CN109993071A - The method and system of discoloration forest automatic identification and investigation based on remote sensing image - Google Patents
The method and system of discoloration forest automatic identification and investigation based on remote sensing image Download PDFInfo
- Publication number
- CN109993071A CN109993071A CN201910190167.1A CN201910190167A CN109993071A CN 109993071 A CN109993071 A CN 109993071A CN 201910190167 A CN201910190167 A CN 201910190167A CN 109993071 A CN109993071 A CN 109993071A
- Authority
- CN
- China
- Prior art keywords
- forest
- discoloration
- remote sensing
- image data
- sensing image
- 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.)
- Granted
Links
Classifications
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The method and system of the present invention provides a kind of discoloration forest automatic identification and investigation based on remote sensing image.High-resolution forest land remote sensing image is obtained using unmanned aerial vehicle remote sensing observation technology, by DeCNN-DCRF model to discoloration forest automatic identification；Recognition result is connected with field work, and recognition result is made really to serve real work, improves the efficiency of pest management and preventing and controlling.
Description
Technical field
The present invention relates to belong to information technology field more particularly to Modern Forestry field.
Background technique
High-incidence, multiple situation is persistently presented in China's pine nematode in recent years, and elimination is not yet eradicated in original epidemic-stricken area, new to send out
Epidemic-stricken area is continuously increased.Increasingly serious prevention and control situation, there is an urgent need to associated mechanisms to improve monitoring, diagnosis density, enhancing processing energy
Power, as far as possible by Risk health behavior in minimum zone or initial phase.In the great attention further aspect of country and local governments at all levels
Current forest pest control work has been achieved for very big effect, although on the other hand traditional ' the light monitoring of prevention and treatment again '
Theory has had some changes but forest monitoring and prediction work is also very weak, and technical system is not perfect, and ground investigation works
Electronic working method is entered from papery, but operating efficiency still has greatly improved space, most area cannot be complete
Accurately and effectively grasp the breaking-out and situation of change of epidemic situation in time or in advance using technological means entirely.Especially for the field of investigation
Fast automatic identification, especially to single plant discoloration forest location information the problem of, do not efficiently solved.
Summary of the invention
Present invention seek to address that problem as described above.Specifically, the present invention provides a kind of discoloration based on remote sensing image
Forest automatic identification and the method for investigation.
The method of discoloration forest automatic identification and investigation based on remote sensing image, comprising:
Acquire the image data of monitoring section；
The image data is analyzed by DeCNN-DCRF model, obtains the recognition result of discoloration forest；
The recognition result is loaded onto the space coordinates of the image data, determines the spatial position number of discoloration forest
According to；
On-site inspection sampling is carried out according to the spatial position data.
Wherein DeCNN-DCRF model includes: to the analysis of image data
The feature class probability value of the image data is extracted by the convolutional network of DeCNN；
The full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Binary function is constructed, the recognition result of discoloration forest is obtained.
On-site inspection sampling includes: to carry out investigation and sampling on the spot, according to the sky of discoloration forest using intelligent mobile terminal
Between position data carry out navigator fix, record forest tarnish reasons；The Disposal Measures of discoloration forest；Obtain the picture of discoloration forest
Information and geographic position data.
According to on-site inspection sampling result, the remote sensing image for the forest that changes colour is marked；To monitoring section discoloration forest feelings
Condition carries out statistics and summarizes.
After the remote sensing image of discoloration forest is marked, it imported into DeCNN-DCRF model sample library.
Image data can carry out the identification of single plant discoloration forest.
The system of discoloration forest automatic identification and investigation based on remote sensing image, comprising:
Image input module: discoloration forest spy is sent to for obtaining monitoring section image data, and by the image data
Sign is extracted and position computation module；
Discoloration forest feature extraction and position computation module: analyzing image data using DeCNN-DCRF model,
The discoloration forest feature in the image data is extracted, the recognition result of discoloration forest is obtained；Image number is loaded to recognition result
According to space coordinates, determine discoloration forest spatial position；
Field investigation module: utilizing mobile terminal, according to the spatial position of discoloration forest, carries out discoloration forest on-site inspection
Sampling；
Data management module: being managed on-site inspection sampled data, and sends statistical analysis mould for management result
Block；
Statistical analysis module: statistics discoloration forest distributed areas, quantity, severity, and related data information is carried out
Analysis, forms the Technical Analysis Report of Monitoring Result.
