CN107515198A - A kind of online test method of starch burn degree and gelatinization point - Google Patents
A kind of online test method of starch burn degree and gelatinization point Download PDFInfo
- Publication number
- CN107515198A CN107515198A CN201710672065.4A CN201710672065A CN107515198A CN 107515198 A CN107515198 A CN 107515198A CN 201710672065 A CN201710672065 A CN 201710672065A CN 107515198 A CN107515198 A CN 107515198A
- Authority
- CN
- China
- Prior art keywords
- mrow
- starch
- msubsup
- msub
- mover
- 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
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
- G01N21/21—Polarisation-affecting properties
- G01N21/23—Bi-refringence
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
Landscapes
- Physics & Mathematics (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)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses the online test method of a kind of starch burn degree and gelatinization point.Micro- sem observation is combined by this method with artificial intelligence image recognition technology, and the principle of unique birefringent characteristic is presented under polarized light using starch granules, the Starch SSD methods based on deep learning is proposed, for monitoring the metamorphosis of starch gelatinization process.The inventive method experimental result collimation is good, and measuring result error, compared to human visual observation, not only eliminates the influence of subjective error within 1%, and whole detection process is time-consuming few, and more operated quickly and conveniently, experimental result is more accurate for operation.The inventive method is based on the target tracking to starch birefringent characteristic in gelatinization process, there is provided a kind of new method of online observation and design control starch reaction degree.
Description
Technical field
The invention belongs to Food Science and Engineering technical field, and in particular to a kind of starch burn degree and gelatinization point
Line detecting method.
Background technology
Starch is natural polymer common in nature, can be widely applied to the research fields such as food, material, medicine.
Starch gelatinization process is the melting process of starch crystallite beam, the hydrogen bond rupture failure of starch molecule, the increase of molecule confusion degree,
The irreversible water swelling of grain, in the process its viscosity, light transmittance and birefringence feature and the reactivity to enzyme or chemical reagent
Significantly changed etc. physicochemical characteristics.
Using starch as in the product processing of raw material, the gelatinization degree and gelatinization point of starch are one very important
Judging index, have a strong impact on its processing technology and the quality of product.Using can be to the paste of starch the methods of physics, chemistry and biochemistry
Change characteristic and carry out analysis measure, such as conventional differentia scanning calorimetry, quick viscosity analytic approach, near infrared spectroscopic method
The methods of.These analysis methods respectively have advantage and disadvantage, but never Tongfang can be characterized in face of Starch paste characters.
PLM with heating stage is to observe starch granules structure, study the common instrument of its Effect On Gelatinization Characteristics.Show in polarised light
Under micro mirror observation, bright birefringence feature is presented in starch granules, it is shown that the anisotropy of starch granules.The gelatinization row of starch
Generally to occur within the temperature range of 10 DEG C or so, conventional method judges the gelatinization point of starch by artificial vision's observation
And gelatinization degree, make that experimental result individual subjective differences are big, and usually conclusion is inconsistent, and takes a substantial amount of time.
Deep learning is the development of artificial neural network, it is intended to the process that human visual system handles visual information is simulated,
So that target identification is more intelligent.Deep learning model has multitiered network structure, can effectively overcome traditional images mesh
The shortcomings that mark recognition methods is present.Birefringence is the key character of starch polarized light microscopy picture, utilizes deep learning image procossing
It is to judge starch gelatinization temperature and gelatinization degree quickly and have that technology carries out analyzing detection to birefringence feature during starch gelatinization
The method of effect.
The content of the invention
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to starch gelatinization temperature can quick and precisely be measured by providing one kind
The method of degree and gelatinization degree, the online test method of specially a kind of starch burn degree and gelatinization point.This method is by microscope
Observation is combined with artificial intelligence image recognition technology, and the original of unique birefringent characteristic is presented under polarized light using starch granules
Reason, proposes the Starch-SSD methods based on deep learning, for monitoring the metamorphosis of starch gelatinization process.
The purpose of the present invention is achieved through the following technical solutions.
