CN106290224A - The detection method of bacon quality - Google Patents
The detection method of bacon quality Download PDFInfo
- Publication number
- CN106290224A CN106290224A CN201610616125.6A CN201610616125A CN106290224A CN 106290224 A CN106290224 A CN 106290224A CN 201610616125 A CN201610616125 A CN 201610616125A CN 106290224 A CN106290224 A CN 106290224A
- Authority
- CN
- China
- Prior art keywords
- bacon
- quality
- nitrite
- image
- information
- 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.)
- Pending
Links
- 235000015241 bacon Nutrition 0.000 title claims abstract description 174
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- IOVCWXUNBOPUCH-UHFFFAOYSA-M Nitrite anion Chemical compound [O-]N=O IOVCWXUNBOPUCH-UHFFFAOYSA-M 0.000 claims abstract description 42
- 238000000034 method Methods 0.000 claims abstract description 40
- 238000001228 spectrum Methods 0.000 claims abstract description 37
- 230000003595 spectral effect Effects 0.000 claims abstract description 31
- 230000002068 genetic effect Effects 0.000 claims abstract description 27
- 238000005457 optimization Methods 0.000 claims abstract description 17
- 230000004927 fusion Effects 0.000 claims abstract description 6
- 238000002474 experimental method Methods 0.000 claims description 19
- 235000013372 meat Nutrition 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 17
- 230000001580 bacterial effect Effects 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000004445 quantitative analysis Methods 0.000 claims description 5
- IOVCWXUNBOPUCH-UHFFFAOYSA-N Nitrous acid Chemical class ON=O IOVCWXUNBOPUCH-UHFFFAOYSA-N 0.000 claims description 4
- 230000002950 deficient Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 3
- 229910002651 NO3 Inorganic materials 0.000 claims 1
- 239000000523 sample Substances 0.000 description 45
- 238000012360 testing method Methods 0.000 description 19
- 238000005516 engineering process Methods 0.000 description 11
- 230000001537 neural effect Effects 0.000 description 10
- 238000000513 principal component analysis Methods 0.000 description 7
- 239000000126 substance Substances 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 108090000623 proteins and genes Proteins 0.000 description 6
- 210000000697 sensory organ Anatomy 0.000 description 6
- 238000010183 spectrum analysis Methods 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 4
- 244000005700 microbiome Species 0.000 description 4
- 230000035772 mutation Effects 0.000 description 4
- GETQZCLCWQTVFV-UHFFFAOYSA-N trimethylamine Chemical compound CN(C)C GETQZCLCWQTVFV-UHFFFAOYSA-N 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 239000006101 laboratory sample Substances 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000001953 sensory effect Effects 0.000 description 3
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 2
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 2
- 241000894006 Bacteria Species 0.000 description 2
- SIKJAQJRHWYJAI-UHFFFAOYSA-N Indole Chemical compound C1=CC=C2NC=CC2=C1 SIKJAQJRHWYJAI-UHFFFAOYSA-N 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000000470 constituent Substances 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 102000004169 proteins and genes Human genes 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000004611 spectroscopical analysis Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 206010064571 Gene mutation Diseases 0.000 description 1
- 238000004497 NIR spectroscopy Methods 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 150000001412 amines Chemical class 0.000 description 1
- 229910021529 ammonia Inorganic materials 0.000 description 1
- 230000000844 anti-bacterial effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000009395 breeding Methods 0.000 description 1
- 230000001488 breeding effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 210000000349 chromosome Anatomy 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 238000010353 genetic engineering Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- PZOUSPYUWWUPPK-UHFFFAOYSA-N indole Natural products CC1=CC=CC2=C1C=CN2 PZOUSPYUWWUPPK-UHFFFAOYSA-N 0.000 description 1
- RKJUIXBNRJVNHR-UHFFFAOYSA-N indolenine Natural products C1=CC=C2CC=NC2=C1 RKJUIXBNRJVNHR-UHFFFAOYSA-N 0.000 description 1
- 238000002329 infrared spectrum Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000000813 microbial effect Effects 0.000 description 1
- 230000007483 microbial process Effects 0.000 description 1
- 238000012009 microbiological test Methods 0.000 description 1
- 238000000386 microscopy Methods 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 229910052757 nitrogen Inorganic materials 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 230000005693 optoelectronics Effects 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 235000013594 poultry meat Nutrition 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 230000014860 sensory perception of taste Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000010561 standard procedure Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
This application discloses the detection method of a kind of bacon quality, comprise the following steps: gather high spectrum image characteristic wavelength information and the EO-1 hyperion content of nitrite spectral information of bacon simultaneously, and gather the microscopic image information that bacon microbe colony is total, obtain the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum；Using the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum as the input of RBF kernel function multi-data fusion forecast model, take genetic optimization method, obtain bacon quality output valve；Judge bacon quality according to described bacon quality output valve and carry out quality grading prediction.The present invention can detect bacon quality and classification quickly and accurately.More convenient discriminating bacon could safe edible.
Description
Technical field
The present invention relates to field of food detection, be specifically related to the detection method of a kind of bacon quality.
Background technology
As an emerging detection technique, EO-1 hyperion (Hyper-spectral) detection technique by optoelectronics, optics,
At electronic information, the advanced technology in the field such as Neo-Confucianism, computer science rolls into one, and organically traditional near infrared spectrum
Technology and two dimensional image technology combine.At visible ray and near infrared region, the sensor of EO-1 hyperion has tens to hundreds of
Individual wave band, its spectral resolution is the highest, and in near-infrared 780-2526nm wave-length coverage, its spectral resolution is generally less than
10nm, can be usually reached 2-3nm.Therefore, in order to improve accuracy of detection, EO-1 hyperion detection technique is applied to agricultural product, poultry
Meat product, the quality of food have the biggest application potential with safety detection.