Change colour in forest feature extraction and position computation module, DeCNN-DCRF model includes: to the analysis of image data
The feature class probability value of the image data is extracted by the convolutional network of DeCNN；
The full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Binary function is constructed, the recognition result of discoloration forest is obtained.
It includes: record forest tarnish reasons that field investigation module, which carries out on-site inspection sampling to discoloration wood,；Record discoloration woods
The Disposal Measures of wood；Obtain the pictorial information and geographic position data of discoloration forest.
Data management module marks the remote sensing image of the discoloration forest according to the on-site inspection sampling result
Note, and the remote sensing image after label is imported into DeCNN-DCRF model sample library.
The present invention obtains high-resolution forest land remote sensing image using unmanned aerial vehicle remote sensing observation technology, passes through DeCNN-DCRF
Recognition result especially for the quick identification of single tree, and is directly applied to field to discoloration forest automatic identification by model
In investigation work, improves the working efficiency of forestry discoloration forest monitoring, improves the accuracy of discoloration forest identification, mention simultaneously
The high timeliness of the prevention and control of plant diseases, pest control and management work.
Being described below for exemplary embodiment is read with reference to the drawings, other property features of the invention and advantage will
It is apparent from.
Detailed description of the invention
It is incorporated into specification and the attached drawing for constituting part of specification shows the embodiment of the present invention, and with
Principle for explaining the present invention together is described.In the drawings, similar appended drawing reference is for indicating similar element.Under
Attached drawing in the description of face is some embodiments of the present invention, rather than whole embodiments.Those of ordinary skill in the art are come
It says, without creative efforts, other drawings may be obtained according to these drawings without any creative labor.
Fig. 1 is the flow chart of discoloration forest automatic identification and investigation method；
Fig. 2 is the structure chart of discoloration forest automatic identification and investigating system.
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.It needs
Illustrate, in the absence of conflict, the features in the embodiments and the embodiments of the present application can mutual any combination.
One aspect of the present invention, a method of discoloration forest automatic identification and investigation, comprising:
Step 101: acquiring the image data of monitoring section；
Step 102: the image data being analyzed by DeCNN-DCRF model, obtains the identification knot of discoloration forest
Fruit；
Step 103: the recognition result being loaded onto the space coordinates of the image data, determines the sky of discoloration forest
Between position data；
Step 104: on-site inspection sampling is carried out according to the spatial position data.
In DeCNN-DCRF model, DeCNN is the abbreviation of Deconvolutional Neural Network, i.e. deconvolution
Neural network；DCRF is the abbreviation of Dense Conditional Random Field, i.e., full condition of contact random field.
Wherein, in step 101, image data can be obtained in any manner, for example, can be by carrying picture pick-up device
The unmanned aerial vehicle remote sensing images data of monitoring section forest are obtained according to predetermined flight plan with the unmanned plane of data transmission set.
In step 102, the image data is analyzed by DeCNN-DCRF model, obtains the identification of discoloration forest
As a result, wherein DeCNN-DCRF model may include: to the analysis of image data
Step 2-1: the feature class probability value of image data obtained in the convolutional network extraction step 1 by DeCNN；
Step 2-2: the full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Step 2-3: building binary function obtains the recognition result of discoloration forest.