A kind of online test method of starch burn degree and gelatinization point, the main collection including starch microphoto, form sediment
The identification and starch gelatinization behavior and the assessment of feature of powder birefringent characteristic.
A kind of online test method of starch burn degree and gelatinization point, specifically comprises the following steps:
(1) film-making:Starch milk is prepared, starch milk is drawn using dropper after being uniformly dispersed, is added dropwise on circular slide, is covered
Upper cover glass makes starch milk be uniformly dispersed and be sealed with glass cement, obtains sample chips;
(2) shoot:The sample chips made are placed in thermal station, under petrographic microscope heat up heating, and use with it is inclined
The connected digital camera of light microscope carries out continuous shooting, collecting, gathers polarized light microscopy photo;
(3) polarisation feature marks:Using Matlab softwares, will there is birefringence feature in the polarized light microscopy photo of collection
Starch granules is labeled with rectangular frame;
(4) training sample and polarisation feature recognition:Polarized light microscopy photo using Starch-SSD methods to completion mark
Middle birefringence feature carries out deep learning, starch birefringence property data base is built, then to the starch gelatinization in test sample
The polarized light microscopy photo of process carries out analysis detection;
(5) calculating of starch burn degree and sign:
DG:The gelatinization degree (the Degree of Gelatinization) of starch;
A0:Starch is original, and the particle number of birefringent phenomenon is presented when not heating up in polarized light microscopy photo;
Aq:During the temperature programming of thermal station to q DEG C of specified temp, of birefringent phenomenon is presented in starch polarized light microscopy photo
Grain number;
(6) determination of starch gelatinization temperature:
Starch initial gelatinization temperature T0, terminate gelatinization point TeDetermined by following relational expression:T0=TDG=5%, wherein
TDG=5%Represent temperature during DG%=5%;Te=TDG=95%, wherein TDG=95%Represent temperature during DG%=95%.
Further, in step (1), the starch is included in the ative starch, modified starch and process of all kinds
Starch.
Further, in step (1), the mass fraction of the starch milk is 0.5%-3%, experimental to require selection
Specific concentration, and starch granules is dispersed in field of microscope and spreads out, be overlapped mutually less as far as possible between particle for
It is good.
Further, in step (2), the initial temperature of the heating heating is no more than 40 DEG C, and maximum heating temperature does not surpass
Cross 100 DEG C.
Further, in step (2), the heating rate of the heating heating is less than 5 DEG C/min.
Further, in step (2), the sampling dot frequency of the shooting, collecting is less than 2 DEG C/time, i.e., in the temperature less than 2 DEG C
One is sampled in degree excursion and is shot.
Further, in step (2), the multiplication factor of the digital camera is 50 × 10 or 20 × 10.
Further, in step (4), the polarized light microscopy photo of the training sample is more than 1000.
Further, in step (4), the algorithm of the deep learning is using image recognition technology to starch birefringence
All algorithm routines that feature is detected.
Further, in step (4), the Starch-SSD methods are based on SSD methods, use single deep layer nerve
Network, it is characterized as detecting mesh calibration method with starch birefringence in petrographic microscope photo, core is to use small convolution filter
Carry out classification fraction and the position skew of one group of default boundary frame fixed on predicted characteristics figure.
SSD (Single Shot Multi-Box Detector framework) is a kind of based on propagated forward convolution god
Method through network, for solving the problems, such as target detection (positioning+classification), that is, a testing image is inputted, exports multiple box
Positional information and classification information, the target detection being mainly used in natural image.
Starch-SSD methods are to be based on SSD methods, with reference to polarized light microscope observing, with starch in petrographic microscope photo
Birefringence is characterized as the method for detecting the detection and research Starch paste characters of target.
