CN109977802A - Crops Classification recognition methods under strong background noise - Google Patents
Crops Classification recognition methods under strong background noise Download PDFInfo
- Publication number
- CN109977802A CN109977802A CN201910175557.1A CN201910175557A CN109977802A CN 109977802 A CN109977802 A CN 109977802A CN 201910175557 A CN201910175557 A CN 201910175557A CN 109977802 A CN109977802 A CN 109977802A
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- background noise
- recognition methods
- strong background
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000003062 neural network model Methods 0.000 claims abstract description 24
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 6
- 238000013526 transfer learning Methods 0.000 claims abstract description 6
- 238000013480 data collection Methods 0.000 claims abstract description 5
- 239000007787 solid Substances 0.000 claims abstract description 4
- 238000010200 validation analysis Methods 0.000 claims abstract description 4
- 239000010410 layer Substances 0.000 claims description 26
- 238000002310 reflectometry Methods 0.000 claims description 6
- 230000005855 radiation Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 239000002356 single layer Substances 0.000 claims description 2
- 241000196324 Embryophyta Species 0.000 description 15
- 238000013135 deep learning Methods 0.000 description 12
- 230000000694 effects Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 238000000701 chemical imaging Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241000209140 Triticum Species 0.000 description 1
- 235000021307 Triticum Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- VMXUWOKSQNHOCA-UKTHLTGXSA-N ranitidine Chemical compound [O-][N+](=O)\C=C(/NC)NCCSCC1=CC=C(CN(C)C)O1 VMXUWOKSQNHOCA-UKTHLTGXSA-N 0.000 description 1
- 238000009394 selective breeding 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
-
- 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
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- 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
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8466—Investigation of vegetal material, e.g. leaves, plants, fruits
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/68—Food, e.g. fruit or vegetables
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Analytical Chemistry (AREA)
- Mathematical Physics (AREA)
- Business, Economics & Management (AREA)
- Immunology (AREA)
- Biomedical Technology (AREA)
- Biochemistry (AREA)
- Evolutionary Computation (AREA)
- Chemical & Material Sciences (AREA)
- Software Systems (AREA)
- Pathology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Animal Husbandry (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- General Business, Economics & Management (AREA)
- Agronomy & Crop Science (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Primary Health Care (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses Crops Classification recognition methods under a kind of strong background noise, each several of the picture for shooting crop of all categories with multispectral camera form pictures；It obtains each pixel NDVI value and is partitioned into plant regional；Non-plant region is replaced with into solid background to protrude plant regional, form multispectral data collection after carrying out picture pretreatment and is divided into 3 training, test, verifying data sets；Pass through the method for transfer learning, training dataset is inputted preset convolutional neural networks model to be trained, convolution prediction neural network model is obtained, test data set input convolution prediction neural network model is subjected to accuracy rate and tests to obtain qualified convolution prediction neural network model；Validation data set is inputted into convolution prediction neural network model, Classification and Identification is carried out to crop therein and obtains classification results.The method reduce the influences that strong background noise generates Crops Classification identification, improve the recognition efficiency and predictive ability of model.
Description
Technical field
The invention belongs to Crop classifications to identify field, and in particular to Crops Classification identification side under a kind of strong background noise
Method.
Background technique
In image classification and retrieval based on deep learning, how to which of image progress feature extraction and extraction image
Feature (color, texture, shape etc.) not only affects the accuracy of image classification, but also rises to content-based image retrieval
Vital effect.Currently, when carrying out crop identification by deep learning using RGB image, under by strong background noise
There is the trend that is decreased obviously in the influence of interference characteristic and self brightness, the precision of deep learning, RGB color model passes through between image
Color, shape, Texture eigenvalue difference achieve the effect that classification and identification Different Crop, but this method cannot give expression to
The information of Color-spatial distribution, only records the information of three wave bands of RGB, and all band is lost, is unfavorable in strong background noise
Under identification to crop.In most of deep learnings, crop image generally passes through artificial selection and evades complicated interference, but real
The influence that may be subjected to strong background noise in the application of border, causes Crops Classification accuracy of identification inadequate.
Summary of the invention
The object of the present invention is to provide Crops Classification recognition methods under a kind of strong background noise, and the method reduce strong noises
Background identifies the influence generated to Crops Classification, improves recognition accuracy, can be applied to agrotype under Small Sample Database
Identification, improves the recognition efficiency and predictive ability of model.
The technical scheme adopted by the invention is that:
Crops Classification recognition methods under a kind of strong background noise, comprising steps of
S1, the picture that crop of all categories is shot with multispectral camera it is each several, form pictures；
S2, radiation calibration and vegetation index calculating are completed with algorithm, obtain each pixel NDVI value and be partitioned into plant regional；
S3, non-plant region is replaced with into solid background to protrude plant regional, forms mostly light after carrying out picture pretreatment
Multispectral data collection is divided into 3 training, test, verifying data sets by spectrum data set；
Training dataset is inputted preset convolutional neural networks model and is trained by S4, the method by transfer learning,
Convolution prediction neural network model is obtained, test data set input convolution prediction neural network model is subjected to accuracy rate test,
Optimize the parameter of convolution prediction neural network model according to test result and obtains qualified convolution prediction neural network model；
S5, validation data set is inputted into convolution prediction neural network model, Classification and Identification is carried out to crop therein and obtained
Take classification results.
In S1, when multispectral camera is shot, it is necessary to based under the good sunshine condition of light, guarantee enough light
Line amount.
In S2, the calculation formula of the NDVI value of each pixel is,
NDVI=(ρnir-ρred)/(ρnir+ρred)
Wherein ρnirReflectivity, the ρ obtained near infrared bandredFor the reflectivity that infrared band obtains, plant regional
NDVI value of the NDVI value obviously than ground object area is high, is calculated automatically from NDVI threshold value by otsu algorithm, is drawn with NDVI threshold value
Divide plant regional and ground object area.
In S3, carry out using TFRecord format system storing data when picture pretreatment.
In S4, preset convolutional neural networks model uses the Inception_v3 of GoogleNet.
Further, for the parameter setting of Inception_v3, retain the ginseng of all convolutional layers in Inception_v3
Number only substitutes the last layer and trains layer entirely, is the extraction to the feature vector of image before the last layer, with the feature extracted
Vector obtains the full Connection Neural Network model of a single layer as output training.
The beneficial effects of the present invention are:
The multispectral image being made of multiple channels, each channel capture the light of specified wavelength, can fully consider figure
The Color-spatial distribution information of picture, multispectral imaging can obtain spectral signature and obtain image information, and this method utilizes more
Spectrum camera shoots crop picture, eliminates pure RGB image background color and obscures feature searching, and it is accurate to improve identification
Rate.
NDVI value (vegetation index) is applied to monitoring vegetation growth state, vegetation coverage and eliminates part
Radiation error helps from comprising water and soil etc. to separate plant in interior ambient enviroment, and this method is multispectral by calculating
Picture each point NDVI value is simultaneously partitioned into plant regional with this, separates crop under strong background noise, reduces strong background noise pair
The influence that Crops Classification identification generates, further extracts and enhances crop feature.
This method obtains convolution prediction neural network model, can be applied to Small Sample Database by the method for transfer learning
The identification of lower agrotype, NDVI value using the accurate position of crop feature, improve the recognition efficiency of model, parameter
Optimization improves the predictive ability of model, can reach better Crops Classification recognition effect based on deep learning, which can
To improve the accuracy rate that Crops Classification identifies under Small Sample Database.
This method is equivalent to delimitation using the thinking for carrying out deep learning after NDVI value cut zone to specified region again
One piece of region is played for deep learning, reduces the workload and possible mistake that characteristic point is found in deep learning.
Detailed description of the invention
Fig. 1 is the principle of the present invention flow chart.
Fig. 2 is specific flow chart of the invention.
Specific embodiment
The present invention is further illustrated with reference to the accompanying drawings and examples.
As depicted in figs. 1 and 2, Crops Classification recognition methods under a kind of strong background noise, comprising steps of
S1, wheat and each 500, picture of other classification plant are shot with multispectral camera, forms pictures.
S2, radiation calibration and vegetation index calculating are completed with algorithm, i.e., each pixel NDVI value is simultaneously partitioned into plant with this
Region, the calculation formula of NDVI value of each pixel in NDVI image are as follows:
NDVI=(ρnir-ρred)/(ρnir+ρred)
Wherein ρnirFor the reflectivity that near infrared band obtains, ρredThe reflectivity obtained for infrared band；Plant regional
NDVI value can be obviously higher than atural object, gives NDVI threshold value, divides plant and ground object area with this.
S3, the replacement of plant regional external application solid background, prominent plant regional, then carry out pretreatment (the progress picture of picture
Using TFRecord format system storing data when pretreatment, when data source is more complicated, sample type increases, each sample
In information it is more complicated after, still be able to effectively record input data in information), formed multispectral data collection, operation
Code realizes the division of multispectral data collection and forming label, and all pictures are divided into 3 training, test, verifying data sets, and
And picture is converted to 299 × 299 × 3 character matrix that Inception_v3 needs from unprocessed form.
Training dataset is inputted preset convolutional neural networks model and is trained by S4, the method by transfer learning,
Convolution prediction neural network model after being trained, detailed process are as follows:
Convolution presets the Inception_v3 that neural network model uses GoogleNet, can be by one biggish two
Dimension convolution kernel splits into two lesser one-dimensional convolution (for example 3 × 3 convolution are splitted into 1 × 3 convolution sum, 3 × 1 convolution), reduces ginseng
Number quantity accelerates operation and alleviates over-fitting, while increasing one layer of nonlinear extensions model tormulation ability；It is used in structure
The convolutional layer of different size convolution kernel does parallel connection, therefore can identify the feature on different scale.
When the Inception_v3 mould that training dataset has been prepared for, and load is defined by Tensorflow-Slim
Type defines preset hyper parameter, and finally the weight parameter of full articulamentum and offset parameter are obtained network by new data training, are saved
Train the path of model.
Simplified model form be divided into from front to back convolutional layer → pond layer → convolutional layer → pond layer → full articulamentum →
Articulamentum → softmax layers complete, size when picture has just inputted convolutional layer is 299 × 299, and convolutional layer will be every in neural network
One fritter carry out deeper into analysis, to obtain the higher feature of level of abstraction, width after convolution and highly usable following
Formula calculates:
W2=(W1-F+2P)/S+1
H2=(H1-F+2P)/S+1
Wherein, W2It is width, the W of Feature Map after convolution1Be image before convolution width, F be filter width,
P be the quantity of Zero Padding, Zero Padding refer to mended around original image a few circles 0 (if value be 1,
With regard to mend 1 circle 0), S be stride, H2It is the height of Feature Map, H after convolution1It is the height of image before convolution.This layer it is defeated
It is out next layer of input.
Pond layer neural network will not change the depth of three-dimensional matrice, but it can reduce the size of matrix, further contract
The number of small last full articulamentum interior joint, achievees the purpose that reduce parameter in entire neural network.
After processing layer by layer by convolutional layer and pond layer, it is higher that the information in image has been abstracted into information content
Feature.Subsequent several layers of full articulamentums are then used to carry out classification task.Last softmax layer is then indicated using a probability
Object to be sorted has much probability to belong to some class.
It can be found that network is restrained quickly after by newdata collection training, training loss is gradually decreased, and accuracy rate is promoted
To 92% or so.
Finally, test data set input convolution prediction neural network model is carried out accuracy rate test, according to test result
Optimize the parameter of convolution prediction neural network model and obtains qualified convolution prediction neural network model；
S5, validation data set is inputted into convolution prediction neural network model, Classification and Identification is carried out to crop therein and obtained
Take classification results.
The multispectral image being made of multiple channels, each channel capture the light of specified wavelength, can fully consider figure
The Color-spatial distribution information of picture, multispectral imaging can obtain spectral signature and obtain image information, and this method utilizes more
Spectrum camera shoots crop picture, eliminates pure RGB image background color and obscures feature searching, and it is accurate to improve identification
Rate.
NDVI value (vegetation index) is applied to monitoring vegetation growth state, vegetation coverage and eliminates part
Radiation error helps from comprising water and soil etc. to separate plant in interior ambient enviroment, and this method is multispectral by calculating
Picture each point NDVI value is simultaneously partitioned into plant regional with this, separates crop under strong background noise, reduces strong background noise pair
The influence that Crops Classification identification generates, further extracts and enhances crop feature.
This method obtains convolution prediction neural network model, can be applied to Small Sample Database by the method for transfer learning
The identification of lower agrotype, NDVI value using the accurate position of crop feature, improve the recognition efficiency of model, parameter
Optimization improves the predictive ability of model, can reach better Crops Classification recognition effect based on deep learning, which can
To improve the accuracy rate that Crops Classification identifies under Small Sample Database.
This method is equivalent to delimitation using the thinking for carrying out deep learning after NDVI value cut zone to specified region again
One piece of region is played for deep learning, reduces the workload and possible mistake that characteristic point is found in deep learning.
It should be understood that for those of ordinary skills, it can be modified or changed according to the above description,
And all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (6)
1. Crops Classification recognition methods under a kind of strong background noise, it is characterised in that: including step,
S1, the picture that crop of all categories is shot with multispectral camera it is each several, form pictures；
S2, radiation calibration and vegetation index calculating are completed with algorithm, obtain each pixel NDVI value and be partitioned into plant regional；
S3, non-plant region is replaced with into solid background to protrude plant regional, forms multispectral number after carrying out picture pretreatment
According to collection, multispectral data collection is divided into 3 training, test, verifying data sets；
Training dataset is inputted preset convolutional neural networks model and is trained, obtained by S4, the method by transfer learning
Test data set input convolution prediction neural network model is carried out accuracy rate test by convolution prediction neural network model, according to
The parameter of test result optimization convolution prediction neural network model simultaneously obtains qualified convolution prediction neural network model；
S5, validation data set is inputted into convolution prediction neural network model, crop therein is carried out Classification and Identification and obtained to divide
Class result.
2. Crops Classification recognition methods under strong background noise as described in claim 1, it is characterised in that: multispectral in S1
When camera is shot, it is necessary to based under the good sunshine condition of light, guarantee enough amount lights.
3. Crops Classification recognition methods under strong background noise as described in claim 1, it is characterised in that: in S2, each picture
The calculation formula of the NDVI value of vegetarian refreshments is,
NDVI=(ρnir-ρred)/(ρnir+ρred)
Wherein ρnirReflectivity, the ρ obtained near infrared bandredFor the reflectivity that infrared band obtains, the NDVI of plant regional
The obvious NDVI value than ground object area of value is high, is calculated automatically from NDVI threshold value by otsu algorithm, is divided and planted with NDVI threshold value
Object area and ground object area.
4. Crops Classification recognition methods under strong background noise as described in claim 1, it is characterised in that: in S3, carry out figure
Using TFRecord format system storing data when piece pre-processes.
5. Crops Classification recognition methods under strong background noise as described in claim 1, it is characterised in that: preset in S4
Convolutional neural networks model uses the Inception_v3 of GoogleNet.
6. Crops Classification recognition methods under strong background noise as claimed in claim 5, it is characterised in that: for Inception_
The parameter setting of v3 retains the parameter of all convolutional layers in Inception_v3, only substitutes the last layer and trains layer entirely, last
It is the extraction to the feature vector of image before layer, uses the feature vector extracted to obtain a single layer as output training and connect entirely
Connect neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910175557.1A CN109977802A (en) | 2019-03-08 | 2019-03-08 | Crops Classification recognition methods under strong background noise |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910175557.1A CN109977802A (en) | 2019-03-08 | 2019-03-08 | Crops Classification recognition methods under strong background noise |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109977802A true CN109977802A (en) | 2019-07-05 |
Family
ID=67078249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910175557.1A Pending CN109977802A (en) | 2019-03-08 | 2019-03-08 | Crops Classification recognition methods under strong background noise |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109977802A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619349A (en) * | 2019-08-12 | 2019-12-27 | 深圳市识农智能科技有限公司 | Plant image classification method and device |
| CN110909820A (en) * | 2019-12-02 | 2020-03-24 | 齐鲁工业大学 | Image classification method and system based on self-supervision learning |
| CN110991454A (en) * | 2019-12-23 | 2020-04-10 | 云南大学 | Blade image recognition method and device, electronic equipment and storage medium |
| CN111428798A (en) * | 2020-03-30 | 2020-07-17 | 北京工业大学 | Plant seedling classification method based on convolutional neural network |
| CN112347894A (en) * | 2020-11-02 | 2021-02-09 | 东华理工大学 | Single-plant vegetation extraction method based on transfer learning and Gaussian mixture model separation |
| CN113489869A (en) * | 2021-07-05 | 2021-10-08 | 深圳市威视佰科科技有限公司 | Clothing material identification method based on hyperspectral camera |
| CN113673490A (en) * | 2021-10-21 | 2021-11-19 | 武汉大学 | Phenological period self-adaptive crop physiological parameter remote sensing estimation method and system |
| WO2023018387A1 (en) * | 2021-08-11 | 2023-02-16 | Agcurate Bilgi Teknolojileri Anonim Sirketi | A crop classification method using deep neural networks |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991439A (en) * | 2017-03-28 | 2017-07-28 | 南京天数信息科技有限公司 | Image-recognizing method based on deep learning and transfer learning |
| CN108710864A (en) * | 2018-05-25 | 2018-10-26 | 北华航天工业学院 | Winter wheat Remotely sensed acquisition method based on various dimensions identification and image noise reduction processing |
| CN109241817A (en) * | 2018-07-02 | 2019-01-18 | 广东工业大学 | A kind of crops image-recognizing method of unmanned plane shooting |
-
2019
- 2019-03-08 CN CN201910175557.1A patent/CN109977802A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991439A (en) * | 2017-03-28 | 2017-07-28 | 南京天数信息科技有限公司 | Image-recognizing method based on deep learning and transfer learning |
| CN108710864A (en) * | 2018-05-25 | 2018-10-26 | 北华航天工业学院 | Winter wheat Remotely sensed acquisition method based on various dimensions identification and image noise reduction processing |
| CN109241817A (en) * | 2018-07-02 | 2019-01-18 | 广东工业大学 | A kind of crops image-recognizing method of unmanned plane shooting |
Non-Patent Citations (1)
| Title |
|---|
| 卫娇娇: "基于遥感的甘肃河西地区绿洲分布提取及时空变化分析", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619349A (en) * | 2019-08-12 | 2019-12-27 | 深圳市识农智能科技有限公司 | Plant image classification method and device |
| CN110909820A (en) * | 2019-12-02 | 2020-03-24 | 齐鲁工业大学 | Image classification method and system based on self-supervision learning |
| CN110991454A (en) * | 2019-12-23 | 2020-04-10 | 云南大学 | Blade image recognition method and device, electronic equipment and storage medium |
| CN111428798A (en) * | 2020-03-30 | 2020-07-17 | 北京工业大学 | Plant seedling classification method based on convolutional neural network |
| CN112347894A (en) * | 2020-11-02 | 2021-02-09 | 东华理工大学 | Single-plant vegetation extraction method based on transfer learning and Gaussian mixture model separation |
| CN112347894B (en) * | 2020-11-02 | 2022-05-20 | 东华理工大学 | Single plant vegetation extraction method based on transfer learning and Gaussian mixture model separation |
| CN113489869A (en) * | 2021-07-05 | 2021-10-08 | 深圳市威视佰科科技有限公司 | Clothing material identification method based on hyperspectral camera |
| WO2023018387A1 (en) * | 2021-08-11 | 2023-02-16 | Agcurate Bilgi Teknolojileri Anonim Sirketi | A crop classification method using deep neural networks |
| CN113673490A (en) * | 2021-10-21 | 2021-11-19 | 武汉大学 | Phenological period self-adaptive crop physiological parameter remote sensing estimation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109977802A (en) | Crops Classification recognition methods under strong background noise | |
| Wang et al. | Identification of tomato disease types and detection of infected areas based on deep convolutional neural networks and object detection techniques | |
| Kong et al. | Multi-stream hybrid architecture based on cross-level fusion strategy for fine-grained crop species recognition in precision agriculture | |
| Zhou et al. | Wheat ears counting in field conditions based on multi-feature optimization and TWSVM | |
| CN111709379B (en) | Remote sensing image-based hilly area citrus planting land plot monitoring method and system | |
| CN110619632B (en) | Mango example confrontation segmentation method based on Mask R-CNN | |
| Gong et al. | Citrus yield estimation based on images processed by an Android mobile phone | |
| CN110569747A (en) | method for rapidly counting rice ears of paddy field rice by using image pyramid and fast-RCNN | |
| Chen et al. | Citrus fruits maturity detection in natural environments based on convolutional neural networks and visual saliency map | |
| CN109829425B (en) | Farmland landscape small-scale ground feature classification method and system | |
| CN114445785A (en) | Internet of things-based litchi insect pest monitoring and early warning method and system and storage medium | |
| CN115311588A (en) | Pine wood nematode disease stumpage detection method and device based on unmanned aerial vehicle remote sensing image | |
| CN111814563B (en) | Method and device for classifying planting structures | |
| Lv et al. | A visual identification method for the apple growth forms in the orchard | |
| CN112560623B (en) | Unmanned aerial vehicle-based rapid mangrove plant species identification method | |
| CN116543316B (en) | Method for identifying turf in paddy field by utilizing multi-time-phase high-resolution satellite image | |
| Hao et al. | Growing period classification of Gynura bicolor DC using GL-CNN | |
| CN110163101A (en) | The difference of Chinese medicine seed and grade quick discrimination method | |
| CN113657158A (en) | Google Earth Engine-based large-scale soybean planting region extraction algorithm | |
| Yang et al. | A comparative evaluation of convolutional neural networks, training image sizes, and deep learning optimizers for weed detection in alfalfa | |
| CN116129260A (en) | Forage grass image recognition method based on deep learning | |
| Du et al. | DSW-YOLO: A detection method for ground-planted strawberry fruits under different occlusion levels | |
| CN113435254A (en) | Sentinel second image-based farmland deep learning extraction method | |
| Li et al. | Maize leaf disease identification based on WG-MARNet | |
| CN115330833A (en) | Fruit yield estimation method with improved multi-target tracking |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190705 |
|
| RJ01 | Rejection of invention patent application after publication |