Wherein, in step 2-1, DeCNN model includes positive convolutional network and deconvolution network.Network finally exports and input
The identical probability graph of image length and width indicates that each pixel is under the jurisdiction of the probability value P (y of defined classificationi)。
In step 2-2, according to above-mentioned class probability value P (yi) obtain unitary item: φi(yi)=- logP (yi), by unitary
Item φi(yi) it is passed to the full condition of contact random field energy function of DCRF,
Step 2-3: building binary function: φij=δ (yi≠yj) k (i, j), wherein k (i, j) is expressed as two Gaussian kernels
The linear combination of function:Wherein, v(m)It is linear group
Weight, the as parameter to be estimated of this are closed, δ () is label indicator function.k(m)(i, j) is Gaussian kernel, whereinpiFor the spatial position of pixel i, fiFor pixel i
Spectral signature.FormulaIn previous item be space length measurement, latter description is special
Similitude is levied, so that the similar pixel of closely located and feature is higher labeled as of a sort probability, this has reacted DCRF
The connection of long range between the pixel of middle foundation；FormulaFor smooth segmentation result, pass through distance
Measurement can remove isolated area to carry out local smoothing method work, reduce noise.Parameters in Formula λα,λβ,λγControl binary gesture
Influence degree of each metric form to result in function.Finally obtain segmentation result:In formula
" Z " be normalization factor.P (y | x) is the segmentation result of x point pixel in image, and y is class label (discoloration forest, non-discoloration
Forest), value is 1 or 0, and P (y | x) is the probability of y.According to most probable value, y value, i.e. segmentation result are determined.When y is 1
When indicate segmentation result be discoloration forest, when y value be 0 when, segmentation result be non-discoloration forest.According to segmentation result P (y | x),
The segmentation result of available each pixel, i.e., the corresponding forest of each pixel are the segmentation of normal forest or the forest that changes colour
As a result, to obtain the recognition result of discoloration forest.
There are 5 convolutional layers, 3 pond layers, 2 full articulamentums, warp in DeCNN model convolutional network part in the present embodiment
Product network portion is the mirror image to positive conventional part, has a series of corresponding anti-ponds, warp lamination.Pass through a series of companies
Continuous anti-pondization and deconvolution operates, and final DeCNN can produce to distinguish different classes of semantic information, generate dense
The class probability figure of Pixel-level.Network exports probability graph identical with input image length and width finally to indicate that each pixel is under the jurisdiction of
The probability of defined classification.
In step 103, by the recognition result of step 2 final output, i.e., probability graph identical with input image length and width, load
The space coordinates of input image data can determine the coordinate position of the corresponding pixel of discoloration forest in recognition result,
So that it is determined that the spatial position data of discoloration forest；
Step 104: on-site inspection sampling is carried out according to the spatial position data；The discoloration forest that will be obtained in step 3
Spatial position data be downloaded to field investigation and verify functional module intelligent mobile terminal on, with the sky for the forest that changes colour
Between position carry out investigation and sampling on the spot, record the cause of death, Disposal Measures as navigation data, and obtain discoloration forest
Pictorial information.
On-site inspection sampling includes: by intelligent mobile terminal (such as mobile phone, PDA, tablet computer), according to discoloration forest
Spatial position data carry out navigator fix, record forest tarnish reasons；The Disposal Measures of discoloration forest；Obtain discoloration forest
Pictorial information records on-site inspection geographical coordinate.
According to on-site inspection sampling result, discoloration forest is marked in remote sensing image, to monitoring section discoloration forest
Situation carries out statistics and summarizes.Determine epidemic situation distribution situation according to the remote sensing monitoring data of verification, calculate withered tree distributed areas,
Quantity, severity, and related data information is extracted and analyzed, summarize the Technical Analysis Report for forming Monitoring Result.
After the remote sensing image of discoloration forest is marked, it imported into DeCNN-DCRF model sample library.For DeCNN-
For DCRF model, the sample of study is abundanter, and the accuracy rate of automatic identification is higher, in the automatic identification mistake of discoloration forest
Sample after label is constantly imported DeCNN-DCRF model sample library by Cheng Zhong, and DeCNN-DCRF model is allowed to continue to learn
It practises, the accuracy of identification is continuously improved.
Image data is the high resolution image data obtained based on unmanned aerial vehicle remote sensing, can carry out the knowledge of single plant discoloration forest
Not.In the present embodiment, using resolution ratio in the unmanned aerial vehicle remote sensing images of 0.1m, training data is specific to single plant trees and delineates
Sample data, the algorithm trained is also specially to be split identification to single plant trees, so as to realize single plant change colour woods
The identification of wood.
Another aspect of the present invention, a kind of system of change colour forest automatic identification and investigation, comprising:
Image input module: discoloration forest spy is sent to for acquiring monitoring section image data, and by the image data
Sign is extracted and position computation module；By carrying the unmanned plane of picture pick-up device and data transmission set according to predetermined in the present embodiment
The unmanned aerial vehicle remote sensing images data of flight plan acquisition monitoring section.And image data is sent to discoloration forest feature extraction and position
Set computing module；
Change colour forest feature extraction and position computation module: the image data that image input module obtains is received, by image
Data input DeCNN-DCRF model, analyze image data, extract the discoloration forest feature in image data, are become
The recognition result of color forest.To the space coordinates of the discoloration forest load raw video in image, the sky of discoloration forest is determined
Between position；
Field investigation module: utilizing mobile terminal, according to the spatial position of discoloration forest, carries out discoloration forest on-site inspection
Sampling；
Data management module: being managed on-site inspection sampled data, sends statistical analysis module for management result；
Statistical analysis module: statistics discoloration forest distributed areas, quantity, severity, and related data information is carried out
Analysis, forms the Technical Analysis Report of Monitoring Result.