Starch-SSD target loss function:
Overall target loss function loses conf and position loss loc weighted sum for confidence;
Forecast confidence after standardization,
l:Prediction block;
g:Give tacit consent to frame;
p:Classification number, p=2;
N:The number of the acquiescence frame to match with the starch birefringence feature frame of mark;
Represent the starch birefringence feature of i-th of acquiescence frame and classification p j-th of mark
Frame matches,Represent to mismatch;
I-th of acquiescence frame does not include any classification, then i ∈ neg, i.e. negative sample negtive;Otherwise, i ∈ pos, i.e., positive sample
This positive;
Starch-SSD positioning, to center c=(cx, cy), width w and the height of the starch birefringence feature frame of mark
Degree h is returned, weight α 1:
Compared with prior art, the invention has the advantages that and beneficial effect:
(1) the inventive method experimental result collimation is good, and measuring result error overcomes traditional pass through and shown within 1%
The shortcomings that micro mirror artificial observation or utilization PLM with heating stage measure starch burn degree and gelatinization point;
(2) the inventive method is identified and detected to starch birefringence feature using computer, is seen compared to artificial vision
Examine, not only eliminate the influence of subjective error, and whole detection process is time-consuming is less than 4s (conventional manual's method of identification is generally time-consuming
65~70s, save the substantial amounts of time, more operated quickly and conveniently, experimental result is also more accurate for operation;
(3) the inventive method is based on the target tracking to starch birefringent characteristic in gelatinization process, there is provided a kind of online
Observation and the new method of design control starch reaction degree.
Brief description of the drawings
Fig. 1 is the structural representation of Starch-SSD methods in the specific embodiment of the invention;
Fig. 2 a are the cornstarch birefringence number and heating-up temperature relation curve in embodiment 1 in polarized light microscopy photo
Figure;
Fig. 2 b are Gelatinization Degree of Maize Starch and heating-up temperature graph of relation in embodiment 1;
Fig. 3 a are the farina birefringence number and heating-up temperature relation curve in embodiment 2 in polarized light microscopy photo
Figure;
Fig. 3 b are farina gelatinization degree and heating-up temperature graph of relation in embodiment 2.
Embodiment
Below in conjunction with embodiment, technical scheme is clearly and completely described.Obviously, it is described
Embodiment is only part of the embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, this area
The every other embodiment that those of ordinary skill is obtained under the premise of creative work is not made, belongs to protection of the present invention
Scope.
In the specific embodiment of the invention, the structural representation of Starch-SSD methods is as shown in figure 1, the pixel of input picture
For 768 × 1024, resonance offset effect obtains characteristic pattern by multiple convolutional layers (solid line layer) and pond layer (dotted line layer);
Wherein, Starch-SSD target loss function:
Overall target loss function loses conf and position loss loc weighted sum for confidence;
Forecast confidence after standardization,
l:Prediction block;
g:Give tacit consent to frame;
p:Classification number, p=2;
N:The number of the acquiescence frame to match with the starch birefringence feature frame of mark;
Represent the starch birefringence feature of i-th of acquiescence frame and classification p j-th of mark
Frame matches,Represent to mismatch;
I-th of acquiescence frame does not include any classification, then i ∈ neg, i.e. negative sample negtive;Otherwise, i ∈ pos, i.e., positive sample
This positive;
Starch-SSD positioning, to center c=(cx, cy), width w and the height of the starch birefringence feature frame of mark
Degree h is returned, weight α 1;
Because all starch sizes are basically identical in experimental data and birefringence feature has square shape, therefore use
The bounding box of three kinds of different sizes (8 × 8,16 × 16,32 × 32 pixel) detects to starch birefringence feature.
