CN109063713A - A kind of timber discrimination method and system based on the study of construction feature picture depth - Google Patents
A kind of timber discrimination method and system based on the study of construction feature picture depth Download PDFInfo
- Publication number
- CN109063713A CN109063713A CN201810800080.7A CN201810800080A CN109063713A CN 109063713 A CN109063713 A CN 109063713A CN 201810800080 A CN201810800080 A CN 201810800080A CN 109063713 A CN109063713 A CN 109063713A
- Authority
- CN
- China
- Prior art keywords
- timber
- image
- convolutional neural
- neural networks
- multilayer convolutional
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present embodiments relate to a kind of timber discrimination methods and system based on the study of construction feature picture depth, which comprises acquisition wood transverse section constructs image data；Described image data are divided into multiple image blocks of the same size；The corresponding training set of described image data and test set are established according to multiple described image blocks；Construct timber image authentication multilayer convolutional neural networks；Deep learning is carried out to the timber image authentication multilayer convolutional neural networks using the training set；It is tested using model of the test set to deep learning, according to test result Optimized model parameter, generates timber image authentication deep learning algorithm model；Identify that deep learning algorithm model identifies timber image data to be identified according to described image.Thus, it is possible to realize the accurate quick identification to timber varieties of trees to be identified.
Description
Technical field
The present embodiments relate to computer vision technique more particularly to it is a kind of based on construction feature picture depth study
Timber discrimination method and system.
Background technique
With the continuous aggravation of the continuous consumption and the demand and supply contraction of timber resources, illegally adopted by the timber that interests drive
The sustainable use for having seriously affected timber resources with trade is cut down, while grave danger is constituted to species conservation and ecological environment.
Traditional Wood Identification Method is established on the basis of Wood Anatomical Structure, can only identify level of the timber to " category ", Er Qiezhou
Phase is long, at high cost, is overly dependent upon the wood identification personnel of profession.Emerging DNA bar code technology, chemical fingerprint technology
Although needing to construct the perfect database of profession Deng timber may be implemented in the identification of " kind " level, expending huge people
Power and financial resources are unfavorable for being widely applied in actual production and life.
Timber is a kind of with anisotropic natural material, and three sections of timber (cross section, radial longitudinal section, tangential section) are in
Existing different anatomical features.It is the key that carry out timber identification that effective identification feature is extracted from Wood Anatomical Structure,
But traditional timber image-recognizing method, by manually extracting feature, it is difficult to extract go out that the effective of different tree species timber can be identified
Construction feature.Meanwhile timber varieties of trees has biggish intraspecific variablity, the anatomic construction of same tree species usually has biggish change
The opposite sex causes traditional wood structure image-recognizing method accuracy rate lower.Thus, by wood structure image to timber varieties of trees
It is identified, is an extremely complex and challenging job.
In recent years, deep learning is developed rapidly in computer vision field, be widely used in recognition of face,
The fields such as security security protection, unmanned, medical diagnosis on disease.Pass through acquisition Wood structure features image data and carry out image segmentation,
The great amount of images block for covering wood variations can be obtained；Extraction timber can be automated by constructing multilayer convolutional neural networks
Identification feature in construction solves the problems, such as that feature extraction difficulty and recognition accuracy are low in traditional images identification technology, are
Wood structure image recognition provides new idea and method.
Summary of the invention
The embodiment of the invention provides a kind of timber discrimination methods and system based on the study of construction feature picture depth, can
To realize the accurate quick identification to timber varieties of trees to be identified.
In a first aspect, the embodiment of the present invention provides a kind of timber discrimination method based on the study of construction feature picture depth,
Include:
Acquire the construction feature image data of wood transverse section；
Described image data are divided into multiple image blocks of the same size；
The corresponding training set of described image data and test set are established according to multiple described image blocks；
Construct timber image authentication multilayer convolutional neural networks；
Deep learning is carried out to the timber image authentication multilayer convolutional neural networks using the training set；
It is tested using model of the test set to deep learning, according to test result Optimized model parameter, is generated
The image recognition deep learning algorithm model of the timber to be identified；
Wood structure image data to be identified is identified using described image identification deep learning algorithm model, is exported
Recognition result and confidence level.