Change colour in forest feature extraction and position computation module, DeCNN-DCRF model includes: to the analysis of image data
The feature class probability value of the image data is extracted by the convolutional network of DeCNN；
The full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Binary function is constructed, the recognition result of discoloration forest is obtained.
Field investigation module to discoloration wood carry out on-site inspection sampling include: by intelligent mobile terminal (such as mobile phone, PDA,
Tablet computer etc.) record forest tarnish reasons；On-site inspection geographical location；The Disposal Measures of record discoloration forest；Obtain discoloration
The pictorial information of forest.
The remote sensing image for the forest that changes colour is marked according to on-site inspection sampling result for data management module；Change colour woods
After the remote sensing image of wood is marked, it imported into DeCNN-DCRF model sample library.By constantly enriching DeCNN-DCRF mould
Type sample database, so that the accuracy of identification can be continuously improved with continuous learning in DeCNN-DCRF model.
Statistical analysis module carries out statistics to monitoring section discoloration forest situation and summarizes, anti-further to formulate forest zone pest and disease damage
It controls scheme and data foundation is provided.
In the present embodiment, image input module is based on unmanned aerial vehicle remote sensing and obtains high resolution image data, using resolution ratio
For the unmanned aerial vehicle remote sensing images of 0.1m, and training data is also specific to the sample data that isolated tree wood is delineated, and trains
Algorithm be also that identification specially is split to single plant trees, so as to realize single plant discoloration forest identification.
In conclusion the present invention uses unmanned aerial vehicle remote sensing observation technology, have the characteristics that monitor real-time, fast and accurate, energy
Enough to obtain high-resolution forest land remote sensing image well, low-altitude aerial remote sensing prison is implemented in key area especially big to area
It surveys, solves the problems such as human resources that prospecting faces are insufficient, coverage rate is low, low efficiency.By DeCNN-DCRF model to discoloration
Forest automatic identification realizes high discrimination particular for this subtle target of dead and drying tree, meanwhile, recognition result is no longer only
Only as it is final as a result, but it is connected with field work, serve recognition result really in real work, improve
The efficiency of pest management work；Using collection result as the data sample of present system, continual enrich is
In system database, to realize stepping up for accuracy.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment comprising a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence " including one ... ", it is not excluded that wrapping
Include in the process, method, article or equipment of the element that there is also other identical elements.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations.Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. the method for discoloration forest automatic identification and investigation based on remote sensing image, comprising:
Acquire the image data of monitoring section；
The image data is analyzed by DeCNN-DCRF model, obtains the recognition result of discoloration forest；
The recognition result is loaded onto the space coordinates of the image data, determines the spatial position data of discoloration forest；
On-site inspection sampling is carried out according to the spatial position data.
2. the method for discoloration forest automatic identification and investigation based on remote sensing image as described in claim 1, which is characterized in that
The DeCNN-DCRF model includes: to the analysis of image data
The feature class probability value of the image data is extracted by the convolutional network of DeCNN；
The full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Binary function is constructed, the recognition result of discoloration forest is obtained.
3. the method for discoloration forest automatic identification and investigation based on remote sensing image as described in claim 1, which is characterized in that
On-site inspection sampling includes: to carry out investigation and sampling on the spot, according to the space bit of discoloration forest using intelligent mobile terminal
It sets data and carries out navigator fix, record forest tarnish reasons, the Disposal Measures of record discoloration forest；Obtain the picture of discoloration forest
Information and geographic position data.
4. the method for discoloration forest automatic identification and investigation based on remote sensing image as claimed in claim 3, which is characterized in that
According to on-site inspection sampling result, the remote sensing image of the discoloration forest is marked；To monitoring section change colour forest situation into
Row statistics summarizes.
