202761A (en) | The portable detection equipment and detection method of a kind of quick detection fruit internal quality | |
| Wang et al. | Rapid detection of protein content in rice based on Raman and near-infrared spectroscopy fusion strategy combined with characteristic wavelength selection | |
| Nturambirwe et al. | Detecting bruise damage and level of severity in apples using a contactless nir spectrometer | |
| CN110264050B (en) | Agricultural product quality analysis method and analyzer | |
| CN111445469A (en) | Hyperspectrum-based apple damage parameter lossless prediction method after impact | |
| Wang et al. | General model of multi-quality detection for apple from different origins by Vis/NIR transmittance spectroscopy | |
| CN102128805A (en) | Method and device for near infrared spectrum wavelength selection and quick quantitative analysis of fruit | |
| CN108169168A (en) | Test and analyze rice grain protein content mathematical model and construction method and application | |
| CN110672578A (en) | Model universality and stability verification method for polar component detection of frying oil | |
| CN109961179A (en) | A kind of aquatic products quality detecting method and portable Raman device | |
| CN112485238B (en) | Method for identifying turmeric essential oil producing area based on Raman spectrum technology | |
| CN106940292A (en) | Bar denier wood raw material quick nondestructive discrimination method of damaging by worms based on multi-optical spectrum imaging technology | |
| CN105158178A (en) | Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology | |
| CN114970675A (en) | Artificial nose refrigerator food freshness detection system and method based on feature selection | |
| CN201051074Y (en) | Multi-spectrum flesh verdure artificial intelligence measurement system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220712 Address after: 100097 No. 9 middle garden, Shuguang garden, Beijing, Haidian District Patentee after: BEIJING ACADEMY OF AGRICULTURE AND FORESTRY SCIENCES Address before: Room 1011, germ plasm building, Beijing Academy of agriculture and Forestry Sciences, 9 Shuguang Huayuan Middle Road, Haidian District, Beijing 100097 Patentee before: BEIJING RESEARCH CENTER FOR AGRICULTURAL STANDARDS AND TESTING |
|
| TR01 | Transfer of patent right |