In a possible embodiment, the building timber image authentication multilayer convolutional neural networks, comprising: building
Multilayer convolutional neural networks VGG16；The multilayer convolutional neural networks VGG16 is instructed in advance using ImageNet data set
Practice；Timber image authentication multilayer convolutional neural networks are constructed according to the pre-training VGG16.
In a possible embodiment, the acquisition wood transverse section construction feature image data to be identified, comprising:
The RGB image of the cross section construction feature of 500 each tree species is acquired by image capture module；Wherein, point of described image
Resolution is 2048*2048, depth is 8.
In a possible embodiment, it is 512*512, the image that quantity is 2000 that described image block, which is resolution ratio,
Block.
In a possible embodiment, the training set image number of blocks accounting is 80%, the test set image
Number of blocks accounting is 20%.
In a possible embodiment, the multilayer convolutional neural networks VGG16 includes: 1 input layer, 13 volumes
Lamination, 5 maximum value pond layers and 1 output layer.
In a possible embodiment, the timber image authentication multilayer convolutional neural networks, comprising: described
7 convolutional layers and 3 pond layers in VGG16, and addition global pool layer, batch normalization layer, Dropout layers and connect entirely
Connect layer；The convolution kernel size of the convolutional layer is 3*3 pixel, using ReLU activation primitive；The size of maximum value pond layer
For 2*2 pixel, step-length is 2 pixels.
In a possible embodiment, the timber image authentication multilayer convolutional neural networks deep learning is using
Habit rate is 10-4, momentum be 0.9 stochastic gradient descent method be iterated.
In a possible embodiment, when the recognition accuracy of the multilayer convolutional neural networks is more than 95%,
Complete the training to the timber image authentication multilayer convolutional neural networks.
In a possible embodiment, the recognition result is the tree species information of timber, when the confidence level is greater than
When equal to 0.95, the timber to be identified is identified successfully.
Second aspect, the embodiment of the present invention provide a kind of timber identification system based on the study of construction feature picture depth,
Include:
Image data module, for acquiring wood transverse section construction feature image data to be identified, by described image data
It is divided into multiple image blocks of the same size；And described image data are established according to the multiple image block of the same size and are corresponded to
Training set and test set；
Timber image authentication multilayer convolutional neural networks module, for establishing described image number according to multiple described image blocks
According to corresponding training set and test set；Construct timber image authentication multilayer convolutional neural networks；Using the training set to described
Timber image authentication multilayer convolutional neural networks carry out deep learning；It is surveyed using model of the test set to deep learning
Examination generates wood structure image recognition deep learning algorithm model according to test result Optimized model parameter；
Picture recognition module for identifying to Wood structure features image data to be identified, and exports recognition result
And confidence level.
In a possible embodiment, described image data module, the resolution ratio of the image data of acquisition are 2048*
2048, depth is 8, color mode RGB；The resolution ratio of image block is 512*512, quantity 2000；Image data is corresponding
The training set image number of blocks accounting 80% of image set, test set image number of blocks accounting 20%.
In a possible embodiment, the timber image authentication multilayer convolutional neural networks module, for constructing
Multilayer convolutional neural networks VGG16；The multilayer convolutional neural networks VGG16 is instructed in advance using ImageNet data set
Practice；Timber image authentication multilayer convolutional neural networks are constructed according to the pre-training VGG16.
In a possible embodiment, the timber image authentication convolutional neural networks module is by input layer, convolution
Layer, maximum value pond layer, global pool layer, batch normalization layer, Dropout layers, full articulamentum and output layer are constituted；The convolution
The convolution kernel size of layer is 3*3 pixel, using ReLU activation primitive；The size of maximum value pond layer is 2*2 pixel, step-length
For 2 pixels.
In a possible embodiment, described image identification module is by deep learning algorithm model to wood to be identified
The construction image of material identified, recognition result is timber varieties of trees information, when the confidence level is more than or equal to 0.95, it is described to
Identify timber to identify successfully.
