CN111044469A - Fruit soluble solid detection method and equipment based on single integrating sphere - Google Patents

Abstract

The invention relates to the technical field of soluble solid detection equipment, and discloses a fruit soluble solid detection method and equipment based on a single integrating sphere, wherein the method comprises the steps of firstly obtaining a pericarp and pulp diffuse reflection spectrum and a pericarp and pulp diffuse reflection reference spectrum of a target fruit, and obtaining a total reflectivity spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the target fruit; and inputting the total reflectance spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit. The method utilizes the influence of the reflection spectrum information of fruit peel and pulp tissues on the near-infrared detection model of the soluble solid matters of the fruits, improves the stability and the applicability of the model, provides a new detection idea for the detection of the content of the soluble solid matters of the fruits in actual production, and improves the prediction precision of the soluble solid matters.

Description

Fruit soluble solid detection method and equipment based on single integrating sphere

Technical Field

The invention relates to the technical field of soluble solid detection equipment, in particular to a fruit soluble solid detection method and equipment based on a single integrating sphere.

Background

The soluble solids content of fruit is one of the important indicators for evaluating its internal quality and ripeness, and in recent years, merchants and consumers are increasingly concerned about this indicator because it is directly related to the sales of fruit and the purchase willingness of consumers. As one of the most widely and mature spectrum techniques, the near infrared spectrum has the advantages of rapidness, no loss, easy maintenance, no need of pretreatment of samples and the like, so that the near infrared spectrum is widely applied to the nondestructive detection of the internal quality (content of soluble solids, acidity, hardness, dry matter and the like) of fruits, but the near infrared spectrum technique is limited by the fact that the near infrared spectrum technique is essentially an empirical technique, the principle of the technique is to establish a relationship between a large amount of spectrum information and corresponding fruit tissue quality parameters through the lambert's law, the lambert's law is a relationship describing the strength of absorption of a substance to light with a certain wavelength and the concentration and thickness of a light absorbing substance, and the lambert's law is limited by the factors such as the concentration, purity and uniformity of the light absorbing substance.

However, the internal tissue of the fruit is a complex and non-uniform medium, and the light and the tissue in the fruit simultaneously contain the interaction of absorption and scattering characteristics, while the existing near infrared spectrum technology ignores the influence of the interaction between the light and the fruit tissue through the method of the lambert's law, thereby having an influence on the prediction result of the soluble solids of the fruit.

Disclosure of Invention

The embodiment of the invention provides a fruit soluble solid detection method and equipment based on a single integrating sphere, which are used for solving the problem that the existing near infrared spectrum technology ignores the interaction between light and fruit tissues and causes large deviation of a prediction result.

The embodiment of the invention provides a fruit soluble solid detection method based on a single integrating sphere, which comprises the following steps:

acquiring a peel and pulp diffuse reflection spectrum and a peel and pulp diffuse reflection reference spectrum of a target fruit;

obtaining a total reflectivity spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum of the target fruit and the pericarp and pulp diffuse reflection reference spectrum;

and inputting the total reflectance spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

The preset soluble solid content prediction model is obtained through the following steps:

acquiring a peel and pulp diffuse reflection spectrum and a peel and pulp diffuse reflection reference spectrum of a first correction set, and obtaining a total reflectivity spectrum of the first correction set based on the peel and pulp diffuse reflection spectrum and the peel and pulp diffuse reflection reference spectrum of the first correction set, wherein the first correction set comprises a fruit sample for establishing a model;

and acquiring the measured value of the soluble solid content of the first correction set, and establishing a soluble solid content prediction model by utilizing a partial least square algorithm based on the measured value of the soluble solid content of the first correction set and the total reflectivity spectrum within the preset waveband range of the first correction set.

The calculation formula of the total reflectivity spectrum specifically comprises the following steps:

wherein R'_cR 'is the pericarp pulp diffuse reflectance spectrum, R'_rThe reference spectrum is the diffuse reflection spectrum of the fruit and peel, and the reference spectrum D is the dark spectrum collected after the light source is closed and the light inlet hole of the spectrometer is covered.

After obtaining the total reflectance spectrum of the first calibration set, before establishing a soluble solid content prediction model by using a partial least squares algorithm based on the measured value of the soluble solid content of the first calibration set and the total reflectance spectrum within a preset waveband range of the first calibration set, the method further includes:

screening a plurality of first characteristic wavelengths from the total reflectivity spectrum within the preset waveband range of the first correction set through a competitive adaptive re-weighting algorithm, and extracting the total reflectivity spectrum at the first characteristic wavelength from the total reflectivity spectrum of the first correction set to be used as a first total reflectivity prediction spectrum; and the first total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

The obtaining of the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the first correction set, and obtaining the total reflectance spectrum of the first correction set based on the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the first correction set, further includes:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of the first correction set, and obtaining a pulp absorption coefficient spectrum of the first correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the first correction set;

screening a plurality of second characteristic wavelengths from pulp absorption coefficient spectra within a preset waveband range of the first correction set through a competitive adaptive re-weighting algorithm, and extracting total reflectivity spectra at the first characteristic wavelength and the second characteristic wavelength from total reflectivity spectra of the first correction set to serve as second total reflectivity prediction spectra; and the second total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

The embodiment of the invention also provides a fruit soluble solid detection method based on the single integrating sphere, which comprises the following steps:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of a target fruit;

obtaining a pulp absorption coefficient spectrum of the target fruit based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the target fruit;

and inputting the pulp absorption coefficient spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

The preset soluble solid content prediction model is obtained through the following steps:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of a second correction set, and obtaining a pulp absorption coefficient spectrum of the second correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the second correction set, wherein the second correction set comprises a fruit sample for modeling;

and acquiring the measured value of the soluble solid content of the second correction set, and establishing a soluble solid content prediction model by using a partial least square algorithm based on the measured value of the soluble solid content of the second correction set and the pulp absorption coefficient spectrum in the preset waveband range of the second correction set.

Wherein the obtaining of the pulp absorption coefficient spectrum of the second correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the second correction set specifically comprises:

the pulp transmittance spectrum and the pulp reflectance spectrum were calculated according to the following formulas,

wherein, T_cIs the pulp transmittance spectrum, T_rFor the flesh transmission reference spectrum, R_cIs the pulp reflectance spectrum, R is the pulp diffuse reflectance spectrum, R_rThe reference spectrum is the pulp diffuse reflection spectrum, and D is a dark spectrum collected after a light source is closed and a light inlet hole of a spectrometer is covered;

subjecting the pulp transmittance spectrum T_cPulp reflectance spectrum R_cWith scattering anisotropy factor g into open sourceCalculating in reverse multiplication algorithm to obtain pulp absorption coefficient spectrum mu_aWherein the scattering anisotropy factor g is constant.