Embodiment 1
Measure the Effect On Gelatinization Characteristics of common corn starch
(1) film-making:Configuration quality concentration (butt) is 0.5% common corn starch suspension, will be avaled with glass bar
Starch stir, make it disperse more uniformly, to draw starch milk rapidly with dropper and be added dropwise and 1 drip in edge in system
The circular slide center of a circle glass cement is coated, another sheet glass is covered after starch fluid is uniformly dispersed, is sealed;Film-making will
The starch in light microscopy field is asked to sprawl uniformly, it is overlapping less between starch granules;
(2) shoot:The piece made is placed in thermal station, under petrographic microscope heat up heating, using with microscope phase
Digital camera even carries out continuous data acquisition;Microscope magnification is adjusted to 50 × 10, sets petrographic microscope aperture
With the photographic exposure time, make photo exposure amount suitable;The rate of heat addition of thermal station is arranged to 2 DEG C/min, 30 DEG C of initial temperature, highest
100 DEG C of heating temperature;During heating starch, the digital camera being connected with microscope is continuously shot in the suitably visual field,
Capture rate is two per minute, and collection filming frequency is 1 DEG C/time;
(3) polarisation feature marks:Using Matlab softwares, will pass through in pretreated cornstarch polarized light microscopy photo
Starch granules with polarisation feature is labeled with rectangular frame;
(4) training sample and polarisation feature recognition:Cornstarch polarisation using Starch-SSD algorithms to completion mark
Picture birefringence feature carries out deep learning, and training sample is 1000, starch birefringence property data base is built, then to leading
One group of corn starch pasting process polarised light displaing micro picture for entering computer is detected;
(5) calculating of Gelatinization Degree of Maize Starch and sign:
DG:Gelatinization degree at a temperature of cornstarch;
A0:The particle number of birefringent phenomenon during the original polarisation picture of cornstarch is presented;
Aq:When thermal station is warming up to q DEG C, the starch granules of birefringent phenomenon is presented in cornstarch polarized light microscopy photo
Number;
(6) determination of corn starch pasting temperature:
Initial gelatinization temperature T0, terminate gelatinization point TeDetermined by following relational expression:T0=TDG=5%;Te=TDG=95%;Its
Middle TDG=5%Represent temperature during DG%=5%;TDG=95Show temperature during DG%=95%.
Embodiment 2
Measure the Effect On Gelatinization Characteristics of farina
(1) film-making:Configuration quality concentration (butt) is 1% potato starch suspension, the shallow lake that will be avaled with glass bar
Powder stirs, and makes it disperse evenly, to draw starch milk rapidly with dropper and dropwise addition 1 drips to and coats one at edge in system
The circular glass piece center of glass cement is enclosed, another sheet glass is covered after starch fluid is uniformly dispersed, is sealed;Film-making requires light
Learn field of microscope in starch sprawl uniformly, it is overlapping less between starch granules;
(2) shoot:The piece made is placed in thermal station, under petrographic microscope heat up heating, using with microscope phase
Digital camera even carries out continuous data acquisition;Microscope magnification is adjusted to 20 × 10, sets petrographic microscope aperture
With the photographic exposure time, make photo exposure amount suitable;The rate of heat addition of thermal station is arranged to 1 DEG C/min, 35 DEG C of initial temperature, highest
Heating-up temperature is 90 DEG C;During heating starch, the digital camera being connected with microscope is continuously clapped in the suitably visual field
Take the photograph, capture rate is 1 per minute, and collection filming frequency is 1 DEG C/time;
(3) polarisation feature marks:Using Matlab softwares, will have in pretreated cornstarch polarized light microscopy photo
The starch granules of polarisation feature is labeled with rectangular frame;
(4) training sample and polarisation feature recognition:It is inclined to the farina for completing mark using Starch-SSD algorithms
Light picture birefringence feature carries out deep learning, and training sample is 1200, builds starch birefringence property data base, then right
The one group of farina gelatinization process polarised light picture for importing computer is detected;
(5) calculating of farina gelatinization degree and sign:
DG:Gelatinization degree at a temperature of farina;
A0:The particle number of birefringent phenomenon is presented in the original polarisation picture of farina;
Aq:When thermal station is warming up to q DEG C, farina polarisation shows the particle number that birefringent phenomenon is presented in photo;
(6) determination of farina gelatinization temperature:
Starch initial gelatinization temperature T0, terminate gelatinization point TeDetermined by following relational expression:T0=TDG=5%;Te=
TDG=95%;Wherein TDG=5%Represent temperature during DG%=5%;TDG=95%Show temperature during DG%=95%.