5. the method for discoloration forest automatic identification and investigation based on remote sensing image as claimed in claim 4, which is characterized in that
After the remote sensing image of the discoloration forest is marked, it imported into DeCNN-DCRF model sample library.
6. the method for discoloration forest automatic identification and investigation based on remote sensing image as described in claim 1, which is characterized in that
The image data can carry out the identification of single plant discoloration forest.
7. the system of discoloration forest automatic identification and investigation based on remote sensing image, comprising:
Image input module: for obtaining monitoring section image data, and the image data is sent to discoloration forest feature and is mentioned
It takes and position computation module；
Discoloration forest feature extraction and position computation module: analyzing image data using DeCNN-DCRF model, extracts
Discoloration forest feature in the image data obtains the recognition result of discoloration forest；To recognition result load image data
Space coordinates determine the spatial position of discoloration forest；
Field investigation module: utilizing mobile terminal, according to the spatial position of discoloration forest, carries out discoloration forest on-site inspection and takes
Sample；
Data management module: being managed on-site inspection sampled data, and sends statistical module for management result；
Statistical analysis module: statistics discoloration forest distributed areas, quantity, severity, and related data information is analyzed,
Form the Technical Analysis Report of Monitoring Result.
8. the system of discoloration forest automatic identification and investigation based on remote sensing image as claimed in claim 7, which is characterized in that
In the discoloration forest feature extraction and position computation module, DeCNN-DCRF model includes: to the analysis of image data
The feature class probability value of the image data is extracted by the convolutional network of DeCNN；
The full condition of contact random field of DCRF is passed to using the class probability value as unitary item；
Binary function is constructed, the recognition result of discoloration forest is obtained.
9. the system of discoloration forest automatic identification and investigation based on remote sensing image as claimed in claim 7, which is characterized in that
It includes: record forest tarnish reasons that the field investigation module, which carries out on-site inspection sampling to discoloration wood,；The disposition of discoloration forest
Measure；Obtain the pictorial information and geographic position data of discoloration forest.
10. the system of discoloration forest automatic identification and investigation based on remote sensing image, feature exist as claimed in claim 9
In, the data management module is marked the remote sensing image of the discoloration forest according to the on-site inspection sampling result,
And the remote sensing image after label is imported into DeCNN-DCRF model sample library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910190167.1A CN109993071B (en) | 2019-03-13 | 2019-03-13 | Method and system for automatically identifying and investigating color-changing forest based on remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910190167.1A CN109993071B (en) | 2019-03-13 | 2019-03-13 | Method and system for automatically identifying and investigating color-changing forest based on remote sensing image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109993071A true CN109993071A (en) | 2019-07-09 |
| CN109993071B CN109993071B (en) | 2021-05-18 |
Family
ID=67130616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910190167.1A Active CN109993071B (en) | 2019-03-13 | 2019-03-13 | Method and system for automatically identifying and investigating color-changing forest based on remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109993071B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111582176A (en) * | 2020-05-09 | 2020-08-25 | 湖北同诚通用航空有限公司 | Visible light remote sensing image withered and dead wood recognition software system and recognition method |
| CN113408474A (en) * | 2021-07-06 | 2021-09-17 | 中国科学院地理科学与资源研究所 | Under-forest composite ecological treatment method, device, medium and terminal equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740894A (en) * | 2016-01-28 | 2016-07-06 | 北京航空航天大学 | Semantic annotation method for hyperspectral remote sensing image |
| CN106708782A (en) * | 2016-12-22 | 2017-05-24 | 贾翔 | Regional pest detection diagnosis judging method based on wavelet analysis |
| CN107516103A (en) * | 2016-06-17 | 2017-12-26 | 北京市商汤科技开发有限公司 | A kind of image classification method and system |