Wherein, after the obtaining of the pulp absorption coefficient spectrum of the second correction set, before the establishing of the soluble solid content prediction model by using the partial least square algorithm based on the measured value of the soluble solid content of the second correction set and the pulp absorption coefficient spectrum within the preset waveband range of the second correction set, the method further comprises:

screening a plurality of third characteristic wavelengths from the pulp absorption coefficient spectra in the preset waveband range of the second correction set through a competitive adaptive re-weighting algorithm, and extracting the pulp absorption coefficient spectra at the third characteristic wavelengths from the pulp absorption coefficient spectra of the second correction set to serve as pulp absorption coefficient prediction spectra; and the pulp absorption coefficient prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

An embodiment of the present invention further provides an electronic device, including:

at least one processor; and at least one memory communicatively coupled to the processor, wherein: the memory stores program instructions executable by the processor, which when called by the processor are capable of performing the method as described above.

According to the fruit soluble solid content detection method and device based on the single integrating sphere, provided by the embodiment of the invention, firstly, a pericarp and pulp diffuse reflection spectrum and a pericarp and pulp diffuse reflection reference spectrum of a target fruit are obtained, then, a total reflectivity spectrum of the target fruit is obtained based on the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the target fruit, and finally, the total reflectivity spectrum of the target fruit is input to a preset soluble solid content prediction model, and the soluble solid content of the target fruit is output. In addition, another fruit soluble solid content detection method based on a single integrating sphere is provided in the embodiments of the present invention, the method includes obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum, and a pulp reflection reference spectrum of a target fruit, obtaining a pulp absorption coefficient spectrum of the target fruit based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum, and the pulp reflection reference spectrum of the target fruit, inputting the pulp absorption coefficient spectrum of the target fruit to a preset soluble solid content prediction model, and outputting a soluble solid content of the target fruit. According to the method, the influence of the reflection spectrum information of the fruit peel and pulp tissues and/or the optical characteristic information of the fruit pulp tissues on the near-infrared detection model of the soluble solid matters of the fruits is utilized, the stability and the applicability of the model are improved, a new detection idea is provided for the detection of the content of the soluble solid matters of the fruits in actual production, and the prediction precision of the soluble solid matters is improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.

FIG. 1 is a schematic flow chart of a method for detecting soluble solids in fruit based on a single integrating sphere according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of the structure of an integrating sphere spectrometer used in an embodiment of the present invention;

FIG. 3 is a schematic diagram of the diffuse reflectance spectrum of the peel and pulp of the embodiment of the present invention;

FIG. 4 is a schematic diagram of the collection of reference spectra of the diffuse reflection of the peel and pulp in the embodiment of the invention;

FIG. 5 is a schematic diagram of the collection of a pulp transmission spectrum in an embodiment of the present invention;

FIG. 6 is a schematic diagram of the acquisition of a reference spectrum for pulp transmission in an embodiment of the present invention;

FIG. 7 is a schematic view of the collection of a pulp reflectance spectrum in an embodiment of the present invention;

FIG. 8 is a schematic diagram of the collection of a reference spectrum of pulp reflectance in an embodiment of the present invention;

FIG. 9 is a scattering distribution diagram of the measured value and the predicted value of the soluble solid content obtained by detecting the prediction set by the first total reflectance prediction spectrum in the embodiment of the present invention;

FIG. 10 is a scattering distribution diagram of the measured value and the predicted value of the soluble solid content measured by the second total reflectance spectrum from the prediction set according to the embodiment of the present invention;

FIG. 11 is a schematic flow chart of another method for detecting soluble solids in fruit based on a single integrating sphere in an embodiment of the present invention;

FIG. 12 is a scattering distribution diagram of the measured value and the predicted value of the soluble solid content obtained by detecting the prediction set by the pulp absorption coefficient prediction spectrum in the embodiment of the present invention;

fig. 13 is a schematic structural diagram of an electronic device according to an embodiment of the present invention;

description of reference numerals:

1. an integrating sphere; 2. An end cap; 3. An optical fiber;

4. a spacer block; 5. A sample; 5-1, No. I fruit blocks;

5-2, II fruit blocks; 5-3, No. III fruit blocks; 6. An adapter;

7. a light source conduit; 8. A sample holder; 9. A holder base;

10. a sliding guide rail; 11. A spectrometer; 12. An optical platform;

13. a collimator; 14. Quartz glass;

501. a processor; 502. A communication interface; 503. A memory;

504. a communication bus.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As shown in fig. 1 to 4, a method for detecting soluble solids of a fruit based on a single integrating sphere according to an embodiment of the present invention includes:

step 110: and acquiring a peel and pulp diffuse reflection spectrum and a peel and pulp diffuse reflection reference spectrum of the target fruit.

To facilitate understanding of the specific processes of the embodiments of the present invention, the following examples are given for illustration:

the fruit is explained by taking a pear as an example, the spectrum acquisition equipment adopts an integrating sphere spectrometer, and the structure of the integrating sphere spectrometer is shown in figures 2 and 3 and comprises the following components: the spectrometer comprises an integrating sphere 1, an end cover 2, an optical fiber 3, a spacing block 4, a sample 5, an adapter 6, a light source guide pipe 7, a sample clamping piece 8, a clamping piece base 9, a sliding guide rail 10, a spectrometer 11, an optical platform 12, a collimator 13 and the like, wherein key equipment models and parameters of the integrating sphere spectrometer are shown in table 1.

TABLE 1 Key Equipment models and parameters used in the experiment

The integrating sphere 1 functions as a light diffuser for scattering light emitted from the sampling port inside the integrating sphere 1 very uniformly after multiple reflections inside the integrating sphere 1. The end cap 2 is used for adjusting the collection mode of the integrating sphere. The optical fiber 3 is used for transmitting an emergent light signal of the integrating sphere 1 to the spectrometer 11. A spacer block 4 is mounted at the end of the collimator 13 facing the sample 5 for limiting the separation between the exit point of the light and the sample 5. The adapter 6 is used to connect the light source guide 7 and the collimator 13 so that the light emitted from the light source guide 7 is entirely guided to the collimator 13. The sample clamping piece 8 is used for clamping the sample 5, the clamping piece base 9 is used for adjusting the position of the sample 5 and is matched with the sample clamping piece 8 to adjust the height of the sample 5, and the sliding guide rail 10 is combined with the clamping piece base 9 to adjust the optical fiber 3, the sample 5 and the integrating sphere 1 to be always kept on the same straight line and to be displaced. The optical platform 12 is used for installing the detection system and accessories thereof built by the whole integrating sphere 1 and the spectrometer 11. It should be noted that, besides the single integrating sphere and the spectrometer, an indirect measurement technique of optical parameters represented by time resolution, spatial resolution, frequency domain resolution, etc. may be adopted as long as the diffuse reflectance spectrum of the fruit and peel pulp and the reference diffuse reflectance spectrum of the fruit and peel pulp can be collected.

When performing diffuse reflectance spectrum collection of peel and pulp on a target fruit, as shown in fig. 3, the sample 5 is placed on a tray (not shown) in close proximity to the end cap 2, which functions as a sample holder 8, and both serve to hold the sample 5. The tray can be respectively through lifter, tray base about with remove to adapt to the not target fruit of equidimension, and closely laminate target fruit in the end cover 2 department on the right side of integrating sphere 1 all the time, need reveal the end cover 2 on right side during the measurement. Then, the adapter 6 extends into the integrating sphere 1 from the port on the left side of the integrating sphere 1, and is close to the end cover 2 on the right side, due to the limiting effect of the spacer block 4, the distance between the collimator 13 at the end of the adapter 6 and the sample 5 is constant, the distance between the adapter 6 and the sample 5 in the embodiment is 8mm, and the distance can be properly adjusted according to the size requirement of the diameter of the light spot projected on the sample, and the limitation is not made here. Then, a light source is started, and the spectrum acquired by the spectrometer is the diffuse reflection spectrum R' of the peel and pulp of the target fruit.

When the reference spectrum of the diffuse reflection of the peel and pulp of the target fruit is collected, as shown in fig. 4, the sample 5 is removed, the end cover 2 is covered, the position relation of other components is kept unchanged, then the light source is turned on, and the spectrum collected by the spectrometer is the reference spectrum R 'of the diffuse reflection of the peel and pulp of the target fruit'_r。

Step 120: and obtaining the total reflectivity spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum of the target fruit and the pericarp and pulp diffuse reflection reference spectrum.

Specifically, the total reflectance spectrum may be obtained by comparing the pericarp and pulp diffuse reflectance spectrum with the pericarp and pulp diffuse reflectance reference spectrum, or by comparing the difference between the pericarp and pulp diffuse reflectance spectrum and the dark spectrum with the difference between the pericarp and pulp diffuse reflectance reference spectrum and the dark spectrum.

Step 130: and inputting the total reflectivity spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

The method for detecting soluble solid content of fruit based on single integrating sphere includes the steps of firstly obtaining a pericarp and pulp diffuse reflection spectrum and a pericarp and pulp diffuse reflection reference spectrum of target fruit, then obtaining a total reflectance spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the target fruit, finally inputting the total reflectance spectrum of the target fruit to a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit. The method utilizes the influence of the reflection spectrum information of fruit peel and pulp tissues on the near-infrared detection model of the soluble solid matters of the fruits, improves the stability and the applicability of the model, provides a new detection idea for the detection of the content of the soluble solid matters of the fruits in actual production, and improves the prediction precision of the soluble solid matters.

On the basis of the above embodiment, the preset soluble solid content prediction model is obtained by the following steps:

step 210: obtaining a peel pulp diffuse reflectance spectrum R ' and a peel pulp diffuse reflectance reference spectrum R ' of a first correction set '_rPericarp-pulp diffuse reflectance spectrum R ' and pericarp-pulp diffuse reflectance reference spectrum R ' based on the first correction set '_rObtaining a total reflectance spectrum R 'of the first correction set'_cThe first correction set includes fruit samples used to model.

Firstly, collecting samples, and purchasing 145 pears in a fruit wholesale market of Beijing, wherein the number of the samples of a first correction set and a prediction set is respectively 90 and 30, the first correction set is used for establishing a soluble solid content prediction model, and the prediction set is used for evaluating the established model. The other 25 pears were used as independent validation sets for the validation of the model created.

Then obtaining a peel pulp diffuse reflection spectrum R ' and a peel pulp diffuse reflection reference spectrum R ' of each pear sample in the first correction set '_rThe collecting manner is the same as that of step 110, and is not described herein again. And then a peel pulp diffuse reflectance spectrum R ' and a peel pulp diffuse reflectance reference spectrum R ' based on the first correction set '_rObtaining a total reflectance spectrum R 'of the first correction set'_cThe calculation method is the same as that of step 120, and is not described herein again. Total reflectance spectrum R 'used in this example'_cThe calculation formula is specifically as follows:

wherein R'_cR 'is the pericarp pulp diffuse reflectance spectrum, R'_rThe reference spectrum is the diffuse reflection spectrum of the fruit and peel, and the reference spectrum D is the dark spectrum collected after the light source is closed and the light inlet hole of the spectrometer is covered.

Step 220: and acquiring the measured value of the soluble solid content of the first correction set, and establishing a soluble solid content prediction model by utilizing a partial least square algorithm based on the measured value of the soluble solid content of the first correction set and the total reflectivity spectrum within the preset waveband range of the first correction set.

Specifically, after the spectrum collection is completed, a traditional destructive experiment is immediately carried out to carry out corresponding pulp cutting and juicing on each pear sample of the first correction set, the pear sample is contained in a beaker, the juice is dripped into a PAL-1 type electronic refractometer (ATAGOCo.Ltd., Tokyo, Japan) and corresponding soluble solid content values are recorded until the soluble solid content measurement of all samples is completed. Then, the total reflectivity spectrum within a preset waveband range is screened out from the total reflectivity spectrum of the full waveband of the first calibration set, specifically, the preset waveband range in this embodiment is 900nm to 1350 nm. And finally, establishing a soluble solid content prediction model by using a Partial Least Squares (PLS) algorithm. The fruit soluble solid content prediction model specifically comprises the following steps:

Y_p＝Σk_λA_λ+Z；

wherein, Y_pIs the predicted value of the soluble solid content of the fruit, lambda is the characteristic wavelength, A_λIs the total reflectance value of pericarp and pulp at lambda wavelength, k_λIs A_λZ is a constant.

Further, the acquired raw spectra may also be pre-processed. The pretreatment mode comprises the following steps: smoothing, first derivative calculation, second derivative calculation, or multivariate scatter correction. The embodiment adopts smoothing to preprocess the original spectrum so as to improve the reliability of the spectrum. The original spectrum collected usually inevitably contains random noise of the instrument, and may have problems such as baseline drift, etc., which affect the establishment of the corresponding relationship between the spectrum and the soluble solid matter, and make the reliability and stability of the model worse. In the embodiment, 7-point Savitzky-Golay smoothing (7-SGS) processing is performed on the original spectrum, and subsequent modeling analysis is performed on the preprocessed spectrum data, so that the reliability of the spectrum is improved.

And carrying out modeling analysis and comparison on the spectrum, wherein the modeling method is a Partial Least Squares (PLS). Respectively establishing soluble solid content prediction models by using the total reflectivity spectrums within different wave band ranges and the total reflectivity spectrums before and after pretreatment, predicting the content of the soluble solids of the concentrated pear, respectively comparing the obtained predicted values with the measured value of the content of the concentrated soluble solids, and completing the verification of the spectrum detection model.

Correlation coefficient of correction (r) in evaluation index_c) Prediction correlation coefficient (r)_p) Verifying the correlation coefficient (r)_v) The larger the corrected Root Mean Square Error (RMSEC), the predicted Root Mean Square Error (RMSEP), the smaller the verified Root Mean Square Error (RMSEV), and the smaller the difference, the better the model prediction performance. The predicted results are shown in table 2.

TABLE 2 Total reflectance spectrum based soluble solid content model prediction results for pears at different wavebands

Compared with the model, the model precision after pretreatment is obviously improved compared with the model precision without pretreatment, and the 7-SGS pretreatment method can better remove interference information such as random noise in the original spectrum and has good effectiveness on the prediction of the soluble solid substance model of the pome. In addition, the 900 nm-1350 nm wave band has good prediction results, on one hand, the C-H chemical bond and H related to the prediction of the content of the soluble solid matter in the 900 nm-1350 nm wave band₂The absorption peak of the O group is most abundant, which indicates that the wave band can reflect the content information of the soluble solid matter most, on the other hand, the acquired spectrum signal is influenced by hardware factors such as the performance of a spectrometer, and as shown in table 1, the NIRQuest512 spectrometer adopted by the invention has larger noise acquired in the wave band of 1350nm to 1900nm, and the noise influence still exists after the pretreatment, which is not beneficial to the prediction of the soluble solid matter, so that the wave band of 900nm to 1350nm has good prediction effect. Subsequent analysis therefore only performed soluble solids prediction model analysis for this band.

In table 2, the correlation coefficient r after the preprocessing is shown_pThe model accuracy after pretreatment is obviously improved compared with the model accuracy without pretreatment under the comprehensive consideration.

Based on the above embodiments, the preset soluble solid content prediction model, after step 210 and before step 220, further includes:

screening a plurality of first characteristic wavelengths from the total reflectivity spectrum within a preset waveband range of the first correction set through a competitive self-adaptive re-weighting algorithm, and extracting the total reflectivity spectrum at the first characteristic wavelength from the total reflectivity spectrum of the first correction set to be used as a first total reflectivity prediction spectrum; and the first total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

Specifically, a Competitive adaptive reweighted sampling (CARS for short) is used to optimize the near-infrared transmission spectrum information within a preset waveband range (900nm to 1350nm) of the pome, and 27 first characteristic wavelengths are preferably selected from 279 wavelengths, wherein the selected first characteristic wavelengths include: 915.60nm, 917.30nm, 925.50nm, 927.10nm, 954.90nm, 961.50nm, 966.40nm, 1016.90nm, 1023.40nm, 1075.40nm, 1083.50nm, 1136.80nm, 1153.00nm, 1159.40nm, 1193.20nm, 1204.50nm, 1209.30nm, 1231.80nm, 1235.00nm, 1249.40nm, 1251.00nm, 1259.00nm, 1260.60nm, 1316.50nm, 1319.70nm, 1345.20nm and 1350.00 nm. Then extracting a total reflectance spectrum at a first characteristic wavelength from the total reflectance spectrum of the first correction set to serve as a first total reflectance prediction spectrum, and further establishing a soluble solid content prediction model by using a Partial Least Squares (PLS), wherein the method specifically comprises the following steps:

Y_pEWs1＝a_915.60A_915.60+a_917.30A_917.30+a_925.50A_925.50+a_927.10A_927.10+a_954.90A_954.90+a_961.50A_961.50+a_966.40A_966.40+a_1016.90A_1016.90+a_1023.40A_1023.40+a_1075.40A_1075.40+a_1083.50A_1083.50+a_1136.80A_1136.80+a_1153.00A_1153.00+a_1159.40A_1159.40+a_1193.20A_1193.20+a_1204.50A_1204.50+a_1209.30A_1209.30+a_1231.80A_1231.80+a_1235.00A_1235.00+a_1249.40A_1249.40+a_1251.00A_1251.00+a_1259.00A_1259.00+a_1260.60A_1260.60+a_1316.50A_1316.50+a_1319.70A_1319.70+a_1345.20A_1345.20+a_1350.00A_1350.00+Z₁

wherein, Y_pEWs1For soluble solids established on the basis of a first characteristic wavelength (EWs1)A material content prediction model; a represents the total reflectance value of the peel pulp at the first characteristic wavelength, a represents the regression coefficient at the first characteristic wavelength, and Z₁Is a constant obtained by regression from a soluble solids content prediction model.

In order to verify the prediction accuracy and stability of the established spectrum detection model, based on the content of the above embodiment, as an optional embodiment, after establishing the spectrum detection model based on the characteristic wavelength, the embodiment of the present invention further includes: respectively establishing a soluble solid content prediction model by using a total reflectivity spectrum and a first total reflectivity prediction spectrum (based on the total reflectivity spectrum under a first characteristic wavelength) within a preset waveband range (900 nm-1350 nm), predicting the content of the soluble solids of the concentrated pear, respectively comparing the obtained predicted values with the measured value of the content of the soluble solids of the concentrated pear, and completing the verification of the spectrum detection model. The predicted results are shown in table 3.

TABLE 3 Total reflectance spectra (EWs1), absorption coefficient (. mu.s) based on CARS-PLS_a) Pear soluble solids prediction results for spectra (EWs2) and their combination spectra (EWs1+ EWs2)

In Table 3, EWs is an abbreviation for Effective Wavelengths, i.e., the Effective wavelength. The data of number 1 in table 3 shows the prediction results of a soluble solid content prediction model established using a total reflectance spectrum in a predetermined wavelength range (900nm to 1350nm), and the number of wavelengths used for modeling is 279. The data in the number 3 shows the prediction result of the soluble solid content prediction model established using the first total reflectance prediction spectrum (based on the total reflectance spectrum at the first characteristic wavelength), and the number of wavelengths used for the modeling is 27. Although the PLS model based on the 900 nm-1350 nm wave band has achieved good prediction results, the large number of the wave lengths participating in modeling can cause problems of long time consumption, large computation amount, low applicability and the like in an actual online system. For the model based on CARS optimization, the prediction accuracy of the model is slightly improved, the number of the wavelengths based on the total reflectivity spectrum is reduced from original 279 to 27, the correlation coefficients of a correction set and a prediction set are both greater than 0.86, the root mean square error of the correction set and the prediction set is both less than 0.55 DEG Brix, and the model is simplified, so that the better prediction accuracy is obtained, and the actual detection accuracy requirement is met.

On the basis of the above embodiment, step 210 further includes:

step 211: and obtaining the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the first correction set, and obtaining the pulp absorption coefficient spectrum of the first correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the first correction set.

Specifically, when pulp transmission spectrum collection is performed for the first calibration set, as shown in fig. 5, the sample 5 is placed in the sample holder 8 at the port immediately on the left side of the integrating sphere 1, and the sample 5 is held by two pieces of quartz glass 14 provided on the sample holder 8. Sample 5 here was taken of the whole pome fruit of size II pulp piece 5-2, and size II pulp piece 5-2 was obtained by slicing the whole fruit of fig. 3 from left to right to cut into size I pulp piece 5-1, size II pulp piece 5-2 and size III pulp piece 5-3, respectively, then removing size II pulp piece 5-2 of approximately 5mm thickness shown in the figure, and performing a pulp transmission spectrum T test by cutting the center pulp of the pulp piece with a custom made cube cutter (30 mm. times.30 mm. times.20 mm, length. times.width. times.height). The pome is sliced to ensure that the side surface of the pulp is smooth during detection, and the slice with the thickness of 5mm is selected by comprehensively considering the transmittance and the reflectivity of the pome pulp. Other varieties of fruit may be appropriately adjusted and are not limited herein.

The sample holder 8 can move up and down and left and right through the holder base 9 and the slide rail 10, respectively. When in measurement, the end cover 2 on the right side needs to be covered, the sample clamping piece 8 is utilized to enable the No. II fruit pulp block 5-2 to be tightly attached to the port of the left side of the integrating sphere 1, then the adapter 6 is close to the sample clamping piece 8, and due to the limiting effect of the spacing block 4, the distance between the collimator 13 at the end part of the adapter 6 and the No. II fruit pulp block 5-2 is constant. And then, turning on a light source, wherein the spectrum acquired by the spectrometer is the pulp transmission spectrum T of the first correction set.

When the pulp transmission reference spectrum T is carried out on the first correction set_rDuring collection, as shown in fig. 6, the sample holder 8 is moved away (or lowered), the positional relationship of other components is kept unchanged, then the light source is turned on, and the spectrum collected by the spectrometer is the pulp transmission reference spectrum T of the first calibration set_r。

When the pulp reflectance spectrum is collected in the first calibration set, as shown in fig. 7, the sample holder 8 holding the No. ii pulp piece 5-2 is placed in the end cap 2 on the right side of the integrating sphere 1, and the end cap 2 on the right side is removed for measurement. Then, the adapter 6 extends into the integrating sphere 1 from the port on the left side of the integrating sphere 1 and is close to the end cover 2 on the right side, due to the limiting effect of the spacing block 4, the distance between the collimator 13 at the end of the adapter 6 and the sample 5 is constant, then the light source is turned on, and the spectrum collected by the spectrometer is the pulp diffuse reflection spectrum R of the first correction set.

When the reference spectrum of the diffuse reflection of the pulp is collected in the first calibration set, as shown in fig. 8, the sample holder 8 is removed (or lowered), the end cap 2 is covered, the position relationship of other components is kept unchanged, then the light source is turned on, and the spectrum collected by the spectrometer is the reference spectrum R of the diffuse reflection of the pulp in the first calibration set_r。

The pulp transmittance spectrum and the pulp reflectance spectrum can then be calculated according to the following equations,

wherein, T_cIs the pulp transmittance spectrum, T_rFor the flesh transmission reference spectrum, R_cIs the pulp reflectance spectrum, R is the pulp diffuse reflectance spectrum, R_rThe reference spectrum is the pulp diffuse reflection spectrum, and the dark spectrum is collected after the light source is closed and the light inlet hole of the spectrometer is covered.

The optical properties of biological tissues such as fruits are mainly used for absorptionCoefficient spectrum mu_aAnd reduced scattering coefficient spectrum mu'_sTo characterize, the pulp transmittance spectrum T_cPulp reflectance spectrum R_cAnd substituting the scattering anisotropy factor g into an open source Inverse doubling-doubling (IAD) algorithm for calculation to obtain a pulp absorption coefficient spectrum mu_aAnd reduced scattering coefficient spectrum mu'_s. Wherein the scattering anisotropy factor g is constant, the scattering anisotropy factor g is also called as mean scattering cosine, g of most biological tissues is between 0.7 and 0.95 in the wavelength range of 600nm to 1300nm, g is 1 to represent complete forward scattering, g is 0 to represent isotropic scattering, g is-1 to represent complete backward scattering, g is always equal to the scattering coefficient mu_sAre combined to utilize the formula mu'_s＝μ_s(1-g) expressed as reduced scattering coefficient μ'_sIn the patent of the invention, g is 0.7.

More specifically, the absorption coefficient spectrum μ of pulp_aAnd reduced scattering coefficient spectrum mu'_sAnd 7-SGS preprocessing is carried out, and the preprocessed spectral data is subjected to subsequent modeling analysis, wherein the modeling method is a partial least squares algorithm (PLS). The predicted results are shown in the following table:

TABLE 4 absorption coefficient spectra μ of respective pulps at different wave bands_aAnd reduced scattering coefficient spectrum mu'_sPear soluble solid content model prediction result

From Table 4, it can be seen that_aThe spectrum was predicted to be good,. mu.,. mu.'_sThe prediction effect of the spectrum is poor, so that only the mu is considered subsequently_aThe model built by the spectrum is further optimized.

Step 212: pulp absorption coefficient spectrum mu in a preset waveband range (900 nm-1350 nm) from the first correction set by a competitive adaptive re-weighting algorithm_aMedium screeningA plurality of second characteristic wavelengths are extracted, and total reflectivity spectrums at the first characteristic wavelength and the second characteristic wavelength are extracted from the total reflectivity spectrum of the first correction set and serve as second total reflectivity prediction spectrums; and the second total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm. The predicted results are shown in table 3.

The data of number 2 in Table 3 show the absorption coefficient spectrum μ for pulp within a predetermined wavelength range (900nm to 1350nm)_aThe number of wavelengths used for modeling the prediction result of the established soluble solid content prediction model is 279. Data No. 4 shows the use of the pulp absorption coefficient spectrum μ based on the second characteristic wavelength_aThe number of the wavelengths used for modeling the prediction result of the established soluble solid content prediction model is 14. Specifically, the second characteristic wavelength includes: 900.90nm, 917.30nm, 932.00nm, 948.40nm, 976.10nm, 1070.50nm, 1104.50nm, 1162.60nm, 1165.80nm, 1183.60nm, 1235.00nm, 1241.40nm, 1251.00nm and 1270.20 nm. The data of reference numeral 5 shows the prediction results of the soluble solid content prediction model using the second total reflectance prediction spectrum (based on the total reflectance spectra at the first characteristic wavelength and the second characteristic wavelength), and the number of wavelengths used for modeling was 38, and the number of wavelengths used was 38 in total because 3 effective wavelengths among the 27 first characteristic wavelengths and the 14 second characteristic wavelengths were the same number of wavelengths (specifically, 917.30nm, 1235.00nm, and 1251.00 nm). The soluble solid content prediction model established based on the total reflectivity spectrum under the first characteristic wavelength and the second characteristic wavelength specifically comprises the following steps:

Y_pEWs1+EWs2＝b_900.90A_900.90+b_915.60A_915.60+b_917.30A_917.30+b_925.50A_925.50+b_927.10A_927.10+b_932.00A_932.00+b_948.40A_948.40+b_954.90A_954.90+b_961.50A_961.50+b_966.40A_966.40+b_976.10A_976.10+b_1016.90A_1016.90+b_1023.40A_1023.40+b_1070.50A_1070.50+b_1075.40A_1075.40+b_1083.50A_1083.50+b_1104.50A_1104.50+b_1136.80A_1136.80+b_1153.00A_1153.00+b_1159.40A_1159.40+b_1162.60A_1162.60+b_1165.80A_1165.80+b_1183.60A_1183.60+b_1193.20A_1193.20+b_1204.50A_1204.50+b_1209.30A_1209.30+b_1231.80A_1231.80+b_1235.00A_1235.00+b_1241.40A_1241.40+b_1249.40A_1249.40+b_1251.00A_1251.00+b_1259.00A_1259.00+b_1260.60A_1260.60+b_1270.20A_1270.20+b_1316.50A_1316.50+b_1319.70A_1319.70+b_1345.20A_1345.20+b_1350.00A_1350.00+Z₂

wherein, Y_pEWs1+EWs2A soluble solids content prediction model established for the first characteristic wavelength (EWs1) and the second characteristic wavelength (EWs 2); a represents the total reflectance value of the peel pulp at a first characteristic wavelength (or a second characteristic wavelength), b represents the regression coefficient at the corresponding characteristic wavelength, Z₂Is a constant obtained by regression from a soluble solids content prediction model.

The second total reflectivity prediction spectrum (based on the total reflectivity spectrum under the first characteristic wavelength and the second characteristic wavelength) after CARS optimization has 38 spectra, the correlation coefficients of a correction set and a prediction set are both more than 0.89, the root mean square errors of the correction set and the prediction set are both less than 0.51 DEG Brix, and good prediction results are also obtained. Compared with the results of the independent verification set, the second total reflectivity prediction spectrum obtains a better result, namely the correlation coefficient of the verification set is 0.81, the root mean square error of the verification set is 0.50 DEG Brix, and the latent variable number is less, so that the combined spectrum model is more stable. By utilizing the combined spectral characteristic wavelength, the effective spectral bands of the pericarp pulp layer and the pulp layer are fully considered, and a single pericarp pulp layer model established in a full band can be improved, so that the model is simplified on the premise of ensuring the prediction precision of the model, and the stability and the applicability of the model are improved. The result scatter point distribution of the prediction set sample modeled by the CARS-based optimized first total reflectivity prediction spectrum and the second total reflectivity prediction spectrum is respectively shown in fig. 9 and fig. 10, and the prediction values are closely distributed on two sides of the regression line, so that the linear prediction effect is good. The results show that the modeling method of the first total reflectivity prediction spectrum and the second total reflectivity prediction spectrum can effectively realize near-infrared detection of soluble solids of the pomes, has a better prediction result, and can realize nondestructive detection of the fruits.

As shown in fig. 11, an embodiment of the present invention further provides a method for detecting soluble solids of a fruit based on a single integrating sphere, including:

step 310: and acquiring a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of the target fruit.

Obtaining a pulp transmission spectrum T and a pulp transmission reference spectrum T of each pear sample in a target fruit_rA pulp diffuse reflectance spectrum R and a pulp diffuse reflectance reference spectrum R_rThe collecting method is the same as step 211, and is not described herein again.

Step 320: and obtaining a pulp absorption coefficient spectrum of the target fruit based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the target fruit.

Pulp transmission spectrum T and pulp transmission reference spectrum T based on target fruit_rA pulp diffuse reflectance spectrum R and a pulp diffuse reflectance reference spectrum R_rObtaining the pulp absorption coefficient spectrum mu of the target fruit_aThe calculation method is the same as that in step 211, and is not described herein again.

Step 330: and inputting the pulp absorption coefficient spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

On the basis of the above embodiment, the preset soluble solid content prediction model is obtained by the following steps:

step 410: and acquiring a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of the second correction set, and acquiring a pulp absorption coefficient spectrum of the second correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the second correction set, wherein the second correction set comprises a fruit sample for establishing a model.

Wherein, the second correction set can be the same as or different from the first correction set. The same will be described in this embodiment as an example. Obtaining a pulp transmission spectrum T and a pulp transmission reference spectrum T of each pear sample in the second correction set_rA pulp diffuse reflectance spectrum R and a pulp diffuse reflectance reference spectrum R_rThe collecting manner is the same as that of step 310, and is not described herein again.

Pulp transmission spectrum T and pulp transmission reference spectrum T based on second correction set_rA pulp diffuse reflectance spectrum R and a pulp diffuse reflectance reference spectrum R_rObtaining a pulp absorption coefficient spectrum mu of the second correction set_aThe calculation method is the same as that of step 320, and is not described herein again.

Step 420: and acquiring the measured value of the soluble solid content of the second correction set, and establishing a soluble solid content prediction model by utilizing a partial least square algorithm based on the measured value of the soluble solid content of the second correction set and the pulp absorption coefficient spectrum in the preset waveband range of the second correction set. The fruit soluble solid content prediction model specifically comprises the following steps:

Y_p＝Σm_λB_λ+N；

wherein, Y_pIs the predicted value of the soluble solid content of the fruit, lambda is the characteristic wavelength, B_λIs the pulp absorption coefficient value at lambda wavelength, m_λIs B_λN is a constant.

Specifically, the preset wavelength range in this embodiment is 900nm to 1350 nm. And finally, establishing a soluble solid content prediction model by using a Partial Least Squares (PLS) algorithm. The detailed manner is similar to step 220, and is not described herein again.

More specifically, the acquired raw spectra may also be preprocessed. The pretreatment mode comprises the following steps: smoothing, first derivative calculation, second derivative calculation, or multivariate scatter correction. The embodiment adopts smoothing to preprocess the original spectrum so as to improve the reliability of the spectrum. The original spectrum collected usually inevitably contains random noise of the instrument, and may have problems such as baseline drift, etc., which affect the establishment of the corresponding relationship between the spectrum and the soluble solid matter, and make the reliability and stability of the model worse. In the embodiment, 7-point Savitzky-Golay smoothing (7-SGS) processing is performed on the original spectrum, and subsequent modeling analysis is performed on the preprocessed spectrum data, so that the reliability of the spectrum is improved. Spectrum mu of absorption coefficient of pulp_aAnd reduced scattering coefficient spectrum mu'_sAnd 7-SGS preprocessing is carried out, and the preprocessed spectral data is subjected to subsequent modeling analysis, wherein the modeling method is a partial least squares algorithm (PLS). The predicted results are shown in Table 4.

On the basis of the above embodiment, after step 410, before step 420, further comprising:

screening a plurality of third characteristic wavelengths from the pulp absorption coefficient spectra in the preset waveband range of the second correction set through a competitive adaptive re-weighting algorithm, and extracting the pulp absorption coefficient spectra at the third characteristic wavelengths from the pulp absorption coefficient spectra of the second correction set to serve as pulp absorption coefficient prediction spectra; the pulp absorption coefficient prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

Specifically, a competitive adaptive re-weighting algorithm (CARS) is used to optimize the near-infrared transmission spectrum information within a preset waveband range (900nm to 1350nm) of the pome, and from 279 wavelengths, 14 third characteristic wavelengths are preferably selected (in this embodiment, the second correction set is the same as the first correction set, and thus the third characteristic wavelengths are the same as the second characteristic wavelengths), where the third characteristic wavelengths include: 900.90nm, 917.30nm, 932.00nm, 948.40nm, 976.10nm, 1070.50nm, 1104.50nm, 1162.60nm, 1165.80nm, 1183.60nm, 1235.00nm, 1241.40nm, 1251.00nm and 1270.20 nm. And then extracting a pulp absorption coefficient spectrum at a third characteristic wavelength from the pulp absorption coefficient spectrum of the second correction set to be used as a pulp absorption coefficient prediction spectrum, and further establishing a soluble solid content prediction model by using a partial least squares algorithm (PLS). The method specifically comprises the following steps:

Y_pEWs3＝m_900.90B_900.90+m_917.30B_917.30+m_932.00B_932.00+m_948.40B_948.40+

m_976.10B_976.10+m_1070.50B_1070.50+m_1104.50B_1104.50+m_1162.60B_1162.60+m_1165.80B_1165.80+m_1183.60B_1183.60+m_1235.00B_1235.00+m_1241.40B_1241.40+m_1251.00B_1251.00+m_1270.20B_1270.20+N

wherein, Y_pEWs3A soluble solids content prediction model established based on the third characteristic wavelength (EWs 3); b represents a pulp absorption coefficient value at the third characteristic wavelength, m represents a regression coefficient at the third characteristic wavelength, and N is a constant obtained by regression from a soluble solid content prediction model.

In order to verify the prediction accuracy and stability of the established spectrum detection model, based on the content of the above embodiment, as an optional embodiment, after establishing the spectrum detection model based on the characteristic wavelength, the embodiment of the present invention further includes: respectively establishing a soluble solid content prediction model by utilizing a pulp absorption coefficient spectrum (see data of serial number 2 in table 3) and a pulp absorption coefficient prediction spectrum (see data of serial number 4 in table 3) within a preset waveband range (900 nm-1350 nm), predicting the content of soluble solids of the concentrated pear, respectively comparing the obtained predicted values with the measured values of the content of the concentrated soluble solids, and finishing the verification of the spectrum detection model. From number 4 in Table 3As can be seen from the data, the characteristic wavelength-screened μ is used_aThe spectrum has better prediction capability and better stability, and the result of the independent verification set is better, namely r_vAnd the value of RMSEV may make a strong explanation of this.

The result scatter distribution of the prediction set sample after the prediction spectrum modeling based on the pulp absorption coefficient after CARS optimization is respectively shown in FIG. 12, and the predicted values are distributed close to the two sides of the regression line, so that the linear prediction effect is good.

Based on the CARS optimized model, the prediction accuracy of the model is slightly improved, and meanwhile, the model is based on the absorption coefficient spectrum mu_aThe number of the wavelengths is reduced from 279 to 14, the correlation coefficients of the correction set and the prediction set are both more than 0.86, the root mean square error of the correction set and the prediction set is both less than 0.55 DEG Brix, meanwhile, the correlation coefficient of the independent verification set reaches 0.91, the root mean square error is less than 0.36 DEG Brix, and the prediction effect is very excellent. Therefore, the method obtains better prediction precision while simplifying the model, and meets the requirement of actual detection precision.

Fig. 13 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and referring to fig. 13, the electronic device may include: a Processor (Processor)501, a communication Interface (Communications Interface)502, a memory (memory)503, and a communication bus 504, wherein the Processor 501, the communication Interface 502, and the memory 503 are configured to communicate with each other via the communication bus 504.

The processor 501 may call logic instructions in the memory 503 to perform the following method: the method comprises the steps of firstly obtaining a pericarp and pulp diffuse reflection spectrum and a pericarp and pulp diffuse reflection reference spectrum of a target fruit, then obtaining a total reflectivity spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum and the pericarp and pulp diffuse reflection reference spectrum of the target fruit, finally inputting the total reflectivity spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

In addition, the processor 501 may also call logic instructions in the memory 503 to perform the following method: the method comprises the steps of firstly obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of a target fruit, then obtaining a pulp absorption coefficient spectrum of the target fruit based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the target fruit, finally inputting the pulp absorption coefficient spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

In addition, the logic instructions in the memory 503 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

In another aspect, embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the method for detecting soluble solids in fruit based on a single integrating sphere provided in the above embodiments.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.

Through the embodiment, the spectrum information of six pear tissues including the peel and pulp diffuse reflection spectrum, the peel and pulp diffuse reflection reference spectrum, the pulp transmission reference spectrum, the pulp reflection reference spectrum and the pulp reflection spectrum of the fruit is respectively collected in a single integrating sphere system mode, and a near infrared transmission spectrum detection model of the soluble solid matter of the pear is established by combining a characteristic wavelength screening algorithm. And establishing a more stable soluble solid prediction model by combining the optical characteristics of the fruits and the spectral information of the reflection spectrum.

Further, aiming at the problems of random noise generated by the original spectrum under the corrected spectrum information, large intensity difference of the transmission spectrum and the like, 7-point Savitzky-Golay smoothing processing is carried out to preprocess the original spectrum, so that the prediction accuracy of the soluble solid of the pome under the reflection spectrum and the optical characteristic information is optimized.

Furthermore, based on a prediction model after pretreatment under a 900 nm-1350 nm wave band, near-infrared characteristic wavelength screening of soluble solids of the pomes is carried out, combined spectrum information of the total reflectivity spectrum and the absorption coefficient spectrum is compared in an overlapping mode, modeling analysis is carried out, and a modeling result is optimized.

The invention provides the influence of the optical characteristic information of fruit pulp tissues and the reflection spectrum information of peel pulp tissues on the near-infrared detection model of the soluble solid content of the pome, improves the stability and the applicability of the model, and provides a new detection idea for detecting the content of the soluble solid content of the pome in actual production. The invention combines the selected characteristic wavelength screening method, establishes the soluble solid near infrared spectrum detection model combining the optical characteristic and the reflection spectrum, improves the model established only by a single reflection spectrum sample in the early-stage literature, considers the spectrum information of the fruit peel, the fruit flesh layer and the fruit flesh layer, and improves the prediction precision of the soluble solid. In addition, the invention combines a characteristic wavelength screening method, ensures that the model prediction precision is slightly improved, simplifies the model, and has certain guiding significance for the optimization of the actual detection model.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A fruit soluble solid detection method based on a single integrating sphere is characterized by comprising the following steps:

acquiring a peel and pulp diffuse reflection spectrum and a peel and pulp diffuse reflection reference spectrum of a target fruit;

obtaining a total reflectivity spectrum of the target fruit based on the pericarp and pulp diffuse reflection spectrum of the target fruit and the pericarp and pulp diffuse reflection reference spectrum;

and inputting the total reflectance spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

2. The method for detecting soluble solids in fruit according to claim 1, wherein the predetermined soluble solids content prediction model is obtained by:

acquiring a peel and pulp diffuse reflection spectrum and a peel and pulp diffuse reflection reference spectrum of a first correction set, and obtaining a total reflectivity spectrum of the first correction set based on the peel and pulp diffuse reflection spectrum and the peel and pulp diffuse reflection reference spectrum of the first correction set, wherein the first correction set comprises a fruit sample for establishing a model;

and acquiring the measured value of the soluble solid content of the first correction set, and establishing a soluble solid content prediction model by utilizing a partial least square algorithm based on the measured value of the soluble solid content of the first correction set and the total reflectivity spectrum within the preset waveband range of the first correction set.

3. The method for detecting soluble solids in fruit according to claim 2, wherein the total reflectance spectrum is calculated by the following formula:

wherein R'_cR 'is the pericarp pulp diffuse reflectance spectrum, R'_rThe reference spectrum is the diffuse reflection spectrum of the fruit and peel, and the reference spectrum D is the dark spectrum collected after the light source is closed and the light inlet hole of the spectrometer is covered.

4. The method for detecting soluble solids in fruit according to claim 2, wherein after obtaining the total reflectance spectrum of the first calibration set, before the establishing the predictive model of the soluble solids content by using partial least squares algorithm based on the measured soluble solids content of the first calibration set and the total reflectance spectrum within the predetermined wavelength band of the first calibration set, the method further comprises:

screening a plurality of first characteristic wavelengths from the total reflectivity spectrum within the preset waveband range of the first correction set through a competitive adaptive re-weighting algorithm, and extracting the total reflectivity spectrum at the first characteristic wavelength from the total reflectivity spectrum of the first correction set to be used as a first total reflectivity prediction spectrum; and the first total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

5. The single-integrating-sphere-based fruit soluble solid detection method according to claim 4, wherein the obtaining the pericarp and pulp diffuse reflectance reference spectrum of the first calibration set and the obtaining the total reflectance spectrum of the first calibration set based on the pericarp and pulp diffuse reflectance reference spectrum of the first calibration set further comprises:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of the first correction set, and obtaining a pulp absorption coefficient spectrum of the first correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the first correction set;

screening a plurality of second characteristic wavelengths from pulp absorption coefficient spectra within a preset waveband range of the first correction set through a competitive adaptive re-weighting algorithm, and extracting total reflectivity spectra at the first characteristic wavelength and the second characteristic wavelength from total reflectivity spectra of the first correction set to serve as second total reflectivity prediction spectra; and the second total reflectivity prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

6. A fruit soluble solid detection method based on a single integrating sphere is characterized by comprising the following steps:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of a target fruit;

obtaining a pulp absorption coefficient spectrum of the target fruit based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the target fruit;

and inputting the pulp absorption coefficient spectrum of the target fruit into a preset soluble solid content prediction model, and outputting the soluble solid content of the target fruit.

7. The fruit soluble solid detection method based on the single integrating sphere of claim 6, wherein the predetermined soluble solid content prediction model is obtained by the following steps:

obtaining a pulp transmission spectrum, a pulp transmission reference spectrum, a pulp reflection spectrum and a pulp reflection reference spectrum of a second correction set, and obtaining a pulp absorption coefficient spectrum of the second correction set based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the second correction set, wherein the second correction set comprises a fruit sample for modeling;

and acquiring the measured value of the soluble solid content of the second correction set, and establishing a soluble solid content prediction model by using a partial least square algorithm based on the measured value of the soluble solid content of the second correction set and the pulp absorption coefficient spectrum in the preset waveband range of the second correction set.

8. The method for detecting soluble solids in fruit according to claim 7, wherein the pulp absorption coefficient spectrum of the second calibration set is obtained based on the pulp transmission spectrum, the pulp transmission reference spectrum, the pulp reflection spectrum and the pulp reflection reference spectrum of the second calibration set, and specifically comprises:

the pulp transmittance spectrum and the pulp reflectance spectrum were calculated according to the following formulas,

wherein, T_cIs the pulp transmittance spectrum, T_rFor the flesh transmission reference spectrum, R_cIs prepared from fruitReflectance spectrum of meat, R is diffuse reflectance spectrum of pulp, R_rThe reference spectrum is the pulp diffuse reflection spectrum, and D is a dark spectrum collected after a light source is closed and a light inlet hole of a spectrometer is covered;

subjecting the pulp transmittance spectrum T_cPulp reflectance spectrum R_cSubstituting the scattering anisotropy factor g into an open source reverse multiplication algorithm for calculation to obtain a pulp absorption coefficient spectrum mu_aWherein the scattering anisotropy factor g is constant.

9. The method for detecting soluble solids in fruit according to claim 8, wherein after obtaining the pulp absorption coefficient spectrum of the second calibration set, before establishing the soluble solids content prediction model by using partial least squares algorithm based on the measured soluble solids content of the second calibration set and the pulp absorption coefficient spectrum within the predetermined band range of the second calibration set, the method further comprises:

screening a plurality of third characteristic wavelengths from the pulp absorption coefficient spectra in the preset waveband range of the second correction set through a competitive adaptive re-weighting algorithm, and extracting the pulp absorption coefficient spectra at the third characteristic wavelengths from the pulp absorption coefficient spectra of the second correction set to serve as pulp absorption coefficient prediction spectra; and the pulp absorption coefficient prediction spectrum is used for establishing a soluble solid content prediction model by combining a partial least square algorithm.

10. An electronic device, comprising:

at least one processor; and

at least one memory communicatively coupled to the processor, wherein:

the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1-9.

Images