Compared with traditional one-dimensional Near Infrared Spectroscopy Detection Technology, integrate spectrum and the Gao Guang of two kinds of technological merits of image
Spectrum detection technique, can not only detect that the pertinent image information of testee information also includes abundant spectral information.High
Spectrum picture technology why can detectable substance inside and outside quality information, being because image detection can reflect outside object comprehensively
In feature, spectral detection then can detect the information such as inherent physical arrangement and the chemical composition of object.Thus, EO-1 hyperion inspection
Survey technology is a kind of easily operated, novel detection technique that accuracy of detection is high, quick and lossless.Research in recent years shows,
EO-1 hyperion detection technique is applied in the Non-Destructive Testing research fields such as quality of agricultural product, Dynamic Non-Destruction Measurement will be sent out future
Exhibition provides a very important research means.The outstanding feature of EO-1 hyperion detection technique is: spectral resolution is high, can obtain
Obtaining many and narrow continuous spectrums of whole wave band, the most hundreds of of wave band number, spectral resolution can reach nanoscale.Collection of illustrative plates closes
One, high spectrum image contains the abundant triple information of space, radiation and spectrum, and spectral band is many, at a certain spectral band model
Can continuous imaging in enclosing.The high-resolution characteristic that EO-1 hyperion detection is had makes its view data adjacent band be spaced relatively
Narrow, there is wave band overlapping region, therefore spectrum channel is the most discrete but continuous print, each pixel of hyperspectral image data
Extracting a complete high-resolution spectroscopy curve, EO-1 hyperion Cleaning Principle is as shown in the figure.Its appearance solves Traditional Scientific
Field " imaging is without spectrum " and the historical problem of " spectrum not imaging ".
Existing bacon detection technique uses method to be generally divided into sensory test, chemical examination and microbial check.Sense organ is examined
The method of looking into is a kind of method being easiest to, and it distinguishes meat smell, color mainly by olfactory sensation, vision, sense of touch and the sense of taste of people
Pool, viscosity and the change of elasticity, thus identify the hygienic quality of meat, simple and easy to do.But, the result amount of being difficult to of sensory test
Change, there is subjectivity and one-sidedness, even if inspection personnel has enough experiences, a lot of in the case of also be difficult to draw correct knot
Opinion, still needs and carries out lab testing；Physical examination is according to breaks down proteins, lower-molecular substance increases, conductivity, viscosity, guarantor
Meat is weighed in the change of the water yield；Chemical examination be with qualitative or quantitative method measure protein breakdown products, as ammonia, amine,
TVB-N (total volatile basic nitrogen), trimethylamine (TMA), indole etc., weigh the degree of metamorphism of meat.Wherein TVB-N is China's detection
The national standard of freshness of meat, TVB-N value can reflect that Meat changes regularly, green meat, secondary fresh meat and rotten
Difference highly significant between meat, and consistent with sense organ change, it is more objective index, but the method equipment needed thereby is complicated, step
The most loaded down with trivial details, detection the cycle the longest, be difficult to the most quickly detect；Microbiological Test is the angle of micro organism quantity from meat
Its pollution situation and putrid and deteriorated degree are described, conventional method has total number of bacteria and the mensuration of coliform approximate number, takes fresh
Meat tabletting microscopy, is not required to increase bacterium and select to cultivate, and simple to operate, result is rapid.Many countries are from the angle of total number of bacterial colonies
Formulate freshness of meat standard, can relatively accurately detect freshness of meat, but result is affected by sampling sites very big,
Particularly city's pin meat situation that is secondary polluted in butchering transport, sales process is more serious, and therefore sampling point difference is tied
Fruit difference is bigger.In traditional microbial process, due to the separation of antibacterial, cultivate more than time-consuming long (24h), technology requirement
High, it is difficult at the scene inspection to promote the use of.
Summary of the invention
In view of drawbacks described above of the prior art or deficiency, it is desirable to provide one can detect bacon quality quickly and accurately
And the detection method of the bacon quality of classification.
To achieve these goals, the present invention adopts the technical scheme that:
The detection method of a kind of bacon quality, comprises the following steps:
Gather high spectrum image characteristic wavelength information and the EO-1 hyperion content of nitrite spectral information of bacon simultaneously, and adopt
Collection bacon microbe colony sum microscopic image information, obtain the characteristics of image wavelength of described bacon, content of nitrite and
Microbe colony sum；
Using the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum as RBF
The input of (RBF, Radial Basis Function) artificial neural network multi-data fusion forecast model, takes genetic optimization
Method, obtains bacon quality output valve；
Judge bacon quality according to described bacon quality output valve and carry out quality grading prediction.
Compared with prior art, the invention has the beneficial effects as follows:
EO-1 hyperion detection technique is used to obtain bacon high spectrum image characteristic wavelength information and EO-1 hyperion nitrous acid first
Salt content spectral information；And collect, in synchronization, same observation station, the microscopic image information that bacon microbe colony is total.
First by the bacon high spectrum image characteristic wavelength information obtained and EO-1 hyperion content of nitrite spectral information；
And it is the most refreshing as RBF according to the microscopic image information gathering synchronization, the bacon microbe colony of same observation station is total
Through networks most according to the input of fusion forecasting model, take genetic optimization method, obtain bacon quality output valve.
First by country's bacon quality Three Estate standard (one-level, two grades, three grades) by genetic optimization intellectual learning
Algorithm, is sub-divided into four grades (top grade, good level, qualified, defective) again, it is judged that bacon quality to carry out quality grading pre-
Survey.More convenient and safe eats.
Accompanying drawing explanation
By the detailed description that non-limiting example is made made with reference to the following drawings of reading, other of the application
Feature, purpose and advantage will become more apparent upon:
The bacon qualitative data based on genetic optimization RBF network that Fig. 1 provides for the embodiment of the present invention merges flow chart；
The spectral curve that Fig. 2 provides for the embodiment of the present invention；
The high spectrum image of the bacon sample that Fig. 3 provides for the embodiment of the present invention；
The high spectrum image of bacon under the different-waveband that Fig. 4 provides for the embodiment of the present invention；
Principal component analysis (PCA) scatterplot of the different brackets bacon that Fig. 5 provides for the embodiment of the present invention；
The flow process of bacon content of nitrite based on the EO-1 hyperion spectral information detection that Fig. 6 provides for the embodiment of the present invention
Figure；
The structure chart of the genetic optimization RBF neural that Fig. 7 provides for the embodiment of the present invention；
The experiment sample characteristic information figure that Fig. 8 provides for the embodiment of the present invention；
The genetic RBF network topological diagram that Fig. 9 provides for the embodiment of the present invention；
The error training curve figure of institute's established model that Figure 10 provides for the embodiment of the present invention；
The test sample the result figure that Figure 11 provides for the embodiment of the present invention；
The single-point crossover process schematic diagram that Figure 12 provides for the embodiment of the present invention.
Detailed description of the invention
With embodiment, the application is described in further detail below in conjunction with the accompanying drawings.It is understood that this place is retouched
The specific embodiment stated is used only for explaining related invention, rather than the restriction to this invention.It also should be noted that, in order to
It is easy to describe, accompanying drawing illustrate only and invent relevant part.
It should be noted that in the case of not conflicting, the embodiment in the application and the feature in embodiment can phases
Combination mutually.Describe the application below with reference to the accompanying drawings and in conjunction with the embodiments in detail.
See Fig. 1, the detection method of a kind of bacon quality, comprise the following steps:
Gather high spectrum image characteristic wavelength information and the EO-1 hyperion content of nitrite spectral information of bacon simultaneously, and adopt
Collection bacon microbe colony sum microscopic image information, obtain the characteristics of image wavelength of described bacon, content of nitrite and
Microbe colony sum；
Using the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum as RBF
The input of (RBF, Radial Basis Function) artificial neural network multi-data fusion forecast model, takes genetic optimization
Method, obtains bacon quality output valve；
Judge bacon quality according to described bacon quality output valve and carry out quality grading prediction.
The present invention can detect bacon quality and classification quickly and accurately.More convenient discriminating bacon could safe edible.
Preferably, the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum be synchronization,
The information that same observation station is gathered.
Preferably, the high spectrum image characteristic wavelength information of described collection bacon is chosen spectral region 1000～
1450nm。
Preferably, the EO-1 hyperion content of nitrite spectral information of described collection bacon includes:
Gather the high-spectral data of bacon, use preprocessing procedures that high-spectral data is carried out pretreatment, then use
High-spectral data after the bacon nitrite actual content value of physical and chemical experiment acquisition and process sets up training set and checking collects, and adopts
With partial least square method (PLS), the high-spectral data of training set is associated with nitrite actual content value, set up bacon
Content of nitrite forecast model；It is estimated with checking set pair built bacon content of nitrite forecast model overall performance,
The purpose of assessment is that model is carried out error analysis, in order to obtain detection range of error within country's examination criteria allowed band
Model.The Quantitative Analysis Model the most finally built, is carried out unknown bacon sample high-spectral data by Quantitative Analysis Model
Prediction, finally gives content of nitrite quantitative result.
Preferably, the microscopic image information of described collection bacon microbe colony sum includes:
By micro-digital camera, bacon surface bacterial plaque digital image information is shot, on computers to bacon sample
This digital image information carries out image denoising and image enhancement operation, finally carries out bacon bacterial plaque information in the image after processing
Identify and counting.
Preferably, described bacon quality grading is top grade, good level, qualified and defective.
Bacon spectral signature wavelength is screened by the present invention first with high-spectral data image information, determines spectrum
Analyst coverage, uses offset minimum binary (PLS) to set up different sense organ grade bacon discrimination model.Next utilizes high-spectral data light
Spectrum information, content of nitrite actual value combines partial least square method and sets up bacon nitrite detection model, building at model
Use different preprocess methods during Li and contrast its modeling result, test result indicate that employing combination pretreatment mode energy
Enough model overall performances that preferably improves, furthermore, for making up high spectrum image deficiency on fine feature is observed, utilize numeral
Bacon microbe colony sum is added up by image processing techniques, and compared with tradition method of counting, it is the most convenient to operate.Finally
Use genetic optimization RBF network that the spectral signature wavelength of bacon, content of nitrite, microbe colony sum are carried out many data
Merging, after training, model is 93.3% to the correct discrimination power of unknown sample.
The EO-1 hyperion of present invention genetic optimization based on characteristics of image wavelength, nitrite and total plate count RBF nerve net
Analytical technology may be used for bacon detection, it is provided that a kind of new quick, accurate, practical meat detection approach.Used
Theoretical and method is equally applicable to Other Meat series products, has certain realistic meaning.
Below by specific embodiment, the present invention is described further:
A kind of detection method of bacon quality, including:
1. use EO-1 hyperion instrument that bacon is carried out high-spectral data collection, by analyzing bacon high-spectral data, choose
The characteristic wavelength of bacon, uses principal component analysis (PCA) to make a distinction the bacon sample of different qualities.
1.1 test materials and instrument
1.1.1 material
Select the bacon of the market of farm produce and large supermarket's circulation as laboratory sample, the bacon sample of buying is loaded fresh-keeping
Bag, puts in calorstat and preserves.Sample is taken out by 12h on pretreatment, carries out sample sections, portioning etc. after reaching room temperature again
Reason.
2.1.2 instrument and equipment
All experiments are all completed in the laboratory that air permeability is good, and room temperature is maintained at about 20 DEG C, and this experiment is used
Hyperspectral imager is Beijing Zolix Instrument Co., Ltd. GaiaSorter " Gai Ya " EO-1 hyperion sorter system.
1.2 experimental technique
Before gathering experimental data, in order to obtain higher-quality pictorial information, need to preset the collection ginseng of image
Number.Major parameter has: the distance between spectrum camera and testing sample, the transfer rate of mobile platform, image and spectral information
Acquisition parameter.
Specific experiment operating procedure is as follows:
1, the bacon sample of buying is taken out, when its temperature reaches consistent with room temperature, to bacon sample from calorstat
Cutting into slices, slice thickness is uniform.
2, experiment needs before starting GaiaSorter series EO-1 hyperion sorter is preheated 10min, opens computer and runs
SpectraSENS high-spectral data acquisition software, sets up with instrument and is connected, and is set experiment acquisition parameter.
3, the laboratory sample after processing uniformly is placed on black standard plate, puts into GaiaSorter series EO-1 hyperion and divides
Select on the object stage of instrument.
4, after confirming that data acquisition parameters is errorless, starting the collection of high-spectral data, every part of sample collecting deadline is big
About need 40 seconds, repeated acquisition 10 times, the curve of spectrum as this sample of averaging.
5, for guaranteeing to gather the accuracy of data, after each sample data collection completes, need to be by residue on object stage
Matter is cleaned out, and can continue the collection of remaining sample.
1.3 analysis of experimental data
High-spectral data wave band owing to collecting is longer, and quantity of information is relatively big, therefrom chooses and can characterize bacon quality
Characteristic wavelength difficulty is relatively big, thus the pretreatment of high-spectral data is necessary.
The high-spectral data that experiment is gathered, wavelength band is at 900nm～1000nm, original spectrum image such as Fig. 2 and Fig. 3
Shown in.In Fig. 2, abscissa is: spectral wavelength；Vertical coordinate is: spectral intensity.
From the point of view of Fig. 2, original spectrum curve has stronger noise, at 1400nm in 900～1000nm spectral regions
～trend is more mild, without obvious absorption signal in 1700nm spectral region.At 1000nm～1100nm and 1200nm～
There is at 1350nm obvious absworption peak, possess the spectral characteristic of quantitative spectrochemical analysis, therefore this experiment choose 1000～
1450nm spectral region is as spectrum analysis region, for the analysis of bacon quality.
1.4 results and analysis
1.4.1 the selection of characteristic wavelength
In spectrum analysis region, the high spectrum image of bacon under observation different-waveband, select to show bacon feature
Image, using its place wave band as bacon characteristic wavelength.Wherein 1003nm, 1103nm, 1200nm, 1306nm, 1397nm,
At 1450nm, bacon high spectrum image is as shown in Figure 3.
From the point of view of Fig. 4, the high spectrum image of 1003nm and 1103nm wave band can reflect the raw information of bacon clearly,
And distortion phenomenon occurs in the image at its all band, can only see the general profile of bacon sample, thus select 1003nm and
1103nm wave band is as two characteristics of image wavelength of bacon image.
Being obtained 8 characteristic wavelengths altogether by the image information of bacon, wherein 6 is principal component analysis gained sample image feature
Wavelength, 2 wave bands being can show completely in all band sample surface information.8 wave bands are respectively 1003nm, 1103nm,
1139nm, 1159nm, 1273nm, 1306nm, 1373nm, 1467nm.
1.4.2 offset minimum binary distinguishes the foundation of model
Due to all corresponding curve of spectrum of every bit in sample image, for ensureing that the curve of spectrum screened can table
Levying the raw information of sample, facilitate the process of data, this experiment uses ENVI to open bacon image, selects bacon texture the most equal
Even region (60px × 60px), as area-of-interest (ROI), uses ROI instrument therefrom to extract the averaged spectrum of ROI region
The curve of spectrum as this bacon sample.
After bacon sample average spectrum has extracted, averaged spectrum curve wave band, at 900nm～1700nm, still contains a large amount of
Redundancy, and the wavelength chosen according to bacon characteristics of image is respectively 1003nm, 1103nm, 1139nm, 1159nm,
1273nm, 1306nm, 1373nm, 1467nm.When setting up offset minimum binary and differentiating model, spectrum analysis region should include above
8 characteristic wavelengths, therefore select 1000nm～1500nm wavelength band as the spectrum analysis band of discriminating model, exist respectively
This spectrum analysis wavelength band chooses continuous print wave band as analytical data, totally 130 wave bands, then choose 9 sense organs one
Level (good) bacon, two grades of (secondary) bacon of 9 sense organs, three grades of (bad) bacon of 6 sense organs carry out principal component analysis to it, analyze software
Use Unscrambler (multivariate data analysis software) software.Analysis result is as it is shown in figure 5, the coordinate of wherein scatterplot
Axle is respectively the first two main constituent, and wherein first principal component is abscissa, and Second principal component, is vertical coordinate.In figure, label 1-9 is
One-level bacon, 10-18 is two grades of bacon, and 19-24 is three grades of bacon, and the result of analysis can see PC1 from figure
The contribution rate of (principal component 1, main constituent 1) is 72%, PC2 (principal component 2, main one-tenth
Points 2) contribution rate is 19%, and total contribution rate is 91%, shows to use that hyperspectral technique distinguishes on different quality bacon is feasible
Property.
By above principal component analysis, it can be seen that the bacon using hyperspectral technique to distinguish different quality is feasible,
For making model be optimized further, therefore training set sample size should be strengthened, after model training completes, then with verifying set pair
The estimated performance of model is passed judgment on.Different brackets bacon checking collection recognition result is as shown in table 1.As can be seen from the table, test
It is 86.67% that card collects overall discrimination power, and its main error results from the erroneous judgement of firsts and seconds.
Table 1 different bacon quality discrimination model checking collection identification result
2. for determine associating between the nitrite actual content value that the high-spectral data of acquisition obtains with National Standard Method
System, takes partial least square method (PLS) to be fitted data, sets up bacon content of nitrite detection model, use simultaneously
Spectroscopic data is processed by multiple preprocess method, by contrasting corresponding model parameter, passes judgment on various pretreatment.
Select the bacon of the market of farm produce and market of farm produce circulation as laboratory sample, gather its high-spectral data, by reason
Change method obtains bacon content of nitrite actual value, and recycling partial least square method sets up bacon nitrite forecast model,
Finally the overall performance of built forecast model is estimated.
During setting up bacon nitrite forecast model, first have to choose enough representative bacon to be measured
Sample, gathers its high-spectral data, and selects preprocessing procedures that high-spectral data carries out pretreatment, then real with physics and chemistry
High-spectral data after testing the bacon nitrite actual content value of acquisition and processing sets up training set and checking collection, uses partially
The high-spectral data of training set is associated by little square law (PLS) with nitrite actual content value, sets up bacon nitrous acid
Salt content forecast model.Being estimated with checking set pair institute established model overall performance, built Quantitative Analysis Model can be to not
The bacon sample high-spectral data known is predicted, and quantitative result.Overall testing process is as shown in Figure 6.
3. use micro imaging system that bacon bacterial plaque data are acquired, utilize digital image processing techniques to detect bacon
Bacterial plaque sum.
Mainly use computer vision technique, by CCD digital camera to bacon surface bacterial plaque digital image information
Shooting, on computers bacon sample digital image information is carried out denoising, enhancing etc. operates, finally to the figure after processing
In Xiang, bacon bacterial plaque information is identified and counts.
Use method for counting colonies respectively three kinds of grade bacon bacterial plaques are counted, and with tradition physical and chemical experiment result pair
Ratio, count results is as follows:
Table 2 colony counting result
Can obtain from table 2 result, use image procossing mode that bacon bacterial plaque is counted, the phase of the best level bacon
Being 5.8% to error, the relative error of secondary bacon is 2.6%, and the relative error of bad level bacon is 1.9%, and three compares,
The detection error of good level bacon its main cause maximum is because good level bacon also in warranty period, although the inside contains one
Slightly biological, but microorganism is the trickleest, it is impossible to it is recognized by the human eye and causes.Three kinds of different grades of bacon bacterial plaque detections
Average relative error value is 3.4%, and testing result is the most satisfied, why there is error with in bacterial plaque adhesion degree and experiment
The setting of threshold value and form factor has the biggest relation, and this error can be eliminated by changing experiment parameter, total
For the detection error of bacterial plaque number within the allowed band of error it was confirmed use image processing method to bacon bacterial plaque figure
It is feasible as carrying out detecting.
4. bacon spectral information and image information after processing, by genetic optimization RBF neural data fusion
Mode, sets up bacon quality identification model.
See Fig. 7, for the structure chart of genetic optimization RBF neural, use genetic algorithm to carry out excellent to RBF neural
Change, RBF neural can be overcome to be easily trapped into the defect of local optimum, improve the adaptive ability of network model with accurate
Degree.
Test the quality evaluation to bacon and have three indexs, respectively sensory evaluation, content of nitrite and microorganism
Total plate count.On the premise of three kinds of index concerned countries standards, to not exceeding standard, bacon quality is repartitioned, and draws
Dividing result as shown in table 3, the bacon opinion rating after repartitioning meets national standard to a certain extent, its nitrite
Harmful material is belonged to, the bacon that edible nitrite or microbe colony exceed standard, it is possible to can be right with microbe colony
The health of human body of consumer causes damage, thus when utilizing the indices after repartitioning that bacon is evaluated, Ying Xuan
Take the final grade as bacon of the minimum opinion rating between the two.
Table 3 bacon grade repartition rear indices
Bacon qualitative data based on genetic optimization RBF neural merges:
As it is shown in figure 1, experiment obtains two class bacon sample datas, respectively bacon high-spectral data and bacon microgram altogether
As data, obtain 8 bacon characteristics of image wavelength altogether by analyzing image information in high-spectral data, utilize EO-1 hyperion simultaneously
Data light spectrum information sets up bacon content of nitrite offset minimum binary forecast model, gets containing of bacon nitrite
Amount.Image information due to high-spectral data can not clearly show the minute information of bacon sample surfaces, thus this experiment
Use again micro-image system that bacon sample has carried out micro-image and gather the details external to obtain bacon, and profit
By digital image processing techniques, the bacon sample micro-image collected is processed, got the microorganism of bacon sample
Bacterial plaque information.
The sample characteristics information chosen as shown in Figure 8, by 8 high spectrum images characteristic wavelength 1003nm, 1103nm,
The spectral value that 1139nm, 1159nm, 1273nm, 1306nm, 1373nm, 1467nm are corresponding, offset minimum binary nitrous acid content mould
Type predictive value, microbe colony sum is as the input sample of heredity RBF neural, after what input sample was corresponding repartitions
Standard class as tutor's signal, top grade, good level, qualified, defective represent with 1,2,3,4 respectively, training sample totally 60,
One-level bacon, two grades of bacon, each 20 of three grades of bacon, test sample totally 15, network topology structure is input layer totally 10 joints
Point, output layer is 1 node, and hidden node number is equal to training sample number, and each training sample is respectively each hidden layer
The data center of node, thus genetic RBF network topological structure is 10 × 60 × 1.As shown in Figure 9.
Use genetic algorithm that RBF neural is optimized, implement step and be divided into the following steps:
(1) determination of initial population
The size of initial population is directly connected to the hunting zone of genetic algorithm.Under normal circumstances, the scale of population be by
How many decisions of parameter (gene) in network, needed for numerical value is about about 50 to 100, algorithm, chromosome is occurred by random number
Device produces.
(2) coded system of genetic algorithm
Genetic algorithm has two kinds of conventional coded systems, respectively binary coding and floating-point encoding, binary coding
Each chromogene is built into a string of binary characters, owing to the charactor comparison of this coding is simple, only 0 and 1,
Thus encoding, decode, intersect, decode during, the speed of service of algorithm is very fast, but in the face of complicated discrete function
Time, this coded system is difficult to reach required precision.Floating-point encoding mode is the numerical value model according to network parameter (gene)
Enclose, take the value as this gene of the floating number in its in-scope, and the code length of gene has certainly in network
The number of the parameter of plan is identical, thus the genetic algorithm under floating-point encoding mode, precision is higher, it is possible to solve complex model
Modeling problem.By the contrast of two kinds of coded systems, select floating-point encoding that model parameter is encoded.Due to RBF net
Decisive parameter in network has a data center, extension constant and at all levels between weights, then the coded system of network is:
According to the data center of first RBF with extend constant accordingly, the regulation weight of the most at all levels
Order encodes.The structure assuming the input layer × hidden layer × output layer of RBF network is m × h × n, xmIt is expressed as input layer
M-th node, chm, dhmRepresent the regulation weights between input layer m-th node and the h node of hidden layer and extension respectively
Constant, wnhIt is expressed as the regulation weights between h node of hidden layer and output layer the n-th node, for this RBF network, according to
Above-mentioned coded system and reference order, genetic algorithm optimization RBF encodes: c11c12c13...c1md11d12d13…d1mc21c22…
dhmw11w12w13…wnh-2wnh-1wnh
(1) determination of fitness function
Fitness function is for weighing the excellent degree of individual in population, and the expression formula of fitness function f is:
In formula, RMS is expressed as root-mean-square error, and n is expressed as output layer neuron number, and P is the sum of training sample, fi
X () is the output valve of i-th network of samples, yiExpected value for i-th network of samples.
(2) selection opertor
Selecting operation to carry out between individuality and individuality, the rule of selection is different, selects operation the most different, originally
The roulette selection method that experiment uses, the performance using the method network is relatively stable efficiently, is first according to certain proportion
(10%) choose the defect individual in parent population, directly carry out the next generation, then this some individuals is used the side of roulette
Formula carries out matching operation, by intersecting between selected individuality, obtains follow-on remainder (90%).
The selected Probability p of chromosomeiExpression formula is:
In formula, r is the individual sum of population, i=1,2,3, L, r；
(3) crossover operator
Seeing Figure 12, for single-point crossover process schematic diagram, so-called intersection operates, and is that the breeding between individuality and individuality is existing
As, according to certain crossover rule (uniform crossover), individual for two parents character of a certain position or the character of some position are entered
Row exchanges, and obtains new individuality.Crossover probability is typically selected between 0.65 and 0.9.
(4) mutation operator
Gene mutation process in the genetics of mutation operation simulation, this operation mainly has two kinds of purposes, and one is for protecting
Card population multiformity, its two be raising algorithm ability of searching optimum.What this experiment was selected is step-by-step variation, mutation probability
Span is 0.001 to 0.1.
Produce new population by above-mentioned genetic manipulation, repeat population iterative process, until the regulation weight of RBF network reaches
To optimum.
Table 4 test sample model testing result
The initiation parameter of heredity is 100, and crossover probability is 0.7, and mutation probability is 0.1, genetic iteration number of times 15000.
It is 95.0% that accuracy is sentenced in average time of built genetic RBF network, and test accuracy is 93.3%.See Figure 10 and Tu
11, RBF network model result figure after genetic optimization, Figure 10 is the error training curve figure of institute's established model, and Figure 11 is test sample
The result.From the point of view of table 4 the result, the predictive value of genetic optimization RBF neural model and expected value keep one substantially
Causing, model prediction is functional, shows that genetic optimization RBF neural may be used for the classification of bacon quality.
In Figure 10: performance represents error amount；Performance is represents that error amount is；50epochs represents
50 frequency of training；Goal is 0 represents that desired value is 0；Train represents training process；Validation represents checking curve；
Test represents test curve.
In Figure 11: Bacon quality level represents bacon quality level；He number of sample represents sample
This quantity；Xpected outputs represents expection output；Actual output represents actual output.
Above description is only the preferred embodiment of the application and the explanation to institute's application technology principle.People in the art
Member should be appreciated that invention scope involved in the application, however it is not limited to the technology of the particular combination of above-mentioned technical characteristic
Scheme, also should contain in the case of without departing from described inventive concept simultaneously, above-mentioned technical characteristic or its equivalent feature carry out
Combination in any and other technical scheme of being formed.Such as features described above has similar merit with (but not limited to) disclosed herein
The technical scheme that the technical characteristic of energy is replaced mutually and formed.
Claims (6)
1. the detection method of a bacon quality, it is characterised in that comprise the following steps:
Gather high spectrum image characteristic wavelength information and the EO-1 hyperion content of nitrite spectral information of bacon simultaneously, and gather cured
The microscopic image information of meat microbe colony sum, obtains the characteristics of image wavelength of described bacon, content of nitrite and micro-life
Thing total plate count；
Using the most refreshing as RBF to the characteristics of image wavelength of described bacon, content of nitrite and microbe colony sum
Through networks most according to the input of fusion forecasting model, take genetic optimization method, obtain bacon quality output valve；
Judge bacon quality according to described bacon quality output valve and carry out quality grading prediction.
The detection method of bacon quality the most according to claim 1, it is characterised in that the characteristics of image ripple of described bacon
The information that length, content of nitrite and microbe colony sum are synchronization, same observation station is gathered.
The detection method of bacon quality the most according to claim 1, it is characterised in that the high-spectrum of described collection bacon
The spectral region 1000～1450nm chosen as characteristic wavelength information.
The detection method of bacon quality the most according to claim 1, it is characterised in that the EO-1 hyperion of described collection bacon is sub-
Nitrate content spectral information includes:
Gather the high-spectral data of bacon, use preprocessing procedures that high-spectral data is carried out pretreatment, then use physics and chemistry
High-spectral data after the bacon nitrite actual content value of experiment acquisition and process sets up training set and checking collects, and uses inclined
The high-spectral data of training set is associated by method of least square with nitrite actual content value, sets up bacon nitrite and contains
Amount forecast model；Being estimated with checking set pair built bacon content of nitrite forecast model overall performance, final foundation is determined
Component analysis model, is predicted unknown bacon sample high-spectral data by Quantitative Analysis Model, finally gives nitrous acid
Salt content quantitative result.
The detection method of bacon quality the most according to claim 1, it is characterised in that described collection bacon microbe colony
The microscopic image information of sum includes:
By micro-digital camera, bacon surface bacterial plaque digital image information is shot, on computers to bacon sample number
Word image information carries out image denoising and image enhancement operation, is finally identified bacon bacterial plaque information in the image after processing
With counting.
6. according to the detection method of the bacon quality described in any one of claim 1-5, it is characterised in that described bacon quality is divided
Level is top grade, good level, qualified and defective.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610616125.6A CN106290224A (en) | 2016-07-29 | 2016-07-29 | The detection method of bacon quality |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610616125.6A CN106290224A (en) | 2016-07-29 | 2016-07-29 | The detection method of bacon quality |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106290224A true CN106290224A (en) | 2017-01-04 |
Family
ID=57663487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610616125.6A Pending CN106290224A (en) | 2016-07-29 | 2016-07-29 | The detection method of bacon quality |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106290224A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107505283A (en) * | 2017-09-04 | 2017-12-22 | 上海交通大学 | The method of nitrate ion content near infrared ray Secondary salinization soil |
| CN107677635A (en) * | 2017-08-31 | 2018-02-09 | 维沃移动通信有限公司 | The detection method and mobile terminal of a kind of carcinogen |
| CN109295159A (en) * | 2018-10-26 | 2019-02-01 | 北京工商大学 | Sausage quality Intelligent detecting method |
| CN109668843A (en) * | 2019-01-07 | 2019-04-23 | 江苏大学 | A method of the Portable multiple spectrum imaging technique based on mobile phone detects bacon quality |
| CN109883966A (en) * | 2019-02-26 | 2019-06-14 | 江苏大学 | A method of Eriocheir sinensis amount of cure is detected based on multispectral image technology |
| CN109975217A (en) * | 2019-03-26 | 2019-07-05 | 贵阳学院 | Plum soluble solid content value detection method based on Hyperspectral imager |
| CN110021338A (en) * | 2017-11-20 | 2019-07-16 | 中国科学院昆明动物研究所 | A method of the then model prediction biological community structure of the power method based on Hill numbers and time relationship |
| CN110021337A (en) * | 2017-09-29 | 2019-07-16 | 中国科学院昆明动物研究所 | A method of the then scale model predictive microbiology structure of community of the power method based on diversity indices and regional scope relationship |
| CN110210680A (en) * | 2019-06-11 | 2019-09-06 | 北京农业信息技术研究中心 | A kind of fish body Noninvasive Measuring Method of Freshness and device based on temperature change |
| CN111597878A (en) * | 2020-04-02 | 2020-08-28 | 南京农业大学 | BSA-IA-BP-based colony total number prediction method |
| CN112417008A (en) * | 2020-12-07 | 2021-02-26 | 渤海大学 | Machine learning-based fish quality judgment method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09286739A (en) * | 1996-04-23 | 1997-11-04 | Tomoya Satou | Planning and design method of bioactivity control substance |
| CN1995987A (en) * | 2007-02-08 | 2007-07-11 | 江苏大学 | Non-destructive detection method and device for agricultural and animal products based on hyperspectral image technology |
| CN101936894A (en) * | 2010-08-23 | 2011-01-05 | 北京工商大学 | Near infrared spectrum and microscopic adipose cell data fusion-based pork freshness non-destructive testing technology |
| CN101943661A (en) * | 2010-08-23 | 2011-01-12 | 北京工商大学 | Near infrared spectrum and microscopic bacterial plaque area data fusion-based pork freshness non-destructive testing technology |
| CN201852784U (en) * | 2010-09-19 | 2011-06-01 | 北京工商大学 | Meat detector |
| CN103558167A (en) * | 2013-10-31 | 2014-02-05 | 华南理工大学 | Method for rapidly measuring content of sodium chloride in salted meat |
| CN103674864A (en) * | 2013-11-12 | 2014-03-26 | 浙江大学 | Fish water content distribution detection method based on hyperspectral imaging technology |
| CN104897607A (en) * | 2015-06-18 | 2015-09-09 | 北京工商大学 | Food modeling and rapid detecting integration method and system adopting portable NIRS (near infrared spectroscopy) |
-
2016
- 2016-07-29 CN CN201610616125.6A patent/CN106290224A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09286739A (en) * | 1996-04-23 | 1997-11-04 | Tomoya Satou | Planning and design method of bioactivity control substance |
| CN1995987A (en) * | 2007-02-08 | 2007-07-11 | 江苏大学 | Non-destructive detection method and device for agricultural and animal products based on hyperspectral image technology |
| CN101936894A (en) * | 2010-08-23 | 2011-01-05 | 北京工商大学 | Near infrared spectrum and microscopic adipose cell data fusion-based pork freshness non-destructive testing technology |
| CN101943661A (en) * | 2010-08-23 | 2011-01-12 | 北京工商大学 | Near infrared spectrum and microscopic bacterial plaque area data fusion-based pork freshness non-destructive testing technology |
| CN201852784U (en) * | 2010-09-19 | 2011-06-01 | 北京工商大学 | Meat detector |
| CN103558167A (en) * | 2013-10-31 | 2014-02-05 | 华南理工大学 | Method for rapidly measuring content of sodium chloride in salted meat |
| CN103674864A (en) * | 2013-11-12 | 2014-03-26 | 浙江大学 | Fish water content distribution detection method based on hyperspectral imaging technology |
| CN104897607A (en) * | 2015-06-18 | 2015-09-09 | 北京工商大学 | Food modeling and rapid detecting integration method and system adopting portable NIRS (near infrared spectroscopy) |
Non-Patent Citations (3)
| Title |
|---|
| 刘硕 等: "基于高光谱图像技术的腊肠酸价含量检测", 《食品工业科技》 * |
| 杨昆程 等: "高光谱技术在腊肉亚硝酸盐含量检测的应用", 《科技通报》 * |
| 王昕琨: "基于多数据融合技术的腊肉品质分级方法", 《食品科学》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107677635A (en) * | 2017-08-31 | 2018-02-09 | 维沃移动通信有限公司 | The detection method and mobile terminal of a kind of carcinogen |
| CN107505283A (en) * | 2017-09-04 | 2017-12-22 | 上海交通大学 | The method of nitrate ion content near infrared ray Secondary salinization soil |
| CN110021337A (en) * | 2017-09-29 | 2019-07-16 | 中国科学院昆明动物研究所 | A method of the then scale model predictive microbiology structure of community of the power method based on diversity indices and regional scope relationship |
| CN110021337B (en) * | 2017-09-29 | 2021-06-25 | 中国科学院昆明动物研究所 | Method for predicting microbial community structure based on power law scale model of diversity index and regional scope relation |
| CN110021338B (en) * | 2017-11-20 | 2021-09-17 | 中国科学院昆明动物研究所 | Method for predicting microbial community structure based on Hill numbers and time relation power law model |
| CN110021338A (en) * | 2017-11-20 | 2019-07-16 | 中国科学院昆明动物研究所 | A method of the then model prediction biological community structure of the power method based on Hill numbers and time relationship |
| CN109295159A (en) * | 2018-10-26 | 2019-02-01 | 北京工商大学 | Sausage quality Intelligent detecting method |
| CN109668843A (en) * | 2019-01-07 | 2019-04-23 | 江苏大学 | A method of the Portable multiple spectrum imaging technique based on mobile phone detects bacon quality |
| CN109883966B (en) * | 2019-02-26 | 2021-09-10 | 江苏大学 | Method for detecting aging degree of eriocheir sinensis based on multispectral image technology |
| CN109883966A (en) * | 2019-02-26 | 2019-06-14 | 江苏大学 | A method of Eriocheir sinensis amount of cure is detected based on multispectral image technology |
| CN109975217A (en) * | 2019-03-26 | 2019-07-05 | 贵阳学院 | Plum soluble solid content value detection method based on Hyperspectral imager |
| CN110210680A (en) * | 2019-06-11 | 2019-09-06 | 北京农业信息技术研究中心 | A kind of fish body Noninvasive Measuring Method of Freshness and device based on temperature change |
| CN111597878A (en) * | 2020-04-02 | 2020-08-28 | 南京农业大学 | BSA-IA-BP-based colony total number prediction method |
| CN112417008A (en) * | 2020-12-07 | 2021-02-26 | 渤海大学 | Machine learning-based fish quality judgment method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106290224A (en) | The detection method of bacon quality | |
| Zhao et al. | Application of invasive weed optimization and least square support vector machine for prediction of beef adulteration with spoiled beef based on visible near-infrared (Vis-NIR) hyperspectral imaging | |
| CN103278464B (en) | Flesh of fish detection method and device | |
| CN108663339B (en) | On-line detection method for mildewed corn based on spectrum and image information fusion | |
| Xing et al. | Combination of chemometric tools and image processing for bruise detection on apples | |
| Shao et al. | Determination of the bruise degree for cherry using Vis-NIR reflection spectroscopy coupled with multivariate analysis | |
| CN107064047A (en) | A kind of Fuji apple quality damage-free detection method based near infrared spectrum | |
| Dixit et al. | A global calibration model for prediction of intramuscular fat and pH in red meat using hyperspectral imaging | |
| CN108875913A (en) | A kind of matsutake Fast nondestructive evaluation system and method based on convolutional neural networks | |
| CN106706546A (en) | Analysis method for artificial intelligence learning materials on basis of infrared and Raman spectrum data | |
| Tello et al. | Application of 2D and 3D image technologies to characterise morphological attributes of grapevine clusters | |
| Xu et al. | Recent trends of multi-source and non-destructive information for quality authentication of herbs and spices | |
| CN106092990A (en) | A kind of three-dimensional fluorescence spectrum discrimination method of lycium barbarum | |
| CN107219184A (en) | A kind of meat discrimination method and device traced to the source applied to the place of production | |
| Liu et al. | Rapid identification of chrysanthemum teas by computer vision and deep learning | |
| Chen et al. | Fast detection of cumin and fennel using NIR spectroscopy combined with deep learning algorithms | |
| Zhang et al. | Visual detection of apple bruises using AdaBoost algorithm and hyperspectral imaging | |
| Huang et al. | Assessment of tomato maturity in different layers by spatially resolved spectroscopy | |
| Ríos-Reina et al. | Application of hierarchical classification models and reliability estimation by bootstrapping, for authentication and discrimination of wine vinegars by UV–vis spectroscopy | |
| Saleh et al. | Palm oil classification using deep learning | |
| Gomes et al. | Prediction of sugar content in port wine vintage grapes using machine learning and hyperspectral imaging | |
| Sarkar et al. | Supervised learning aided multiple feature analysis for freshness class detection of Indian gooseberry (Phyllanthus emblica) | |
| CN103278467A (en) | Rapid nondestructive high-accuracy method with for identifying abundance degree of nitrogen element in plant leaf | |
| Huang et al. | Quality inspection of nectarine based on hyperspectral imaging technology | |
| Jian et al. | Predicting bruise susceptibility in apples using Vis/SWNIR technique combined with ensemble learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |
|
| RJ01 | Rejection of invention patent application after publication |