| CN107590813A (en) * | 2017-10-27 | 2018-01-16 | 深圳市唯特视科技有限公司 | A kind of image partition method based on deep layer interactive mode geodesic distance |
| CN108694391A (en) * | 2018-05-16 | 2018-10-23 | 黄铁成 | Populus Euphratica spring looper disaster monitoring method based on high-spectrum remote-sensing |
-
2019
- 2019-03-13 CN CN201910190167.1A patent/CN109993071B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740894A (en) * | 2016-01-28 | 2016-07-06 | 北京航空航天大学 | Semantic annotation method for hyperspectral remote sensing image |
| CN107516103A (en) * | 2016-06-17 | 2017-12-26 | 北京市商汤科技开发有限公司 | A kind of image classification method and system |
| CN106708782A (en) * | 2016-12-22 | 2017-05-24 | 贾翔 | Regional pest detection diagnosis judging method based on wavelet analysis |
| CN107590813A (en) * | 2017-10-27 | 2018-01-16 | 深圳市唯特视科技有限公司 | A kind of image partition method based on deep layer interactive mode geodesic distance |
| CN108694391A (en) * | 2018-05-16 | 2018-10-23 | 黄铁成 | Populus Euphratica spring looper disaster monitoring method based on high-spectrum remote-sensing |
Non-Patent Citations (3)
| Title |
|---|
| CHRISTINE STONE等: "Application of Remote Sensing Technologies for Assessing Planted Forests Damaged by Insect Pests and Fungal Pathogens: a Review", 《CURRENT FORESTRY REPORTS》 * |
| WEI HU等: "Automatic recognition of diseased trees based on the vertical structure of airborne point clouds: A case study of diseased trees of Great Smoky Mountains", 《2016 4TH INTERNATIONAL WORKSHOP ON EARTH OBSERVATION AND REMOTE SENSING APPLICATIONS (EORSA)》 * |
| 任鹏洲: "高光谱遥感技术在林业监测中的应用", 《新科技》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111582176A (en) * | 2020-05-09 | 2020-08-25 | 湖北同诚通用航空有限公司 | Visible light remote sensing image withered and dead wood recognition software system and recognition method |
| CN113408474A (en) * | 2021-07-06 | 2021-09-17 | 中国科学院地理科学与资源研究所 | Under-forest composite ecological treatment method, device, medium and terminal equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109993071B (en) | 2021-05-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mei et al. | Hyperspectral image classification using group-aware hierarchical transformer | |
| US11521380B2 (en) | Shadow and cloud masking for remote sensing images in agriculture applications using a multilayer perceptron | |
| CN108846835B (en) | Image change detection method based on depth separable convolutional network | |
| CN111222545B (en) | Image classification method based on linear programming incremental learning | |
| CN111639587A (en) | Hyperspectral image classification method based on multi-scale spectrum space convolution neural network | |
| CN113657158B (en) | Google EARTH ENGINE-based large-scale soybean planting area extraction algorithm | |
| Chen et al. | Spectral unmixing using a sparse multiple-endmember spectral mixture model | |
| CN113378785A (en) | Forest type identification method and device | |
| Zhu et al. | Identifying carrot appearance quality by an improved dense CapNet | |
| Feng et al. | A novel saliency detection method for wild animal monitoring images with WMSN | |
| Farooq et al. | Transferable convolutional neural network for weed mapping with multisensor imagery | |
| Kerdegari et al. | Smart monitoring of crops using generative adversarial networks | |
| CN113807278A (en) | Deep learning-based land use classification and change prediction method | |
| Nie et al. | Adap-EMD: Adaptive EMD for aircraft fine-grained classification in remote sensing | |
| CN109993071A (en) | The method and system of discoloration forest automatic identification and investigation based on remote sensing image | |
| KR20200058278A (en) | Apparatus and method for analyzing spatio temporal data for geo-location | |
| Li et al. | HRVQA: A Visual Question Answering benchmark for high-resolution aerial images | |
| Srivastava et al. | Feature-Based Image Retrieval (FBIR) system for satellite image quality assessment using big data analytical technique | |
| Silverman et al. | Predicting origins of coherent air mass trajectories using a neural network—the case of dry intrusions | |
| CN114943290B (en) | Biological intrusion recognition method based on multi-source data fusion analysis | |
| Heidary et al. | Urban change detection by fully convolutional siamese concatenate network with attention | |
| KR102576427B1 (en) | Real-time Rainfall Prediction Device using Cloud Images, and Rainfall Prediction Method using the same, and a computer-readable storage medium | |
| CN112613371A (en) | Hyperspectral image road extraction method based on dense connection convolution neural network | |
| Peña et al. | Unmixing low-ratio endmembers in hyperspectral images through Gaussian synapse ANNs | |
| Keswani et al. | Land Cover Classification from Time Series Satellite Images |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |