CN105158178B - Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area - Google Patents
Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area Download PDFInfo
- Publication number
- CN105158178B CN105158178B CN201510644358.2A CN201510644358A CN105158178B CN 105158178 B CN105158178 B CN 105158178B CN 201510644358 A CN201510644358 A CN 201510644358A CN 105158178 B CN105158178 B CN 105158178B
- Authority
- CN
- China
- Prior art keywords
- navel orange
- hyperion
- sample
- collection
- illustrative plates
- 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.)
- Expired - Fee Related
Links
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a kind of navel orange pol based on EO-1 hyperion through-transmission technique spectral peak area to detect rapid modeling method, is related to fruit internal quality technical field of nondestructive testing.This law is：1. obtain navel orange sample half transmitting EO-1 hyperion collection of illustrative plates；2. utilize the pol value of chemical gauging navel orange sample；3. choose navel orange sample average EO-1 hyperion collection of illustrative plates；4. calculate EO-1 hyperion collection of illustrative plates spectral peak area；5. establishing navel orange sample glucose prediction model, Quality Detection is carried out.The present invention gathers navel orange sample EO-1 hyperion collection of illustrative plates based on hyperspectral technique by half transmitting mode can effectively obtain navel orange inside quality information, improve the detection level and detection efficiency of fruit internal quality；The modeling modeling efficiency is high, accuracy rate is high, and model calculation speed is fast, can quickly detect the pol Internal quality index of fruit, the inside quality of fruit is evaluated.
Description
Technical field
The present invention relates to fruit quality technical field of nondestructive testing, more particularly to one kind to be based on EO-1 hyperion through-transmission technique spectral peak
The navel orange pol detection rapid modeling method of area.
Background technology
China is Production of fruit big country, and yield ranks the first in the world throughout the year, accounts for the 19% of world's fruit total output.Fruit
Industry development is rapid, and per the trend for presenting liter every year, fruit industry has been developed as the branch of many local peasant's extra earnings in China
Column industry, to agricultural and economy greatly facilitates and development function.Although China is fruit industry production big country, on the whole
Export volume still very little, not fruit powerful country of commerce.Citrus fruit is the big fruit of the first in the world, and the current orange yield in China exists
The 3rd is occupied between countries in the world（Account for 10.8%）, front two is Brazil respectively（Account for 23.7%）And the U.S.（Account for 15.9%）；To mesh
Before untill, China's fruit exit amount is also not less than the 4% of fruit year total output.At the same time, with internal people's level of consumption
Raising, the consumption status of consumer evolves to the requirement of confrontation by the requirement to product volume, and when buying fruit, people are not only
Only focus on the external sort of fruit, for the inside quality of fruit, component especially in such as pol, mouthfeel and nutritional ingredient
The requirement of matter also accordingly improves.In order to promote fruit to market, product increment is improved, commercial treatment is carried out to fruit i.e. to fruit
Reprocessed and reprocessed after harvesting, be to realize fruit quality, lift the market competitiveness, expand outlet, improve application
The indispensable means with operator's economic benefit；Wherein how to carry out Fast nondestructive evaluation to fruit inside quality is carry out one
The important step of serial fruit commercial treatment.
Due to machine vision technique and spectral technique have the advantages that it is quick, lossless and reliable, it is lossless in agricultural product at present
It is used widely in detection.High light spectrum image-forming technology is a kind of integration technology of image and spectrum, can obtain detection at the same time
The space of object and spectral information.View data and spectroscopic data are combined, can be good at combine surface have be directed to
Being obtained to interior of articles physical arrangement and chemical composition information for property, has good application prospect.It is but thick to navel orange class
For severe edema due to hypofunction of the spleen fruit, it is more difficult to obtain internal information, while high spectrum image information data volume is larger, adds the difficulty of data processing
And the time.
The content of the invention
In order to solve navel orange skin depth, internal information hardly possible obtains, hyperspectral image data amount is big, based on high spectrum image skill
It is difficult to accurately quickly detect the problems such as its pol during the quick online detection of art, the present invention provides one kind to be based on EO-1 hyperion
The navel orange pol detection rapid modeling method of through-transmission technique spectral peak area, establishes prediction model to navel orange pol, effectively extracts bloom
Spectrum information, improves modeling efficiency and accuracy of detection.
The object of the present invention is achieved like this：
Since navel orange pol is to weigh an important indicator of navel orange quality, transmission method can both make and navel orange pol phase
The inside spectral information of pass is effectively obtained, without causing navel orange internal injury because light source power is excessive；Utilize
High-spectrum spectral peak area carries out navel orange pol differentiation, and the modeling process of less spectral variables complexity, calculating speed is fast, accuracy rate
It is higher, it can meet the requirement to navel orange pol Fast nondestructive evaluation.
Specifically, this method comprises the following steps：
1. obtain navel orange sample half transmitting EO-1 hyperion collection of illustrative plates
Using bloom spectrometer detection navel orange sample EO-1 hyperion collection of illustrative plates, setting acquisition mode, the time for exposure, light source power,
Wave-length coverage, resolution ratio and picking rate；
2. utilize the pol value of chemical gauging navel orange sample
According to the pol value of assay method measure navel orange sample described in national standard GB/T8210；
3. choose navel orange sample average EO-1 hyperion collection of illustrative plates
Navel orange mean height spectrum atlas is chosen, it is bent to spectrum under MATLAB environment according to navel orange high-spectrum spectral property
Line carries out Pretreated spectra, remove non-targeted information, instrument noise, background interference and remove the irrelevant variable information such as water peak it
Afterwards, simplify spectral information, retain important information；
4. calculate EO-1 hyperion collection of illustrative plates spectral peak area
Under MATLAB environment, the curve of spectrum is fitted, crest value and valley value under adaptive chosen spectrum region,
Pass through integral and calculating EO-1 hyperion collection of illustrative plates spectral peak area；
5. establishing navel orange sample glucose prediction model, Quality Detection is carried out
The input of the ratio between left and right spectrogram spectral peak area as model becomes under by the use of the adaptive chosen spectrum region of navel orange sample
Amount, carries out pol Quality Detection to unknown navel orange sample, linear regression is established to navel orange pol and quantifies detection model, utilizes prediction
Collect related coefficient and forecast set root-mean-square error to evaluate the precision of detection model.
The present invention has following advantages and good effect：
1. by gathering the half transmitting EO-1 hyperion of navel orange sample, it is difficult to so as to effectively solve the thicker spectroscopic methodology of Peel of Navel Orange
Obtain the difficult point of inside quality；
2. using high-spectrum spectral peak area and the dependency relation of navel orange sample pol change, reduce high-spectral data amount
Greatly to the influence of model calculation speed；
3. the rapid modeling method can meet the requirement of high speed pachydermia class fruit on-line checking, completion obtains in a short time
Effective spectral signal simultaneously establishes regression model, improves detection efficiency and accuracy of detection.
In short, the present invention navel orange sample EO-1 hyperion collection of illustrative plates is gathered by half transmitting mode based on hyperspectral technique can be effective
Navel orange inside quality information is obtained, improves the detection level and detection efficiency of fruit internal quality；Modeling modeling efficiency height,
Accuracy rate is high, and model calculation speed is fast, can quickly detect the pol Internal quality index of fruit, to the inside quality of fruit into
Row evaluation.
Brief description of the drawings
Fig. 1 is the structure diagram of the present apparatus, in figure：
0-navel orange sample,
1-computer, 2-babinet, 3-hyperspectral imager, 4-electricity driving displacement platform,
5-hand-operated lifting platform, 6-light source, 7-sample room；
The step of Fig. 2 is this law is schemed；
Fig. 3 is the half transmitting averaged spectrum curve of single navel orange sample；
Fig. 4 is the navel orange curve of spectrum after curve matching；
Fig. 5 is the related coefficient figure of navel orange sample modeling collection glucose prediction value and measured value；
Fig. 6 is the related coefficient figure of navel orange sample forecast set glucose prediction value and measured value.
Embodiment
Describe in detail with reference to the accompanying drawings and examples：
First, device
Such as Fig. 1, the present apparatus includes target --- navel orange sample 0；
It is provided with computer 1, babinet 2, hyperspectral imager 3, electricity driving displacement platform 4, hand-operated lifting platform 5, light source 6 and sample room
7；
Its position and connection relation are：
Bottom in babinet 2 is provided with electricity driving displacement platform 4, and the top of electricity driving displacement platform 4 is provided with hand-operated lifting platform
5, light source 6 is provided with sample room 7, navel orange sample 0 is placed with above hand-operated lifting platform 5；
Top in babinet 2 is provided with hyperspectral imager 3；
Electricity driving displacement platform 4 and hyperspectral imager 3 are connected with the computer 1 outside babinet 2 respectively.
Above-mentioned each functional component is general part.
Its working mechanism is：Computer 1 and hyperspectral imager 3 are opened, connects the light source 6 of sample room 5, is preheated 30 minutes,
The collection speed that time for exposure, wave-length coverage and resolution ratio and electricity driving displacement platform 4 are set on software is carried in hyperspectral imager 3
Degree and collection position, adjust the object lens of hyperspectral imager 3, obtain the image of clearly navel orange sample 0；At the same time navel orange sample
0 is positioned over hand-operated lifting platform 5, adjusts the position of the navel orange sample 0 and light source 6 in sample room 7, click on software start collection by
Button, electricity driving displacement platform 4 move, and complete the collection of navel orange sample high spectrum image.
2nd, method
Such as Fig. 2, this law comprises the following steps：
1. obtain navel orange sample half transmitting EO-1 hyperion collection of illustrative plates A
Using the EO-1 hyperion collection of illustrative plates of bloom spectrometer detection navel orange sample, acquisition mode is half transmitting mode, spectroscopic light source 4
Small cup 50W halogen lamp, light source general power be 200W, and setting time for exposure 100ms, wave-length coverage is 300~1100nm, picking rate
For 0.2cm/s, resolution ratio 32cm-1；The transmission high-spectrum spectrum information at navel orange sample equator position is gathered, it is automatic respectively to obtain
Pixel size is the spectral value in the image range of 120 × 120pixels in the middle part of image；
2. utilize the pol value B of chemical gauging navel orange sample
According to the pol value of assay method measure navel orange sample described in national standard GB/T8210；
3. choose navel orange sample average EO-1 hyperion collection of illustrative plates C
Navel orange sample average EO-1 hyperion collection of illustrative plates is chosen, according to navel orange high-spectrum spectral property, to light in MATLAB softwares
Modal data carries out S-G and smoothly pre-processes, and removes non-targeted information, instrument noise, background interference and removes the unrelated changes such as water peak
After measuring information, simplify spectral information, retain important information；
4. calculate EO-1 hyperion collection of illustrative plates spectral peak area D
The curve of spectrum is fitted in MATLAB softwares, adaptively chooses wave crest under 390~1055nm SPECTRAL REGIONs
Value and valley value, pass through integral and calculating EO-1 hyperion collection of illustrative plates spectral peak area；
5. establishing navel orange sample glucose prediction model, Quality Detection E is carried out
The ratio between spectrogram spectral peak area under 390~1055nm SPECTRAL REGIONs is adaptively chosen by the use of navel orange sample and is used as model
Input variable, carries out pol Quality Detection to unknown navel orange sample, linear regression is established to pol and quantifies detection model, using pre-
Collection related coefficient and forecast set root-mean-square error are surveyed to evaluate the precision of detection model.
3rd, result of the test
By taking the high spectrum image collection of illustrative plates of navel orange sample as an example, after carrying out integral and calculating by high-spectrum spectral peak area, build
Vertical navel orange pol detection model, and model inspection precision is analyzed.
The step of Fig. 2 is this law is schemed, and Fig. 3 is the half transmitting averaged spectrum curve of single navel orange sample, spectral region 390
~1055nm；Navel orange sample high-spectrum spectrum information is gathered, acquisition mode is half transmitting mode, and spectroscopic light source is 4 50W halogens
Lamp, light source general power are 200W, set time for exposure 100ms, wave-length coverage is 300~1100nm, picking rate 0.2cm/
S, resolution ratio 32cm-1；The transmission high-spectrum spectrum information at navel orange sample equator position is gathered, it is automatic respectively to obtain in the middle part of image
Pixel size is the average light spectrum in the image range of 120 × 120pixels；Spectrum is gathered, according to national standard GB/T8210
The assay method measures navel orange sample pol value, and pol value scope is 9.6~12.6OBrix；By all samples according to 2：1
Ratio cut partition is modeling collection and forecast set, and the pol value of modeling collection sample includes the pol value of forecast set sample.
The navel orange pol fast modeling method, picture is chosen in ENVI softwares from navel orange high spectrum image centre
Plain size is the image of 120 × 120pixels, and extracts the averaged spectrum of each image, and spectral region is 390~1055nm.
Gained navel orange sample spectra progress Savitzky-Golay is smoothly pre-processed, curve is carried out in Matlab softwares
It is fitted and obtains the wave crest and wave trough position of spectrum in curve, calculates the peak face of two peak regions in navel orange EO-1 hyperion collection of illustrative plates
Product.
The ratio between left and right peak area in gained collection of illustrative plates will be calculated and, as the input variable of navel orange glucose prediction, establish linear regression
Model；It is respectively 0.312 to establish linear model RMSEC and RMSEPOBrix and 0.306OBrix, both relatively build at this time
The related coefficient of mould collection and forecast set is respectively 0.832 and 0.846.
Fig. 5,6 are linear model watermelon glucose prediction value and the related coefficient figure of measured value；It can be seen from the figure that built
Linear model obtained preferably prediction result.
Claims (1)
1. a kind of navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area, it is characterised in that including under
Row step：
1. obtain navel orange sample half transmitting EO-1 hyperion collection of illustrative plates (A)
Using the EO-1 hyperion collection of illustrative plates of bloom spectrometer detection navel orange sample, setting acquisition mode, time for exposure, light source power, wavelength
Scope, resolution ratio and picking rate；
2. utilize the pol value (B) of chemical gauging navel orange sample
According to the pol value of assay method measure navel orange sample described in national standard GB/T8210；
3. choose navel orange sample average EO-1 hyperion collection of illustrative plates (C)
Choose navel orange mean height spectrum atlas, according to navel orange high-spectrum spectral property, under MATLAB environment to the curve of spectrum into
Row Pretreated spectra, remove non-targeted information, instrument noise, background interference and remove water peak irrelevant variable information after,
Simplify spectral information, retain important information；
4. calculate EO-1 hyperion collection of illustrative plates spectral peak area (D)
Under MATLAB environment, the curve of spectrum is fitted, crest value and valley value under adaptive chosen spectrum region, pass through
Integral and calculating EO-1 hyperion collection of illustrative plates spectral peak area；
5. establishing navel orange sample glucose prediction model, Quality Detection (E) is carried out
Under by the use of the adaptive chosen spectrum region of navel orange sample the ratio between left and right spectrogram spectral peak area as model input variable, it is right
Unknown navel orange sample carries out pol Quality Detection, and linear regression is established to navel orange pol and quantifies detection model, utilizes forecast set phase
Relation number and forecast set root-mean-square error evaluate the precision of detection model；
1. its device used is the step：
Bottom in babinet (2) is provided with electricity driving displacement platform (4), and the top of electricity driving displacement platform (4) is provided with sample room
(7), light source (6) is provided with sample room (7), navel orange sample (0) is placed with above hand-operated lifting platform (5)；
Top in babinet (2) is provided with hyperspectral imager (3)；
The computer (1) of electricity driving displacement platform (4) and hyperspectral imager (3) respectively with babinet (2) outside is connected；
1. its parameter set is the step：
Using the EO-1 hyperion collection of illustrative plates of bloom spectrometer detection navel orange sample, acquisition mode is half transmitting mode, and spectroscopic light source is 4
50W halogen lamp, light source general power are 200W, set time for exposure 100ms, wave-length coverage is 300~1100nm, and picking rate is
0.2cm/s, resolution ratio 32cm-1；The transmission high-spectrum spectrum information at navel orange sample equator position is gathered, it is automatic respectively to obtain figure
The spectral value in image range for being 120 × 120pixels as middle part pixel size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510644358.2A CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510644358.2A CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105158178A CN105158178A (en) | 2015-12-16 |
| CN105158178B true CN105158178B (en) | 2018-04-20 |
Family
ID=54799113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510644358.2A Expired - Fee Related CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105158178B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108872085B (en) * | 2018-05-09 | 2023-04-25 | 南京林业大学 | Navel orange heart rot nondestructive testing method based on blue wave band transmission image detection |
| CN108846203A (en) * | 2018-06-13 | 2018-11-20 | 深圳市华慧品牌管理有限公司 | The method and device of fruit non-destructive testing |
| CN111795932B (en) * | 2020-06-15 | 2022-11-15 | 杭州电子科技大学 | Hyperspectrum-based nondestructive testing method for sugar acidity of waxberry fruits |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW200504347A (en) * | 2003-07-25 | 2005-02-01 | Taoyuan Distr Agricultural Res And Extension Station | Photoelectric fruit sugar degree detecting technique |
| WO2005108956A1 (en) * | 2004-05-10 | 2005-11-17 | Hiroshi Maeda | On-line internal quality examining method and device |
| CN201233362Y (en) * | 2008-06-04 | 2009-05-06 | 江西农业大学 | Multi optical spectrum imaging device for detecting fruit quality |
| CN103472031A (en) * | 2013-09-20 | 2013-12-25 | 华东交通大学 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
| CN103822879B (en) * | 2014-02-24 | 2016-08-17 | 西北农林科技大学 | A kind of Fructus actinidiae chinensis based on high light spectrum image-forming technology expand fruit lossless detection method |
-
2015
- 2015-10-08 CN CN201510644358.2A patent/CN105158178B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN105158178A (en) | 2015-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101308086B (en) | Fruit internal quality on-line checking apparatus based on near infrared spectra technology | |
| Li et al. | Recent advances in nondestructive analytical techniques for determining the total soluble solids in fruits: a review | |
| CN105044021B (en) | A kind of mid-autumn crisp jujube pol lossless detection method | |
| CN104062263B (en) | The near-infrared universal model detection method of light physical property close fruit quality index | |
| CN103674864B (en) | A kind of fish water content distribution detection method based on high light spectrum image-forming technology | |
| CN106000914A (en) | Apple sorting device for on-line nondestructive test on moldy core | |
| CN104977258B (en) | The quality detecting methods such as tealeaves/cosmetics based on two-dimensional correlation spectra | |
| CN101949686A (en) | Online nondestructive testing (NDT) method and device for comprehensive internal/external qualities of fruits | |
| Yu et al. | Nondestructive determination of SSC in Korla fragrant pear using a portable near-infrared spectroscopy system | |
| CN106841083A (en) | Sesame oil quality detecting method based on near-infrared spectrum technique | |
| CN108956545B (en) | Fruit internal quality nondestructive testing model establishing method and system | |
| CN108846203A (en) | The method and device of fruit non-destructive testing | |
| CN106872370A (en) | Anthocyanin content method for measuring in a kind of red bayberry based on EO-1 hyperion | |
| CN104965973B (en) | A kind of Apple Mould Core multiple-factor Non-Destructive Testing discrimination model and method for building up thereof | |
| CN104280349A (en) | Method for identifying hollowness of white radishes based on hyperspectral image | |
| CN105092579A (en) | Mango quality non-destructive testing device | |
| CN107515203A (en) | The research of near infrared technology quantitative analysis rice single grain amylose content | |
| CN105158178B (en) | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area | |
| CN110646407A (en) | Method for rapidly detecting content of phosphorus element in aquatic product based on laser-induced breakdown spectroscopy technology | |
| CN105044024A (en) | Method for nondestructive testing of grape berries based on near infrared spectrum technology | |
| CN205762405U (en) | Sorting unit for online Non-Destructive Testing Apple Mould Core equipment | |
| CN103487422A (en) | Cloud traditional Chinese medicine quality detection system of multi-wavelength LED fluorescence spectrum and method | |
| CN110487746A (en) | A method of baby cabbage quality is detected based near infrared spectrum | |
| CN101614663A (en) | As seen and near infrared spectrum bee pollen variety discriminating | |
| Wang et al. | Development of near‐infrared online grading device for long jujube |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180420 Termination date: 20181008 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |