CN106067173A - The Complexity Measurement lossless detection method of citrusfruit pol - Google Patents
The Complexity Measurement lossless detection method of citrusfruit pol Download PDFInfo
- Publication number
- CN106067173A CN106067173A CN201610366953.9A CN201610366953A CN106067173A CN 106067173 A CN106067173 A CN 106067173A CN 201610366953 A CN201610366953 A CN 201610366953A CN 106067173 A CN106067173 A CN 106067173A
- Authority
- CN
- China
- Prior art keywords
- pol
- citrusfruit
- image
- training sample
- complexity measurement
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20152—Watershed segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses the Complexity Measurement lossless detection method of a kind of citrusfruit pol, including step: gather M training sample and N number of test samples；Gather training sample and the original image of test samples；Original image is cut；Image after cutting is gone background process；Image after going background process is carried out rim detection and fruit extracted region, RGB image is converted to HSI image；Try to achieve Complexity Measurement C (Y) and comentropy H (Y) of HSI image；The average pol of training sample is measured with saccharometer；Set up citrusfruit pol Nondestructive Testing Model；By H (Y) and C (Y) input nondestructive detection model, the pol that output detections sample is corresponding of detection sample.The present invention can carry out Non-Destructive Testing to citrusfruit pol, practical, provides foundation for producing with Citrus inside quality classification in sales process.
Description
Technical field
The present invention relates to the Complexity Measurement lossless detection method of a kind of citrusfruit pol.
Background technology
In the prior art, need to realize fruit breakage to be measured to the pol mensuration of citrusfruit, this method
Practicality is the strongest.
Due to orange peel be inside quality with presentation quality comprehensively map body, the color of peel and citrusfruit pol
Between there is relatedness.If can be by a kind of lossless method detection citrusfruit pol, inside for citrusfruit in producing
Quality grading provides foundation.Based on this point, the Complexity Measurement lossless detection method of a kind of citrusfruit pol is proposed.
Summary of the invention
It is an object of the invention to, for above-mentioned the deficiencies in the prior art, it is provided that the complexity of a kind of citrusfruit pol
Estimate lossless detection method.
For solving above-mentioned technical problem, the technical solution adopted in the present invention is:
The Complexity Measurement lossless detection method of a kind of citrusfruit pol, comprises the following steps:
Step one, M citrusfruit to be detected of collection, as training sample, gathers N number of citrus fruit implementation to be detected
For test samples, the carpopodium part that training sample and test samples exceed fruit face is removed；
Step 2, after being carried out training sample and test samples and dry up process, gathers training sample and inspection sample
This original image；
Step 3, cuts original image；
Step 4, goes background process to the image after cutting；
Step 5, carries out rim detection and fruit extracted region to the image after going background process, is converted to by RGB image
HSI image；
Step 6, carries out the even partition of a length of 1 °, forms 120 the tone of described HSI image interval [0,120 °]
Individual subinterval yi=[ai, ai+1], wherein i=0,1 ..., 119；a0=0, a120=120；Add up pixel n in each subintervali, meter
Calculate pixel distribution probabilityTrying to achieve Complexity Measurement C (Y)=H (Y) × D (Y), in formula, Y is stochastic variable,
Balanced distributionN=a120-a0, comentropy
Step 7, takes out the sarcocarp of each training sample and stirs respectively, measures each training with saccharometer
The average pol of sample；
Step 8, comentropy H (Y) with training sample is corresponding as input quantity, with training sample with Complexity Measurement C (Y)
Average pol train neutral net as output, set up citrusfruit pol Nondestructive Testing Model；
Step 9, inputs described Nondestructive Testing Model by comentropy H (Y) and the Complexity Measurement C (Y) of detection sample, defeated
Go out to detect the pol that sample is corresponding.
As a kind of optimal way, described step 2 includes: training sample and test samples are respectively placed in 500 ×
500×500mm3Lighting box bottom center, background black, digital camera is in lighting box center of top, and camera lens is away from Guo Ding
460～490mm, case top is symmetrical uniform 4 60w electric filament lamp, the digital picture of collecting fruit centered by camera lens.
As a kind of optimal way, described step 3 includes: utilize the digital imaging processing software original graph to gathering
As carrying out cutting of 1024 × 1024 pixel sizes.
As a kind of optimal way, described step 4 includes: statistics cuts the brightness Y=of rear citrusfruit image
0.1770R+0.8124G+0.0106B rectangular histogram, wherein, R, G, B respectively cut the redness of rear citrusfruit image, green and
Blue component；The trough brightness cut off value of bimodal of brightness histogram of extraction, as threshold value T, is set up luminance segmentation function, is less than
The gray scale of cut off value puts 1, constant higher than the gray scale of cut off value.
As a kind of optimal way, in described step 5, the method for rim detection and fruit extracted region includes: to removing the back of the body
Image after scape carries out the Prewitt operator filtering of horizontal and vertical directions, obtains filtering image ghAnd gv；To described filter
Ripple image carries out Euclidean Distance Transform and obtains receiving basin distance d to watershedf；To dfCarry out watershed detection, labelling df's
External constraint em, with local luminance gradient maximum size as condition, dynamically adjusts threshold value, filters out and goes background higher than threshold value
Gray level image gray scale maximum is extended maximum conversion, calculates dfInternal constraint im；Em and im is utilized to reconstruct gradient map
g2；To g2Do watershed detection, merge perimeter and interior zone, complete fruit margin detection, connect border, labelling fruit
Boundary profile, extracts fruit region.
There is more serious over-segmentation problem in traditional watershed algorithm, in the present invention, modified model watershed algorithm is passing
Carry out Grads threshold process on the basis of system watershed algorithm and internal constraint dynamically adjusts, overcome this shortcoming.
As a kind of optimal way, the method setting up citrusfruit pol Nondestructive Testing Model in described step 8 includes:
With comentropy H (Y) of training sample with Complexity Measurement C (Y) as input quantity, using average pol corresponding to training sample as defeated
Output sets up 3 layers of feedforward neural network citrusfruit pol Nondestructive Testing Model, and in described neutral net, hidden layer node is 5
Individual, hidden layer node uses tansig to transmit function, and output layer node uses purelin to transmit function, and network training is just using
Then change algorithm, by output error 10-4As terminating training criterion.
As a kind of optimal way, described step 9 detects the pol that sample is correspondingFormula
In,f1=tansig transmits function, f2=purelin transmits function, wijFor input layer to hidden
Connection weights containing layer, vjFor the connection weights of hidden layer to output layer,、b2It is respectively hidden layer and the threshold value of output layer,
x1=H (Y), x2=C (Y).
Compared with prior art, the present invention can carry out Non-Destructive Testing to citrusfruit pol, practical, for produce and
In sales process, the classification of Citrus inside quality provides foundation.
Accompanying drawing explanation
Fig. 1 is the image after cutting.
Fig. 2 is brightness histogram.
Fig. 3 is the image after background process.
Fig. 4 is fruit margin and fruit administrative division map.
Detailed description of the invention
The present invention is the Complexity Measurement lossless detection method of a kind of citrusfruit pol, using river, palace satsuma orange as quilt
Survey object, comprise the following steps:
Step one, gathers 100 citrusfruits to be detected as training sample, gathers 100 citrus fruits to be detected
It is implemented as test samples, by concordant for carpopodium fruit face, the carpopodium part that training sample and test samples exceed fruit face is deducted；
Step 2, after being carried out training sample and test samples and dry up process, gathers training sample and inspection sample
This original image；
Training sample and test samples are respectively placed in 500 × 500 × 500mm3Lighting box bottom center, background is black
Color, digital camera is in lighting box center of top, and camera lens is away from fruit top 460～490mm, and case top is symmetrical centered by camera lens
Uniform 4 60w electric filament lamp, the digital picture of collecting fruit.
Step 3, cuts original image；
Utilize digital imaging processing software that the original image gathered is carried out cutting of 1024 × 1024 pixel sizes, obtain
Image as shown in Figure 1.
Step 4, goes background process to the image after cutting；
Statistics cuts the brightness Y=0.1770R+0.8124G+0.0106B rectangular histogram of rear citrusfruit image, wherein, R,
G, B respectively cut redness, green and the blue component of rear citrusfruit image；The trough extracting bimodal of brightness histogram is bright
Degree cut off value, as threshold value T, is set up luminance segmentation function, is put 1 less than the gray scale of cut off value, constant higher than the gray scale of cut off value.
Brightness histogram is as shown in Figure 2.After treatment, as it is shown on figure 3, eliminate major part citrusfruit region outside background.
Step 5, carries out rim detection and fruit extracted region to the image after going background process, is converted to by RGB image
HSI image；
The method of rim detection and fruit extracted region includes: the image after going background carries out two sides of horizontal and vertical
To Prewitt operator filtering, obtain filtering image ghAnd gv；Computed range functionTo described filtering image
Carry out Euclidean Distance Transform and obtain receiving basin distance d to watershedf；To dfCarry out watershed detection, labelling dfOutside about
Bundle em, with local luminance gradient maximum size as condition, dynamically adjusts threshold value, filters out and removes background gray-scale map higher than threshold value
As gray scale maximum is extended maximum conversion, calculate dfInternal constraint im；Em and im is utilized to reconstruct gradient map g2；To g2
Doing watershed detection, merge perimeter and interior zone, complete fruit margin detection, connect border, labelling fruit border is taken turns
Exterior feature, extracts fruit region.As shown in Figure 4, after treatment, fruit margin is coherent uninterrupted, and fruit region is complete without hole.
Step 6, carries out the even partition of a length of 1 °, forms 120 the tone of described HSI image interval [0,120 °]
Individual subinterval yi=[ai, ai+1], wherein i=0,1 ..., 119；a0=0, a120=120；Add up pixel n in each subintervali, meter
Calculate pixel distribution probabilityTrying to achieve Complexity Measurement C (Y)=H (Y) × D (Y), in formula, Y is stochastic variable, flat
Weighing apparatus distributionN=a120-a0, comentropy
Step 7, takes out the sarcocarp of each training sample and stirs respectively, measures with the hand-held saccharometer of WYT-4 type
The average pol of each training sample；
Step 8, comentropy H (Y) with training sample is corresponding as input quantity, with training sample with Complexity Measurement C (Y)
Average pol train neutral net as output, set up citrusfruit pol Nondestructive Testing Model；
The method setting up citrusfruit pol Nondestructive Testing Model in described step 8 includes: with 100 training samples
Comentropy H (Y) is input quantity with Complexity Measurement C (Y), setting up using the average pol that training sample is corresponding as output can be real
3 layers of feedforward neural network citrusfruit pol Nondestructive Testing Model of existing complex mappings, in described neutral net, for making network tie
Structure is unlikely to too much redundancy, and taking hidden layer node is 5, and input layer is 2, and output layer node is 1, hidden layer node
Using conventional tansig to transmit function, output layer node uses conventional purelin to transmit function, and network training uses weights
The regularization algorithm few with number of threshold values redundancy.Pol accuracy of detection is affected, during the least then network training of error owing to error is the biggest
Between oversize, get after arithmetic point 4 appropriate, will output error 10-4As terminating training criterion, when network reaches output by mistake
Differ from 10-4Shi Xunlian terminates, and the pol of comentropy, Complexity Measurement and correspondence couples power by node and memorizes.
Step 9, inputs described Nondestructive Testing Model by comentropy H (Y) and the Complexity Measurement C (Y) of detection sample, defeated
Go out to detect the pol that sample is corresponding.
Described step 9 detects the pol that sample is correspondingIn formula,
f1=tansig transmits function, f2=purelin transmits function, wijFor the connection weights of input layer to hidden layer, vjFor hidden layer
To the connection weights of output layer,、b2It is respectively hidden layer and the threshold value of output layer, x1=H (Y), x2=C (Y).
For verifying the effectiveness of the inventive method, first it is utilized respectively the method for the invention 100 test samples of detection
Pol, then the sarcocarp of 100 test samples taken out and stir respectively, measuring with the hand-held saccharometer of WYT-4 type every
The average pol of one test samples, finally compares the pol that the pol recorded with saccharometer records with the present invention, finds
In the range of ± 1 ° of Brix degree, pol judgment accuracy is 83%, and in the range of ± 0.5 ° of Brix degree, pol judgment accuracy is
67%, therefore the method for the invention can be effective as the non invasive estimation of citrusfruit pol.
Claims (7)
1. the Complexity Measurement lossless detection method of a citrusfruit pol, it is characterised in that comprise the following steps:
Step one, M citrusfruit to be detected of collection, as training sample, gathers N number of citrusfruit to be detected as inspection
Test sample, the carpopodium part that training sample and test samples exceed fruit face is removed；
Step 2, after being carried out training sample and test samples and dry up process, gathers training sample and test samples
Original image；
Step 3, cuts original image；
Step 4, goes background process to the image after cutting；
Step 5, carries out rim detection and fruit extracted region to the image after going background process, RGB image is converted to HSI
Image；
Step 6, carries out the even partition of a length of 1 °, forms 120 sons the tone of described HSI image interval [0,120 °]
Interval yi=[ai, ai+1], wherein i=0,1 ..., 119；a0=0, a120=120；Add up pixel n in each subintervali, calculate picture
Element distribution probabilityTrying to achieve Complexity Measurement C (Y)=H (Y) × D (Y), in formula, Y is stochastic variable, balance
DistributionN=a120-a0, comentropy
Step 7, takes out the sarcocarp of each training sample and stirs respectively, measuring each training sample with saccharometer
Average pol；
Step 8, with comentropy H (Y) of training sample with Complexity Measurement C (Y) as input quantity, with corresponding flat of training sample
All pols train neutral net as output, set up citrusfruit pol Nondestructive Testing Model；
Step 9, inputs described Nondestructive Testing Model, output inspection by comentropy H (Y) and the Complexity Measurement C (Y) of detection sample
The pol of this correspondence of test sample.
2. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 1, it is characterised in that described step
Rapid two include: training sample and test samples are respectively placed in 500 × 500 × 500mm3Lighting box bottom center, background
Black, digital camera is in lighting box center of top, and camera lens is away from fruit top 460～490mm, and case top is right centered by camera lens
Claim uniform 4 60w electric filament lamp, the digital picture of collecting fruit.
3. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 1, it is characterised in that described step
Rapid three include: utilize digital imaging processing software that the original image gathered is carried out cutting of 1024 × 1024 pixel sizes.
4. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 1, it is characterised in that described step
Rapid four include: statistics cuts the brightness Y=0.1770R+0.8124G+0.0106B rectangular histogram of rear citrusfruit image, wherein,
R, G, B respectively cut redness, green and the blue component of rear citrusfruit image；Extract the trough of bimodal of brightness histogram
Brightness cut off value, as threshold value T, sets up luminance segmentation function, puts 1 less than the gray scale of cut off value, higher than cut off value gray scale not
Become.
5. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 1, it is characterised in that described step
In rapid five, the method for rim detection and fruit extracted region includes: the image after going background is carried out horizontal and vertical directions
Prewitt operator filtering, obtain filtering image ghAnd gv；Described filtering image is carried out Euclidean Distance Transform and obtains reception basin
Ground is to distance d in watershedf；To dfCarry out watershed detection, labelling dfExternal constraint em, with local luminance gradient maximum
Size is condition, dynamically adjusts threshold value, filters out and is extended greatly higher than the background gray level image gray scale maximum of going of threshold value
Value conversion, calculates dfInternal constraint im；Em and im is utilized to reconstruct gradient map g2；To g2Do watershed detection, merge perimeter
And interior zone, complete fruit margin detection, connect border, labelling fruit boundary profile, extract fruit region.
6. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 1, it is characterised in that described step
The method setting up citrusfruit pol Nondestructive Testing Model in rapid eight includes: comentropy H (Y) and complexity with training sample are surveyed
Degree C (Y) is input quantity, sets up 3 layers of feedforward neural network citrusfruit using the average pol that training sample is corresponding as output
Pol Nondestructive Testing Model, in described neutral net, hidden layer node is 5, and hidden layer node uses tansig to transmit function,
Output layer node uses purelin to transmit function, and network training uses regularization algorithm, by output error 10-4As terminating instruction
Practice criterion.
7. the Complexity Measurement lossless detection method of citrusfruit pol as claimed in claim 6, it is characterised in that described step
The pol that sample is corresponding is detected in rapid nineIn formula,f1=
Tansig transmits function, f2=purelin transmits function, wijFor the connection weights of input layer to hidden layer, vjFor hidden layer extremely
The connection weights of output layer,b2It is respectively hidden layer and the threshold value of output layer, x1=H (Y), x2=C (Y).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610366953.9A CN106067173A (en) | 2016-05-30 | 2016-05-30 | The Complexity Measurement lossless detection method of citrusfruit pol |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610366953.9A CN106067173A (en) | 2016-05-30 | 2016-05-30 | The Complexity Measurement lossless detection method of citrusfruit pol |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106067173A true CN106067173A (en) | 2016-11-02 |
Family
ID=57420988
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610366953.9A Pending CN106067173A (en) | 2016-05-30 | 2016-05-30 | The Complexity Measurement lossless detection method of citrusfruit pol |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106067173A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108319973A (en) * | 2018-01-18 | 2018-07-24 | 仲恺农业工程学院 | Citrusfruit detection method on a kind of tree |
| CN109001244A (en) * | 2018-09-07 | 2018-12-14 | 华中农业大学 | Device and its detection method based on infrared thermal imaging technique detection Potato Quality |
| CN110780044A (en) * | 2019-10-11 | 2020-02-11 | 华中农业大学 | Nondestructive testing method for sugar-acid ratio of fresh citrus fruits |
| CN112184627A (en) * | 2020-09-03 | 2021-01-05 | 华南农业大学 | Citrus fresh-keeping quality detection method based on image processing and neural network and application |
| CN113933305A (en) * | 2021-11-12 | 2022-01-14 | 江南大学 | Thin-skinned fruit sugar content nondestructive measurement method and system based on smart phone |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN103472031A (en) * | 2013-09-20 | 2013-12-25 | 华东交通大学 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
| CN104574341A (en) * | 2013-10-11 | 2015-04-29 | 中国林业科学研究院资源信息研究所 | Method and device for determining sugar degree of fruit |
-
2016
- 2016-05-30 CN CN201610366953.9A patent/CN106067173A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN103472031A (en) * | 2013-09-20 | 2013-12-25 | 华东交通大学 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
| CN104574341A (en) * | 2013-10-11 | 2015-04-29 | 中国林业科学研究院资源信息研究所 | Method and device for determining sugar degree of fruit |
Non-Patent Citations (4)
| Title |
|---|
| 曹乐平 等: "基于色调分形维数的柑橘糖度和有效酸度检测", 《农业机械学报》 * |
| 曹乐平: "基于周长面积分形维数的柑橘品种机器识别", 《农业工程学报》 * |
| 温芝元，曹乐平: "基于复杂性测度的柑橘病虫害识别", 《中国农学通报》 * |
| 温芝元，曹乐平: "基于补偿模糊神经网络的脐橙不同病虫害图像识别", 《农业工程学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108319973A (en) * | 2018-01-18 | 2018-07-24 | 仲恺农业工程学院 | Citrusfruit detection method on a kind of tree |
| CN108319973B (en) * | 2018-01-18 | 2022-01-28 | 仲恺农业工程学院 | Detection method for citrus fruits on tree |
| CN109001244A (en) * | 2018-09-07 | 2018-12-14 | 华中农业大学 | Device and its detection method based on infrared thermal imaging technique detection Potato Quality |
| CN110780044A (en) * | 2019-10-11 | 2020-02-11 | 华中农业大学 | Nondestructive testing method for sugar-acid ratio of fresh citrus fruits |
| CN110780044B (en) * | 2019-10-11 | 2022-04-19 | 华中农业大学 | Nondestructive testing method for sugar-acid ratio of fresh citrus fruits |
| CN112184627A (en) * | 2020-09-03 | 2021-01-05 | 华南农业大学 | Citrus fresh-keeping quality detection method based on image processing and neural network and application |
| CN113933305A (en) * | 2021-11-12 | 2022-01-14 | 江南大学 | Thin-skinned fruit sugar content nondestructive measurement method and system based on smart phone |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106067173A (en) | The Complexity Measurement lossless detection method of citrusfruit pol | |
| Nuske et al. | Yield estimation in vineyards by visual grape detection | |
| CN105354865B (en) | The automatic cloud detection method of optic of multispectral remote sensing satellite image and system | |
| CN101085442B (en) | Method for treating and classifying orange image based on RGB composite model | |
| CN106855951A (en) | A kind of grain kind quality detecting method based on computer vision | |
| CN104732229B (en) | A kind of dividing method for being used for overlapping cell in cervical smear image | |
| CN109118548A (en) | A kind of comprehensive intelligent water quality recognition methods | |
| CN106296670B (en) | A kind of Edge detection of infrared image based on the watershed Retinex--Canny operator | |
| CN103394472B (en) | A kind of greening potato based on machine vision detects stage division | |
| CN111652326A (en) | Improved fruit maturity identification method and identification system based on MobileNet v2 network | |
| WO2024021359A1 (en) | Built environment dominant color measurement method and system based on image eeg sensitivity data | |
| Grocholsky et al. | A camera and laser system for automatic vine balance assessment | |
| CN111476153A (en) | Method for calculating fruit maturity | |
| CN101936894B (en) | Near infrared spectrum and microscopic adipose cell data fusion-based pork freshness non-destructive testing technology | |
| CN106645180A (en) | Method for checking defects of substrate glass, field terminal and server | |
| CN112613338A (en) | Wheat leaf layer nitrogen content estimation method based on RGB image fusion characteristics | |
| CN105427279B (en) | A kind of grassland Drought Information Monitoring System and method based on computer vision and Internet of Things | |
| CN108596244A (en) | A kind of high spectrum image label noise detecting method based on spectrum angle density peaks | |
| CN106023223A (en) | Orange fruit size describing method and organic fruit size grading method | |
| CN111860125A (en) | Method for identifying fruit maturity based on machine vision | |
| CN106157335B (en) | Bed material surface layer grading observation and analysis method and system based on digital picture | |
| CN109272484B (en) | Rainfall detection method based on video image | |
| CN101943661B (en) | Near infrared spectrum and microscopic bacterial plaque area data fusion-based pork freshness non-destructive testing technology | |
| CN116824405A (en) | Green-peel walnut target detection method based on improved YOLOv8 under unmanned aerial vehicle aerial photographing visual angle | |
| CN113324928B (en) | Intelligent nondestructive detection method for nutrient content of dish food based on map data |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161102 |