Timber discrimination method and system provided in an embodiment of the present invention based on the study of construction feature picture depth, passes through figure
It as data module, obtains good quality wood cross section and constructs image, create training set and test set after carrying out image segmentation；Building
Multilayer convolutional neural networks VGG16；Using ImageNet data set pre-training multilayer convolutional neural networks VGG16；It is basic herein
On, construct timber image authentication multilayer convolutional neural networks；Using wood structure training set of images to timber image authentication multilayer
Convolutional neural networks carry out deep learning, test simultaneously Optimal Parameters to deep learning model using test set；To construct
Generalization ability is strong, and the good deep learning algorithm model of robustness can accurately identify wood structure image to be identified, solution
The problems such as traditional timber of having determined image-recognizing method image characteristics extraction difficulty and recognition accuracy are low.Meanwhile there is operation letter
It is single, easy to use, the advantages that recognition accuracy is high, and recognition speed is fast, and robustness is good, and generalization ability is strong, can customs inspection,
The fields such as timber-trade and wood identification are widely applied.
Detailed description of the invention
Fig. 1 is a kind of stream of timber discrimination method based on the study of construction feature picture depth provided in an embodiment of the present invention
Journey schematic diagram；
Fig. 2 is a kind of multilayer convolutional neural networks structural schematic diagram provided in an embodiment of the present invention；
Fig. 3 is a kind of knot of timber identification system based on the study of construction feature picture depth provided in an embodiment of the present invention
Structure schematic diagram.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
In order to facilitate understanding of embodiments of the present invention, it is further explained below in conjunction with attached drawing with specific embodiment
Bright, embodiment does not constitute the restriction to the embodiment of the present invention.
Fig. 1 is that a kind of process of discrimination method based on the study of construction feature picture depth provided in an embodiment of the present invention is shown
It is intended to, is mainly used in 6 kinds of Pterocarpus timber (red sandalwood, dyestuff red sandalwood, India red sandalwood, hedgehog red sandalwood, African padank, peace
Brother draws red sandalwood) as shown in Figure 1, this method specifically includes:
The cross section construction image data of S101, acquisition 6 kinds of Pterocarpus timber to be identified.
Wherein, described image data are the construction image of 6 kinds of Dalbergia timber to be identified, point of described image data
Resolution is 2048*2048, and depth is 8, color mode RGB；Quantity is that each tree species acquire 500.
S102, described image data are divided into multiple image blocks of the same size.
Described image block is that resolution ratio is 512*512, the image block that quantity is 2000.
S103, the corresponding training set of described image data and test set are established according to multiple described image blocks.
Image block after segmentation is allocated at random, 80% is training set, and 20% is test set.
S104, building multilayer convolutional neural networks VGG16.
Referring to Fig. 2, the multilayer convolutional neural networks VGG16 includes: 1 input layer, 16 convolutional layers, 5 maximum values
Pond layer and 3 full articulamentums.
S105, pre-training multilayer convolutional neural networks VGG16.
The multilayer convolutional neural networks VGG16 pre-training is developed using Google open source system TensorFlow,
And algorithm acceleration is carried out using NVIDIA GPU.
S106, building timber image authentication multilayer convolutional neural networks.
Referring to Fig. 2, retain 2-11 layer (7 convolutional layers and the 3 maximum value pond of the multilayer convolutional neural networks VGG16
Change layer), and global pool layer, batch normalization layer, Dropout layers and full articulamentum are added, construct 6 kinds of Pterocarpus timber image mirror
Other multilayer convolutional neural networks.
S107, deep learning is carried out to the timber image authentication multilayer convolutional neural networks using the training set.
The timber image authentication multilayer convolutional neural networks deep learning uses learning rate for 10-4, momentum be 0.9 with
Machine gradient descent method is iterated.
S108, it is tested using model of the test set to deep learning, according to test result Optimized model parameter,
Generate the image recognition deep learning algorithm model of the timber to be identified.
The recognition accuracy of multilayer convolutional neural networks is tested using the test set and carries out arameter optimization, until the net
Until the recognition accuracy of network reaches 95% or more, the weight of the multilayer convolutional neural networks is finally determined, construct deep learning
Algorithm model.
S109, identify that deep learning algorithm model carries out 6 kinds of Pterocarpuses timber image data according to described image
Identification, obtains recognition result and confidence level.
The recognition result is the tree species information of timber to be identified, and the confidence level is more than or equal to 0.95 to identify successfully.
Timber discrimination method provided in an embodiment of the present invention based on the study of construction feature picture depth, passes through 6 kinds of acquisition
Pterocarpus wood structure image data constructs training set and test set after carrying out image segmentation；Construct multilayer convolutional neural networks
VGG16；Pass through ImageNet pre-training convolutional neural networks VGG16；On this basis, building identifies 6 kinds of Pterocarpus timber
Multilayer convolutional neural networks；Multilayer convolutional neural networks model is carried out using 6 kinds of Pterocarpus wood structure training set of images deep
Degree study is tested and is optimized using test the set pair analysis model.It is strong to construct generalization ability, the good deep learning algorithm mould of robustness
Type can accurately identify 6 kinds of Pterocarpus timber.Meanwhile there is easy to operate, easy to use, recognition accuracy height, knowledge
The advantages that other speed is fast, and robustness is good, and generalization ability is strong, can be wide in fields such as customs inspection, timber-trade and wood identifications
General application.
Fig. 3 is a kind of knot of timber identification system based on the study of construction feature picture depth provided in an embodiment of the present invention
Structure schematic diagram, as shown in figure 3, the system specifically includes:
Image data module 301, for acquiring wood transverse section construction feature image data to be identified, by described image number
According to being divided into multiple image blocks of the same size；And described image data pair are established according to the multiple image block of the same size
The training set and test set answered；
Timber image authentication multilayer convolutional neural networks module 302, for establishing the figure according to multiple described image blocks
As the corresponding training set of data and test set；Construct timber image authentication multilayer convolutional neural networks；Using the training set pair
The timber image authentication multilayer convolutional neural networks carry out deep learning；Using the test set to the model of deep learning into
Row test generates wood structure image recognition deep learning algorithm model according to test result Optimized model parameter；
Picture recognition module 303 for identifying to Wood structure features image data to be identified, and exports identification knot
Fruit and confidence level.
In a possible embodiment, the resolution ratio of described image data module 301, the image data of acquisition is
2048*2048, depth are 8, color mode RGB；The resolution ratio of the image block of image segmentation is 512*512, and quantity is
2000；Image set module training set accounting 80% (1600), test set accounting 20% (400).
In a possible embodiment, the timber image authentication multilayer convolutional neural networks module, for constructing
Multilayer convolutional neural networks VGG16；The multilayer convolutional neural networks VGG16 is instructed in advance using ImageNet data set
Practice；Timber image authentication multilayer convolutional neural networks are constructed according to the pre-training VGG16.
In a possible embodiment, the timber image authentication multilayer convolutional neural networks module by input layer,
Convolutional layer, maximum value pond layer, global pool layer, batch normalization layer, Dropout layers, full articulamentum and output layer are constituted；It is described
The convolution kernel size of convolutional layer is 3*3 pixel, using ReLU activation primitive；The size of maximum value pond layer is 2*2 pixel,
Step-length is 2 pixels.
In a possible embodiment, the recognition result is the tree species information of timber to be identified, the confidence level
It is to identify successfully more than or equal to 0.95.
Timber identification system provided in an embodiment of the present invention based on the study of construction feature picture depth, image data module
For acquiring wood transverse section construction image data to be identified；Obtain multiple images comprising wood transverse section construction feature
Block；Construct the corresponding training set of described image data and test set；Timber image authentication multilayer convolutional neural networks module is used for
Building identifies the deep learning algorithm model of timber；Picture recognition module is for knowing the construction image of timber to be identified
Not.This system image data module, which can get, largely covers the multifarious image block of wood transverse section construction feature；Timber image
Identify multilayer convolutional neural networks module to be made of multilayer convolutional neural networks, can automate to extract from wood structure image and know
Other feature；Picture recognition module can quick and precisely be identified to timber is identified.The simple, convenient carrying of this system, knowledge
Other accuracy rate is high, recognition speed is fast, can be widely applied in fields such as customs inspection, timber-trade and wood identifications.
Professional should further appreciate that, described in conjunction with the examples disclosed in the embodiments of the present disclosure
Unit and algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, hard in order to clearly demonstrate
The interchangeability of part and software generally describes each exemplary composition and step according to function in the above description.
These functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.
Professional technician can use different methods to achieve the described function each specific application, but this realization
It should not be considered as beyond the scope of the present invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can be executed with hardware, processor
The combination of software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only memory
(ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field
In any other form of storage medium well known to interior.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (15)
1. a kind of timber discrimination method based on the study of construction feature picture depth characterized by comprising
Acquire wood transverse section construction feature image data to be identified；
Described image data are divided into multiple image blocks of the same size；
The corresponding training set of described image data and test set are established according to multiple described image blocks；
Construct timber image authentication multilayer convolutional neural networks；
Deep learning is carried out to the timber image authentication multilayer convolutional neural networks using the training set；
It is tested using model of the test set to deep learning, according to test result Optimized model parameter, generates timber
Construct image recognition deep learning algorithm model；
Wood structure image data to be identified is identified using described image identification deep learning algorithm model, output identification
And confidence level as a result.
2. the method according to claim 1, wherein the building timber image authentication multilayer convolutional Neural net
Network, comprising:
Construct multilayer convolutional neural networks VGG16；
Pre-training is carried out to the multilayer convolutional neural networks VGG16 using ImageNet data set；
Timber image authentication multilayer convolutional neural networks are constructed according to the pre-training VGG16.
3. the method according to claim 1, wherein the acquisition wood transverse section construction feature image to be identified
Data, comprising:
The RGB image of the cross section construction feature of 500 each tree species is acquired by image capture module；
Wherein, the resolution ratio of described image is 2048*2048, depth is 8.
4. quantity is the method according to claim 1, wherein it is 512*512 that described image block, which is resolution ratio,
2000 image block.
5. the method according to claim 1, wherein the image number of blocks accounting of the training set be 80%, institute
The image number of blocks accounting for stating test set is 20%.
6. the method according to claim 1, wherein the multilayer convolutional neural networks VGG16 includes: 1 defeated
Enter layer, 13 convolutional layers, 5 maximum value pond layers and 1 output layer.
7. the method according to claim 1, wherein the timber image authentication multilayer convolutional neural networks, packet
It includes: 7 convolutional layers and 3 maximum value pond layers in the pre-training VGG16, and global pool layer, batch normalizing of addition
Change layer, Dropout layers and full articulamentum；The convolution kernel size of the convolutional layer is 3*3 pixel, using ReLU activation primitive；Institute
The size for stating maximum value pond layer is 2*2 pixel, and step-length is 2 pixels.
8. the method according to claim 1, wherein the timber image authentication multilayer convolutional neural networks depth
Study uses learning rate for 10-4, momentum be 0.9 stochastic gradient descent method be iterated.
9. the method according to claim 1, wherein when the timber image authentication multilayer convolutional neural networks
When recognition accuracy is more than 95%, the training to the timber image authentication multilayer convolutional neural networks is completed.
10. the method according to claim 1, wherein the recognition result be timber to be identified tree species information,
When the confidence level is more than or equal to 0.95, the timber to be identified is identified successfully.
11. a kind of timber identification system based on the study of construction feature picture depth characterized by comprising
Image data module divides described image data for acquiring wood transverse section construction feature image data to be identified
For multiple image blocks of the same size；And the corresponding instruction of described image data is established according to the multiple image block of the same size
Practice collection and test set；
Timber image authentication multilayer convolutional neural networks module, for establishing described image data pair according to multiple described image blocks
The training set and test set answered；Construct timber image authentication multilayer convolutional neural networks；Using the training set to the timber
Image authentication multilayer convolutional neural networks carry out deep learning；It is tested using model of the test set to deep learning,
According to test result Optimized model parameter, wood structure image recognition deep learning algorithm model is generated；
Picture recognition module for identifying to Wood structure features image data to be identified, and exports recognition result and sets
Reliability.
12. system according to claim 11, which is characterized in that point of the image data of described image data module acquisition
Resolution is 2048*2048；The resolution ratio that described image divides the image block of module segmentation is 512*512, and quantity 2000 is described
Training set image number of blocks accounting 80%, the test set image number of blocks accounting 20%.
13. system according to claim 11, which is characterized in that the timber image authentication multilayer convolutional neural networks mould
Block, for constructing multilayer convolutional neural networks VGG16；Using ImageNet data set to the multilayer convolutional neural networks
VGG16 carries out pre-training；Timber image authentication multilayer convolutional neural networks are constructed according to the pre-training VGG16.
14. system according to claim 11, which is characterized in that the timber image authentication convolutional neural networks module by
Input layer, convolutional layer, maximum value pond layer, global pool layer, batch normalization layer, Dropout layers, full articulamentum and output layer structure
At.The convolution kernel size of the convolutional layer is 3*3 pixel, using ReLU activation primitive；The size of maximum value pond layer is
2*2 pixel, step-length are 2 pixels.
15. system according to claim 11, which is characterized in that the recognition result is timber varieties of trees information, when described
When confidence level is more than or equal to 0.95, the timber to be identified is identified successfully.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810800080.7A CN109063713A (en) | 2018-07-20 | 2018-07-20 | A kind of timber discrimination method and system based on the study of construction feature picture depth |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810800080.7A CN109063713A (en) | 2018-07-20 | 2018-07-20 | A kind of timber discrimination method and system based on the study of construction feature picture depth |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109063713A true CN109063713A (en) | 2018-12-21 |
Family
ID=64817618
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810800080.7A Pending CN109063713A (en) | 2018-07-20 | 2018-07-20 | A kind of timber discrimination method and system based on the study of construction feature picture depth |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109063713A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858521A (en) * | 2018-12-29 | 2019-06-07 | 国际竹藤中心 | A kind of bamboo category identification method based on artificial intelligence deep learning |
| CN110059549A (en) * | 2019-03-11 | 2019-07-26 | 齐鲁工业大学 | A kind of thin wood plate categorizing system and algorithm based on deep learning |
| CN110503051A (en) * | 2019-08-27 | 2019-11-26 | 西南林业大学 | A kind of precious timber identifying system and method based on image recognition technology |
| CN111144244A (en) * | 2019-12-13 | 2020-05-12 | 唐美(北京)文化科技发展有限公司 | Furniture wood type identification method and device |
| CN111652307A (en) * | 2020-05-29 | 2020-09-11 | 广西大学 | Intelligent nondestructive identification method and device for redwood furniture based on convolutional neural network |
| CN112149630A (en) * | 2020-10-20 | 2020-12-29 | 云南省林业和草原科学院 | Wood microscopic identification system |
| CN112767387A (en) * | 2021-01-29 | 2021-05-07 | 中华人民共和国张家港海关 | Automatic wood image identification method based on block gradient weighting |
| CN112991312A (en) * | 2021-03-29 | 2021-06-18 | 中国农业科学院果树研究所 | Pear variety seedling identification method based on AI deep learning |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005078652A1 (en) * | 2004-02-12 | 2005-08-25 | Carl Henrik Grunditz | Method, device, computer program product and integrated circuit for surface inspection using a multi-tier neural network |
| CN106462549A (en) * | 2014-04-09 | 2017-02-22 | 尹度普有限公司 | Authenticating physical objects using machine learning from microscopic variations |
| CN107392896A (en) * | 2017-07-14 | 2017-11-24 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of Wood Defects Testing method and system based on deep learning |
-
2018
- 2018-07-20 CN CN201810800080.7A patent/CN109063713A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005078652A1 (en) * | 2004-02-12 | 2005-08-25 | Carl Henrik Grunditz | Method, device, computer program product and integrated circuit for surface inspection using a multi-tier neural network |
| CN106462549A (en) * | 2014-04-09 | 2017-02-22 | 尹度普有限公司 | Authenticating physical objects using machine learning from microscopic variations |
| CN107392896A (en) * | 2017-07-14 | 2017-11-24 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of Wood Defects Testing method and system based on deep learning |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858521A (en) * | 2018-12-29 | 2019-06-07 | 国际竹藤中心 | A kind of bamboo category identification method based on artificial intelligence deep learning |
| CN110059549A (en) * | 2019-03-11 | 2019-07-26 | 齐鲁工业大学 | A kind of thin wood plate categorizing system and algorithm based on deep learning |
| CN110503051A (en) * | 2019-08-27 | 2019-11-26 | 西南林业大学 | A kind of precious timber identifying system and method based on image recognition technology |
| CN110503051B (en) * | 2019-08-27 | 2023-04-07 | 西南林业大学 | Precious wood identification system and method based on image identification technology |
| CN111144244A (en) * | 2019-12-13 | 2020-05-12 | 唐美(北京)文化科技发展有限公司 | Furniture wood type identification method and device |
| CN111652307A (en) * | 2020-05-29 | 2020-09-11 | 广西大学 | Intelligent nondestructive identification method and device for redwood furniture based on convolutional neural network |
| CN112149630A (en) * | 2020-10-20 | 2020-12-29 | 云南省林业和草原科学院 | Wood microscopic identification system |
| CN112767387A (en) * | 2021-01-29 | 2021-05-07 | 中华人民共和国张家港海关 | Automatic wood image identification method based on block gradient weighting |
| CN112767387B (en) * | 2021-01-29 | 2024-04-30 | 中华人民共和国张家港海关 | Automatic wood image recognition method based on block gradient weighting |
| CN112991312A (en) * | 2021-03-29 | 2021-06-18 | 中国农业科学院果树研究所 | Pear variety seedling identification method based on AI deep learning |
| CN112991312B (en) * | 2021-03-29 | 2023-11-07 | 中国农业科学院果树研究所 | Pear variety seedling identification method based on AI deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109063713A (en) | A kind of timber discrimination method and system based on the study of construction feature picture depth | |
| US11113809B2 (en) | Group sparsity model for image unmixing | |
| CN109766872B (en) | Image recognition method and device | |
| CN108537191B (en) | Three-dimensional face recognition method based on structured light camera | |
| CN104573688B (en) | Mobile platform tobacco laser code intelligent identification Method and device based on deep learning | |
| CN108596046A (en) | A kind of cell detection method of counting and system based on deep learning | |
| CN105760439B (en) | A kind of personage's cooccurrence relation map construction method based on specific behavior co-occurrence network | |
| CN109815785A (en) | A kind of face Emotion identification method based on double-current convolutional neural networks | |
| CN108334848A (en) | A kind of small face identification method based on generation confrontation network | |
| CN110222701A (en) | A kind of bridge defect automatic identifying method | |
| CN109190643A (en) | Based on the recognition methods of convolutional neural networks Chinese medicine and electronic equipment | |
| CN109117897A (en) | Image processing method, device and readable storage medium storing program for executing based on convolutional neural networks | |
| US20150186755A1 (en) | Systems and Methods for Object Identification | |
| CN106897681A (en) | A kind of remote sensing images comparative analysis method and system | |
| CN107292346B (en) | A kind of MR image hippocampus partitioning algorithm based on Local Subspace study | |
| CN110110610A (en) | A kind of event detecting method for short-sighted frequency | |
| CN108073947A (en) | A kind of method for identifying blueberry kind | |
| CN106529470A (en) | Gesture recognition method based on multistage depth convolution neural network | |
| Weaver et al. | From leaves to labels: Building modular machine learning networks for rapid herbarium specimen analysis with LeafMachine2 | |
| CN109145955B (en) | A kind of Wood Identification Method and system | |
| Pondenkandath et al. | Identifying cross-depicted historical motifs | |
| CN110399804A (en) | A kind of food inspection recognition methods based on deep learning | |
| CN114118186A (en) | Calligraphy image style classification method based on directional feature enhancement | |
| CN103617417A (en) | Automatic plant identification method and system | |
| CN110660068B (en) | Semi-automatic brain region segmentation method for three-dimensional cell construction image |
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: 20181221 |
|
| RJ01 | Rejection of invention patent application after publication |