Gelatinization degree
The number of starch presentation birefringence feature is drawn using the characteristic of experiment gained in embodiment 1 and embodiment 2
The curve map changed with gelatinization degree with heating-up temperature;
Cornstarch birefringence number and heating-up temperature graph of relation such as Fig. 2 a in embodiment 1 in polarized light microscopy photo
Shown, from Fig. 2 a, when heating-up temperature is higher than 60 DEG C in thermal station, with the rise of temperature, it is special that birefringence is presented in cornstarch
The particle of sign is reduced;When heating-up temperature is more than 63 DEG C, the number that birefringence feature is presented in starch granules is reduced rapidly;When thermal station plus
When hot temperature is close to 70 DEG C, corn starch granules lose birefringent feature completely.
In embodiment 1 Gelatinization Degree of Maize Starch and heating-up temperature graph of relation as shown in Figure 2 b, from Fig. 2 b, in heat
After platform heating schedule is warming up to 60 DEG C, with the rise of heating-up temperature, the gelatinization degree of cornstarch raises.Thermal station temperature programming is arrived
At 70 DEG C, the gelatinization degree of cornstarch reaches 100%;
Farina birefringence number in embodiment 2 in polarized light microscopy photo is with heating-up temperature graph of relation as schemed
Shown in 3a, from Fig. 3 a, when heating-up temperature is higher than 55 DEG C in thermal station, with the rise of temperature, farina presents two-fold
The particle for penetrating feature is reduced;When heating-up temperature is more than 56 DEG C, the number that birefringence feature is presented in starch granules is reduced rapidly;Work as heat
When platform heating-up temperature is close to 67 DEG C, potato starch particle loses birefringent feature completely.
In embodiment 2 farina gelatinization degree and heating-up temperature graph of relation as shown in Figure 3 b, from Fig. 3 b,
After thermal station heating schedule is warming up to 55 DEG C, with the rise of heating-up temperature, the gelatinization degree of cornstarch raises.Thermal station temperature programming
During to 67 DEG C, the gelatinization degree of farina reaches 100%;
Gelatinization point
The cornstarch of embodiment 1 and initial gelatinization temperature, final gelatinization point and the survey of the farina of embodiment 2
It is as shown in table 1 to try the deadline.
Starch initial gelatinization temperature, final gelatinization point and the test deadline of the embodiment 1~2 of table 1
As shown in Table 1, corn is obtained using micro- sem observation combination artificial intelligence energy image recognition technology in embodiment 1 to form sediment
Powder initial gelatinization temperature T0=63 DEG C, terminate gelatinization point Te=69 DEG C, the averagely time-consuming 3.35s of test, measurement error 0.47%;
In embodiment 2 initial gelatinization temperature of farina is obtained using micro- sem observation combination artificial intelligence energy image recognition technology
T0=57 DEG C, terminate gelatinization point Te=66 DEG C, the averagely time-consuming 3.67s of test, measurement error 0.74%.
The above described is only a preferred embodiment of the present invention, being not the limitation for making other forms to the present invention, appoint
What those skilled in the art is changed or is modified as the equivalent reality of equivalent variations possibly also with above-mentioned technology contents
Apply example.But it is every without departing from technical solution of the present invention content, above example is made according to technical spirit of the invention
Any simple modification, equivalent variations and remodeling, still fall within the protection domain of technical solution of the present invention.
Claims (9)
1. the online test method of a kind of starch burn degree and gelatinization point, it is characterised in that comprise the following steps:
(1) film-making:Starch milk is prepared, starch milk is drawn using rubber head dropper after being uniformly dispersed, is added dropwise on circular slide, is covered
Upper cover glass makes starch milk be uniformly dispersed and be sealed with glass cement, obtains sample chips;
(2) shoot:The sample chips made are placed in thermal station, heat up heating under petrographic microscope, and use shows with polarisation
The connected digital camera of micro mirror carries out continuous shooting, collecting, gathers polarized light microscopy photo;
(3) polarisation feature marks:Using Matlab softwares, will there is the starch of birefringence feature in the polarized light microscopy photo of collection
Particle is labeled with rectangular frame;
(4) training sample and birefringence identification:It is double in the polarized light microscopy photo after marking to completing using Starch-SSD methods
Refracting feature carries out deep learning, starch birefringence property data base is built, then to the starch gelatinization process in test sample
Polarized light microscopy photo carry out analysis detection;
(5) calculating of starch burn degree and sign:
<mrow>
<mi>D</mi>
<mi>G</mi>
<mi>%</mi>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mfrac>
<msub>
<mi>A</mi>
<mi>q</mi>
</msub>
<msub>
<mi>A</mi>
<mn>0</mn>
</msub>
</mfrac>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<mn>100</mn>
<mi>%</mi>
</mrow>
DG:The gelatinization degree of starch;
A0:Starch is original, and the particle number of birefringent phenomenon is presented when not heating up in polarized light microscopy photo;
Aq:During the temperature programming of thermal station to q DEG C of specified temp, the particle of birefringent phenomenon is presented in starch polarized light microscopy photo
Number;
(6) determination of starch gelatinization temperature:
The initial gelatinization temperature T of starch0, terminate gelatinization point TeDetermined by following relational expression:T0=TDG=5%, wherein TDG=5%
Represent temperature during DG%=5%;Te=TDG=95%, wherein TDG=95%Represent temperature during DG%=95%.
2. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (1), the starch includes the starch in the ative starch, modified starch and process of all kinds.
3. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (1), the mass fraction of the starch milk is 0.5%-3%.
4. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (2), the initial temperature of the heating heating is no more than 40 DEG C, and maximum heating temperature is no more than 100 DEG C.
5. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (2), the heating rate of the heating heating is less than 5 DEG C/min.
6. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (2), the sampling dot frequency of the shooting, collecting is less than 2 DEG C/time, i.e., samples one in the range of temperature less than 2 DEG C
It is individual and shot.
7. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (4), the polarized light microscopy photo of the training sample is more than 1000.
8. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (4), the algorithm of the deep learning owns using what image recognition technology was detected to starch birefringence feature
Algorithm routine.
9. the online test method of a kind of starch burn degree according to claim 1 and gelatinization point, it is characterised in that step
Suddenly in (4), the Starch-SSD methods are based on SSD methods, using single deep-neural-network, with petrographic microscope photo
Middle starch birefringence is characterized as detection mesh calibration method, and this method core is to be come using small convolution filter on predicted characteristics figure admittedly
The classification fraction of one group of fixed default boundary frame and position skew;
Starch-SSD target loss function:
<mrow>
<mi>L</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mover>
<mi>c</mi>
<mo>^</mo>
</mover>
<mo>,</mo>
<mi>l</mi>
<mo>,</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>L</mi>
<mrow>
<mi>c</mi>
<mi>o</mi>
<mi>n</mi>
<mi>f</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mover>
<mi>c</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>&alpha;L</mi>
<mrow>
<mi>l</mi>
<mi>o</mi>
<mi>c</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>l</mi>
<mo>,</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>L</mi>
<mrow>
<mi>c</mi>
<mi>o</mi>
<mi>n</mi>
<mi>f</mi>
</mrow>
</msub>
<mo>=</mo>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<mrow>
<mo>(</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>&Element;</mo>
<mi>p</mi>
<mi>o</mi>
<mi>s</mi>
</mrow>
<mi>N</mi>
</msubsup>
<msubsup>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mi>p</mi>
</msubsup>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<msubsup>
<mover>
<mi>c</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
<mi>p</mi>
</msubsup>
<mo>)</mo>
<mo>+</mo>
<msub>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>&Element;</mo>
<mi>n</mi>
<mi>e</mi>
<mi>g</mi>
</mrow>
</msub>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<msubsup>
<mover>
<mi>c</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
<mn>0</mn>
</msubsup>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
Overall target loss function loses conf and position loss loc weighted sum for confidence;
Forecast confidence after standardization,
l:Prediction block;
g:Give tacit consent to frame;
p:Classification number, p=2;
N:The number of the acquiescence frame to match with the starch birefringence feature frame of mark;
Represent the starch birefringence feature frame phase of i-th of acquiescence frame and classification p j-th of mark
Matching,Represent to mismatch;
I-th of acquiescence frame does not include any classification, then i ∈ neg, i.e. negative sample negtive;Otherwise, i ∈ pos, i.e. positive sample
positive;
Starch-SSD positioning, to center c=(cx, cy), width w and the height h of the starch birefringence feature frame of mark
Returned, weight α 1:
<mrow>
<msub>
<mi>L</mi>
<mrow>
<mi>l</mi>
<mi>o</mi>
<mi>c</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>l</mi>
<mo>,</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>&Element;</mo>
<mi>p</mi>
<mi>o</mi>
<mi>s</mi>
</mrow>
<mi>N</mi>
</msubsup>
<msub>
<mi>&Sigma;</mi>
<mrow>
<mi>m</mi>
<mo>&Element;</mo>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mi>x</mi>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>w</mi>
<mo>,</mo>
<mi>h</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<msubsup>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mi>p</mi>
</msubsup>
<msub>
<mi>smooth</mi>
<msub>
<mi>L</mi>
<mn>1</mn>
</msub>
</msub>
<mrow>
<mo>(</mo>
<msubsup>
<mi>l</mi>
<mi>i</mi>
<mi>m</mi>
</msubsup>
<mo>-</mo>
<msubsup>
<mover>
<mi>g</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>j</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
<mi>m</mi>
</msubsup>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mover>
<mi>g</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>j</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
<mrow>
<mi>c</mi>
<mi>x</mi>
</mrow>
</msubsup>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msubsup>
<mi>g</mi>
<mi>j</mi>
<mrow>
<mi>c</mi>
<mi>x</mi>
</mrow>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mrow>
<mi>c</mi>
<mi>x</mi>
</mrow>
</msubsup>
<mo>)</mo>
</mrow>
<mo>/</mo>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mi>w</mi>
</msubsup>
</mrow>
<mrow>
<msubsup>
<mover>
<mi>g</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>j</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
<mrow>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msubsup>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msubsup>
<mi>g</mi>
<mi>j</mi>
<mrow>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mrow>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msubsup>
<mo>)</mo>
</mrow>
<mo>/</mo>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mi>h</mi>
</msubsup>
</mrow>
<mrow>
<msubsup>
<mover>
<mi>g</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>j</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
<mi>w</mi>
</msubsup>
<mo>=</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mfrac>
<msubsup>
<mi>g</mi>
<mi>j</mi>
<mi>w</mi>
</msubsup>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mi>w</mi>
</msubsup>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mover>
<mi>g</mi>
<mo>^</mo>
</mover>
<mrow>
<mi>j</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
<mi>h</mi>
</msubsup>
<mo>=</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mfrac>
<msubsup>
<mi>g</mi>
<mi>j</mi>
<mi>h</mi>
</msubsup>
<msubsup>
<mi>d</mi>
<mi>i</mi>
<mi>h</mi>
</msubsup>
</mfrac>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710672065.4A CN107515198A (en) | 2017-08-08 | 2017-08-08 | A kind of online test method of starch burn degree and gelatinization point |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710672065.4A CN107515198A (en) | 2017-08-08 | 2017-08-08 | A kind of online test method of starch burn degree and gelatinization point |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107515198A true CN107515198A (en) | 2017-12-26 |
Family
ID=60723007
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710672065.4A Pending CN107515198A (en) | 2017-08-08 | 2017-08-08 | A kind of online test method of starch burn degree and gelatinization point |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107515198A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112102238A (en) * | 2020-08-10 | 2020-12-18 | 华南理工大学 | Method for detecting swelling capacity of starch granules in gelatinization process based on computer vision |
CN112881458A (en) * | 2021-02-03 | 2021-06-01 | 新疆农业科学院综合试验场 | Method for rapidly determining gelatinization degree of potato flour by using differential scanning calorimeter |
CN113214156A (en) * | 2021-05-13 | 2021-08-06 | 井冈山大学 | Rotor type fluorescent molecule for starch gelatinization degree detection and preparation and application thereof |
-
2017
- 2017-08-08 CN CN201710672065.4A patent/CN107515198A/en active Pending
Non-Patent Citations (1)
Title |
---|
JINXUAN TAO等: "A new methodology combining microscopy observation with Artificial Neural Networks for the study of starch gelatinization", 《FOOD HYDROCOLLOIDS》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112102238A (en) * | 2020-08-10 | 2020-12-18 | 华南理工大学 | Method for detecting swelling capacity of starch granules in gelatinization process based on computer vision |
CN112102238B (en) * | 2020-08-10 | 2023-06-20 | 华南理工大学 | Method for detecting swelling capacity of starch particles in gelatinization process based on computer vision |
CN112881458A (en) * | 2021-02-03 | 2021-06-01 | 新疆农业科学院综合试验场 | Method for rapidly determining gelatinization degree of potato flour by using differential scanning calorimeter |
CN113214156A (en) * | 2021-05-13 | 2021-08-06 | 井冈山大学 | Rotor type fluorescent molecule for starch gelatinization degree detection and preparation and application thereof |
CN113214156B (en) * | 2021-05-13 | 2022-05-31 | 井冈山大学 | Rotor type fluorescent molecule for starch gelatinization degree detection and preparation and application thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107657603A (en) | A kind of industrial appearance detecting method based on intelligent vision | |
CN109977780A (en) | A kind of detection and recognition methods of the diatom based on deep learning algorithm | |
CN107515198A (en) | A kind of online test method of starch burn degree and gelatinization point | |
WO2017067023A1 (en) | Method for detecting body fluid based on special test paper | |
CN109948469A (en) | The automatic detection recognition method of crusing robot instrument based on deep learning | |
CN107229930A (en) | A kind of pointer instrument numerical value intelligent identification Method and device | |
WO2021063062A1 (en) | Live-line testing system for power grid apparatus, thermal infrared imager, and method | |
CN109886925A (en) | A kind of aluminium material surface defect inspection method that Active Learning is combined with deep learning | |
CN112102238B (en) | Method for detecting swelling capacity of starch particles in gelatinization process based on computer vision | |
CN108764134A (en) | A kind of automatic positioning of polymorphic type instrument and recognition methods suitable for crusing robot | |
CN106526177A (en) | Biomarker detection system and method based on colloidal gold test strip | |
CN109827957A (en) | A kind of rice leaf SPAD value estimating and measuring method based on computer vision and system | |
CN110472581A (en) | A kind of cell image analysis method based on deep learning | |
CN104252056A (en) | Detection method and device of substrate | |
CN102419305B (en) | Method for determining gelatinization temperature and gelatinization degree of starch | |
CN102509096B (en) | Extracting and processing method for inclination angles of corn plant leaves | |
CN106053727A (en) | Standard curve correction method and system | |
Zhihong et al. | Smartphone-based visual measurement and portable instrumentation for crop seed phenotyping | |
CN106682695A (en) | County cultivated land natural quality elevating method based on support vector machine | |
CN106022354A (en) | SVM-based image MTF measurement method | |
CN108362652A (en) | A kind of object freshness lossless detection method based on evidence theory | |
CN107526294A (en) | A kind of Nonlinear Delay dynamic system model INTELLIGENT IDENTIFICATION method | |
Yin et al. | A Neural Network Method for Inversion of Turbulence Strength | |
Jiang et al. | Two new methods for severity assessment of wheat stripe rust caused by Puccinia striiformis f. sp. tritici | |
Li et al. | Hyperspectral characteristics and scale effects of leaf and canopy of summer maize under continuous water stresses |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171226 |
|
WD01 | Invention patent application deemed withdrawn after publication |