CN109934095A - A kind of remote sensing images Clean water withdraw method and system based on deep learning - Google Patents
A kind of remote sensing images Clean water withdraw method and system based on deep learning Download PDFInfo
- Publication number
- CN109934095A CN109934095A CN201910059294.8A CN201910059294A CN109934095A CN 109934095 A CN109934095 A CN 109934095A CN 201910059294 A CN201910059294 A CN 201910059294A CN 109934095 A CN109934095 A CN 109934095A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- deep learning
- sensing images
- clean water
- 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.)
- Pending
Links
Abstract
The remote sensing images Clean water withdraw method and system based on deep learning that the invention discloses a kind of, this method is by utilizing China's high score -1/No. 2 (GF-1/GF-2) and resource -3 (ZY-3) satellite image, in conjunction with the advantage of convolutional neural networks (CNNs) structure of two kinds of deep learnings of U-net and Densenet, a kind of CNNs model of novel extraction water body is established.By being labeled to the water body in a large amount of high score remote sensing images, the result of water body mark in image is obtained as training set;High score remote sensing images are trained with the CNNs model of foundation and tests and obtains the binary image of water body;The pixel value identified in bianry image is identified to obtain water body target.The present invention can be realized Water-Body Information in high precision, in the rapidly extracting high score remote sensing image of automation, improves the intelligence, automation level and extraction accuracy of Clean water withdraw, can be widely applied for the relevant every field of Clean water withdraw.
Description
Technical field
The present invention relates to technical field of remote sensing image processing, more particularly to a kind of remote sensing images water based on deep learning
Body extracting method and system.
Background technique
With the development of China's space flight and aviation cause, earth observation acquisition terrestrial object information is also more and more, satellite image, boat
The image resolution ratio for clapping the acquisitions such as image is also higher and higher.The water body for extracting earth surface is always the basic of remote Sensing Image Analysis
Task.As the key components of hydrologic cycle, the range of water is for monitoring extreme event and the mechanism for understanding climate change
It is most important.
For region, the variation of water has significant impact to the ecosystem and human lives.To the standard of water position status
Really for successful quantitation assessment space-time monitoring, variation detection and flood warning etc. are very crucial for perception.Therefore, water level is effectively extracted
It is most important further to monitor its dynamic or quantify its volume.
Existing Clean water withdraw remote sensing technique depends on the complicated spectrum analysis limited by experience and through excessive
Amount trains obtained optimal threshold, and optimal threshold changes greatly different images, generally automation and intelligent level
It is not high, it is not able to satisfy high spatial resolution images automation, the requirement that high precision image is quickly handled, water body precisely detects.
Summary of the invention
The remote sensing images Clean water withdraw method and system based on deep learning that the purpose of the present invention is to provide a kind of, can
Nicety of grading and intelligent level are improved, precisely identifies surface water body information.
According to an aspect of the invention, there is provided a kind of remote sensing images Clean water withdraw method based on deep learning, packet
Include following steps:
Obtain remote sensing images;
Water body in part high score image in the remote sensing images is marked in advance, as training sample;
The deep learning model of convolutional neural networks CNNs is established, and according to the training sample to the deep learning mould
Type is trained;
Clean water withdraw is carried out to remote sensing images according to the deep learning model that training obtains, obtains binary image;
Clean water withdraw and target designation are carried out to the binary image.
The acquisition remote sensing images, comprising:
The remote sensing images of target area are obtained by satellite or aerial photography device.
Water body in the part high score image in the remote sensing images is marked in advance, comprising:
Manually amendment mark is combined according to supervised classification, obtains water body annotation results, water body and non-water body are in image
In pixel value it is different.
The deep learning model of the convolutional neural networks CNNs, comprising:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, establish convolutional neural networks CNNs's
Deep learning model.
The deep learning model of the convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes to adopt under space
Sample;
Output layer, for exporting the category score result.
It is described that the deep learning model is trained according to the training sample, comprising:
Convolutional neural networks initialization;
The calculating that moves ahead is carried out according to the training sample;
Using the principle adjustment weight and biasing for minimizing residual error.
The convolutional neural networks initialization, comprising:
Initialize convolution kernel, weight and the size of biasing.
It is described that the calculating that moves ahead is carried out according to the training sample, comprising:
The water body sample is exported in the form of vectors by input layer;
The inner product of itself and the input water body sample is calculated by convolutional layer, and the result of inner product is exported;
Aggregate statistics are carried out by result of the pond layer to the inner product, complete the down-sampling in space.
According to another aspect of the present invention, it provides a kind of remote sensing images water body based on deep learning and automatically extracts and be
System, including image acquisition unit, in advance mark unit, model training unit, extraction unit and calibration unit, wherein
Described image acquiring unit, for obtaining remote sensing images;
The pre- mark unit, for being marked in advance to the water body in the part high score image in the remote sensing images,
As training sample;
The model training unit, for establishing the deep learning model of convolutional neural networks CNNs, and according to the instruction
Practice sample to be trained the deep learning model;
The extraction unit, the deep learning model for being obtained according to training carry out Clean water withdraw to remote sensing images, obtain
To binary image;
The calibration unit, for carrying out Clean water withdraw and target designation to the binary image.
The model training unit, is also used to:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, establish convolutional neural networks CNNs's
Deep learning model;
The deep learning model of the convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes to adopt under space
Sample;
Output layer, for exporting the category score result.
Using technical solution of the present invention, by utilizing China's high score -1/No. 2 (GF-1/GF-2) and resource -3
(ZY-3) satellite image, in conjunction with the advantage of convolutional neural networks (CNNs) structure of two kinds of deep learnings of U-net and Densenet,
Establish a kind of CNNs model of novel extraction water body.By being labeled to the water body in a large amount of high score remote sensing images,
The result of water body mark in image is obtained as training set;Test is trained to high score remote sensing images with the CNNs model of foundation
And obtain the binary image of water body;The pixel value identified in bianry image is identified to obtain water body target.The present invention
It can be realized Water-Body Information in high precision, in the rapidly extracting high score remote sensing image of automation, improve the intelligence of Clean water withdraw
Change, the gentle extraction accuracy of Automated water, can be widely applied for the relevant every field of Clean water withdraw.
The present invention establishes a kind of novel convolutional neural networks CNNs structure, combines U-net and two kinds of Densenet
The structural advantage feature of convolutional neural networks, thus original image is input in the convolutional neural networks of training completion, thus
Automation, the extracted with high accuracy for realizing water body can be used for water transportation rapid drafting, water body variation detection, resource and environment tune
Look into equal fields.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Detailed description of the invention
Attached drawing is used to provide further understanding of the present invention, and constitutes part of specification, with reality of the invention
It applies example to be used to explain the present invention together, not be construed as limiting the invention.In the accompanying drawings:
Fig. 1 is the remote sensing images Clean water withdraw Method And Principle flow chart based on deep learning in the embodiment of the present invention one;
Fig. 2 is the remote sensing images Clean water withdraw method and step schematic diagram based on deep learning in the embodiment of the present invention one;
Fig. 3 is the step flow chart of convolutional neural networks training process in the embodiment of the present invention one;
Fig. 4 is the remote sensing images Clean water withdraw system structure diagram based on deep learning in the embodiment of the present invention two.
Specific embodiment
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings, it should be understood that preferred reality described herein
Apply example only for the purpose of illustrating and explaining the present invention and is not intended to limit the present invention.
Deep learning is to obtain new knowledge expertise using computer simulation mankind learning behavior, is reorganized existing
The structure of knowledge continuouslys optimize knowledge base, finally makes optimizing decision.Deep learning image recognition is to utilize convolutional neural networks
A pocket is randomly selected from image as training sample, learns the feature to Some features information from the sample, so
Make operation for these features as filter and original image afterwards, to obtain the different characteristic in original image in any position
Activation value, then value is inputted in classifier and is trained, it can be achieved that image classification, finally the modes such as made an uproar through connected region, filter
Nicety of grading can be improved, to identify terrestrial object information.
Fig. 1 is the remote sensing images Clean water withdraw Method And Principle flow chart based on deep learning in the embodiment of the present invention one.Such as
Shown in Fig. 1, should remote sensing images Clean water withdraw process based on deep learning the following steps are included:
Step 101 obtains remote sensing images.
The remote sensing images of target area are obtained by satellite or aerial photography device.
Step 102 marks the water body in the part high score image in the remote sensing images in advance, as training sample
This.
Step 103, the deep learning model for establishing convolutional neural networks CNNs, and according to the training sample to the depth
Degree learning model is trained.
Step 104 carries out Clean water withdraw to remote sensing images according to the deep learning model that training obtains, and obtains binary picture
Picture.
Step 105 carries out Clean water withdraw and target designation to the binary image.
Specifically, the water body in the part high score image in the remote sensing images is marked in advance, comprising:
Manually amendment mark is combined according to supervised classification, obtains water body annotation results, water body and non-water body are in image
In pixel value it is different.
The deep learning model of the convolutional neural networks CNNs, comprising:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, establish convolutional neural networks CNNs's
Deep learning model.
The deep learning model of the convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes to adopt under space
Sample;
Output layer, for exporting the category score result.
It is described that the deep learning model is trained according to the training sample, comprising:
Convolutional neural networks initialization;
The calculating that moves ahead is carried out according to the training sample;
Using the principle adjustment weight and biasing for minimizing residual error.
The convolutional neural networks initialization, comprising:
Initialize convolution kernel, weight and the size of biasing.
It is described that the calculating that moves ahead is carried out according to the training sample, comprising:
The water body sample is exported in the form of vectors by input layer;
The inner product of itself and the input water body sample is calculated by convolutional layer, and the result of inner product is exported;
Aggregate statistics are carried out by result of the pond layer to the inner product, complete the down-sampling in space.
As shown in Fig. 2, for the remote sensing images Clean water withdraw method and step signal in the embodiment of the present invention based on deep learning
Figure.Wherein,
The first step obtains remote sensing images by satellite or aerial photography device;
Second step is labeled the water body in high score image, as training sample;
Third step establishes new deep learning convolutional neural networks CNNs model, in conjunction with U-net and two kinds of Densenet
The advantage of model, convolutional neural networks specifically include that
Input layer, for exporting input picture in the form of vectors
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes to adopt under space
Sample;
Output layer, for exporting the category score result.
4th step marks sample using the deep learning model and water body of foundation, is trained test to high score image;Such as
Shown in Fig. 3, wherein the training process are as follows:
S1, convolutional neural networks initialization;
S2, the calculating that moves ahead is carried out;
S3, weight and biasing are adjusted using the principle for minimizing residual error.
Wherein, in S1, convolutional neural networks initialization mainly includes initializing convolution kernel, weight and the size of biasing,
The calculating that move ahead described in S2 specifically includes that
S21, the water body sample is exported in the form of vectors by input layer;
S22, the inner product that itself and the input water body sample are calculated by convolutional layer, and the result of inner product is exported;
S23, aggregate statistics are carried out by result of the pond layer to the inner product, completes the down-sampling in space.
5th step, the deep learning neural network model obtained with training carry out Clean water withdraw to high score remote sensing images and obtain
To binary image;
6th step carries out water body target designation to binary image.
The embodiment of the present invention establishes a kind of CNNs deep learning model of novel high score remote sensing Clean water withdraw;
It is labelled with a large amount of remote sensing images and obtains water body annotation results;
Multi- source Remote Sensing Data data is trained using the deep learning CNNs of foundation;
Clean water withdraw is carried out to multi-source high score remote sensing images with the CNNs that training obtains and obtains bianry image;
The pixel value identified in bianry image is identified;
The CNNs of foundation can be applied to GF-1/-2, ZY-3 multi-source high-resolution remote sensing image;
Realize the high-precision of water body, the extraction of automation.
The deep learning model is convolutional neural networks (CNNs).
Obtained a large amount of remote sensing images mark includes being combined manually amendment to mark according to supervised classification to obtain
Water body annotation results, water body are different from the pixel value of non-water body in the picture.
It is with the water body result of mark for input that the deep learning, which extracts the training test of water body, using the CNNs of foundation
Model is trained.
Method of the present invention and traditional water body index method NDWI method are compared, to analyze nicety of grading,
Evaluated using three Kappa coefficient, F-1 value and overall accuracy indexs two methods as a result, being shown in Table 1.
The traditional remote sensing water body index method (NDWI) of table 1 and deep learning method accuracy comparison
The atural objects such as cloud present in image, shade, massif and building can influence nicety of grading, but pass through table 1
Know that two methods can come out water body information, method of the present invention is in the consistency (Kappa for extracting water body
Coefficient) and two disaggregated model accuracy (combining the accuracy rate of disaggregated model and the F-1 score of recall rate) on it is all significantly high
In NDWI water body method, and the overall accuracy ratio NDWI high 2.4% of method of the present invention.
As shown in figure 4, being a kind of remote sensing images Clean water withdraw system based on deep learning provided in an embodiment of the present invention
Structural schematic diagram, including image acquisition unit 201, pre- mark unit 202, model training unit 203, extraction unit 204 and mark
Order member 205, wherein
Described image acquiring unit 201, for obtaining remote sensing images;
The pre- mark unit 202, for being marked in advance to the water body in the part high score image in the remote sensing images
Note, as training sample;
The model training unit 203, for establishing the deep learning model of convolutional neural networks CNNs, and according to described
Training sample is trained the deep learning model;
The extraction unit 204, the deep learning model for being obtained according to training carry out Clean water withdraw to remote sensing images,
Obtain binary image;
The calibration unit 205, for carrying out Clean water withdraw and target designation to the binary image.
The model training unit 203, is also used to:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, establish convolutional neural networks CNNs's
Deep learning model;
The deep learning model of the convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes to adopt under space
Sample;
Output layer, for exporting the category score result.
The each embodiment of the present invention, by utilizing China's high score -1/No. 2 (GF-1/GF-2) and-No. 3 (ZY- of resource
3) satellite image is built in conjunction with the advantage of convolutional neural networks (CNNs) structure of two kinds of deep learnings of U-net and Densenet
A kind of CNNs model of novel extraction water body is found.By being labeled to the water body in a large amount of high score remote sensing images, obtain
The result that water body marks into image is as training set;Test is trained simultaneously to high score remote sensing images with the CNNs model of foundation
Obtain the binary image of water body;The pixel value identified in bianry image is identified to obtain water body target.Energy of the present invention
It is enough realize high-precision, automation rapidly extracting high score remote sensing image in Water-Body Information, improve Clean water withdraw intelligence,
The gentle extraction accuracy of Automated water can be widely applied for the relevant every field of Clean water withdraw.
The present invention establishes a kind of novel convolutional neural networks CNNs structure, combines U-net and two kinds of Densenet
The structural advantage feature of convolutional neural networks, thus original image is input in the convolutional neural networks of training completion, thus
Automation, the extracted with high accuracy for realizing water body can be used for water transportation rapid drafting, water body variation detection, resource and environment tune
Look into equal fields.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The shape for the computer program product implemented in usable storage medium (including but not limited to magnetic disk storage and optical memory etc.)
Formula.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that can be realized by computer program instructions each in flowchart and/or the block diagram
The combination of process and/or box in process and/or box and flowchart and/or the block diagram.It can provide these computers
Processor of the program instruction to general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices
To generate a machine, so that generating use by the instruction that computer or the processor of other programmable data processing devices execute
In the dress for realizing the function of specifying in one or more flows of the flowchart and/or one or more blocks of the block diagram
It sets.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of remote sensing images Clean water withdraw method based on deep learning, which comprises the following steps:
Obtain remote sensing images;
Water body in part high score image in the remote sensing images is marked in advance, as training sample;
Establish the deep learning model of convolutional neural networks CNNs, and according to the training sample to the deep learning model into
Row training;
Clean water withdraw is carried out to remote sensing images according to the deep learning model that training obtains, obtains binary image;
Clean water withdraw and target designation are carried out to the binary image.
2. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 1, which is characterized in that institute
State acquisition remote sensing images, comprising:
The remote sensing images of target area are obtained by satellite or aerial photography device.
3. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 1, which is characterized in that institute
It states and the water body in the part high score image in the remote sensing images is marked in advance, comprising:
Manually amendment mark is combined according to supervised classification, obtains water body annotation results, water body and non-water body are in the picture
Pixel value is different.
4. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 1, which is characterized in that institute
State the deep learning model of convolutional neural networks CNNs, comprising:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, the depth of convolutional neural networks CNNs is established
Learning model.
5. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 4, which is characterized in that institute
State the deep learning model of convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes the down-sampling in space;
Output layer, for exporting the category score result.
6. a kind of remote sensing images Clean water withdraw method based on deep learning, feature exist according to claim 1 or 5
In described to be trained according to the training sample to the deep learning model, comprising:
Convolutional neural networks initialization;
The calculating that moves ahead is carried out according to the training sample;
Using the principle adjustment weight and biasing for minimizing residual error.
7. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 6, which is characterized in that institute
State convolutional neural networks initialization, comprising:
Initialize convolution kernel, weight and the size of biasing.
8. a kind of remote sensing images Clean water withdraw method based on deep learning according to claim 7, which is characterized in that institute
It states and the calculating that moves ahead is carried out according to the training sample, comprising:
The water body sample is exported in the form of vectors by input layer;
The inner product of itself and the input water body sample is calculated by convolutional layer, and the result of inner product is exported;
Aggregate statistics are carried out by result of the pond layer to the inner product, complete the down-sampling in space.
9. a kind of remote sensing images Clean water withdraw system based on deep learning, which is characterized in that including image acquisition unit, pre- mark
Infuse unit, model training unit, extraction unit and calibration unit, wherein
Described image acquiring unit, for obtaining remote sensing images;
The pre- mark unit, for being marked in advance to the water body in the part high score image in the remote sensing images, as
Training sample;
The model training unit, for establishing the deep learning model of convolutional neural networks CNNs, and according to the trained sample
This is trained the deep learning model;
The extraction unit, the deep learning model for being obtained according to training carry out Clean water withdraw to remote sensing images, obtain two
Value image;
The calibration unit, for carrying out Clean water withdraw and target designation to the binary image.
10. a kind of remote sensing images Clean water withdraw system based on deep learning according to claim 9, which is characterized in that
The model training unit, is also used to:
In conjunction with the structure feature of two kinds of convolutional neural networks of U-net and Densenet, the depth of convolutional neural networks CNNs is established
Learning model;
The deep learning model of the convolutional neural networks CNNs, comprising:
Input layer, for exporting input picture in the form of vectors;
Convolutional layer is calculated its inner product with image by way of sliding window, and the result of inner product is exported;
Pond layer is located at after convolutional layer, carries out aggregate statistics for the result to the inner product, completes the down-sampling in space;
Output layer, for exporting the category score result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910059294.8A CN109934095A (en) | 2019-01-22 | 2019-01-22 | A kind of remote sensing images Clean water withdraw method and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910059294.8A CN109934095A (en) | 2019-01-22 | 2019-01-22 | A kind of remote sensing images Clean water withdraw method and system based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109934095A true CN109934095A (en) | 2019-06-25 |
Family
ID=66985183
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910059294.8A Pending CN109934095A (en) | 2019-01-22 | 2019-01-22 | A kind of remote sensing images Clean water withdraw method and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109934095A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111104889A (en) * | 2019-12-04 | 2020-05-05 | 山东科技大学 | Water body remote sensing identification method based on U-net |
CN111127484A (en) * | 2019-12-25 | 2020-05-08 | 北京小白世纪网络科技有限公司 | Liver CT image segmentation method and device based on deep learning neural network |
CN111539965A (en) * | 2020-04-10 | 2020-08-14 | 中国资源卫星应用中心 | Urban surface water domestic high-resolution remote sensing self-adaptive extraction method cooperating with multi-source information |
CN111914686A (en) * | 2020-07-15 | 2020-11-10 | 云南电网有限责任公司带电作业分公司 | SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition |
CN112560716A (en) * | 2020-12-21 | 2021-03-26 | 浙江万里学院 | High-resolution remote sensing image water body extraction method based on low-level feature fusion |
CN113297965A (en) * | 2021-05-25 | 2021-08-24 | 浙江树人学院(浙江树人大学) | Remote sensing image target extraction method and device based on deep learning |
CN115423829A (en) * | 2022-07-29 | 2022-12-02 | 江苏省水利科学研究院 | Method and system for rapidly extracting water body from single-band remote sensing image |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105809140A (en) * | 2016-03-18 | 2016-07-27 | 华南农业大学 | Method and device for extracting surface water body information based on remote sensing model |
CN108052940A (en) * | 2017-12-17 | 2018-05-18 | 南京理工大学 | SAR remote sensing images waterborne target detection methods based on deep learning |
CN108229425A (en) * | 2018-01-29 | 2018-06-29 | 浙江大学 | A kind of identifying water boy method based on high-resolution remote sensing image |
CN108256424A (en) * | 2017-12-11 | 2018-07-06 | 中交信息技术国家工程实验室有限公司 | A kind of high-resolution remote sensing image method for extracting roads based on deep learning |
CN108710862A (en) * | 2018-05-24 | 2018-10-26 | 河海大学 | A kind of high-resolution remote sensing image Clean water withdraw method |
CN108921225A (en) * | 2018-07-10 | 2018-11-30 | 深圳市商汤科技有限公司 | A kind of image processing method and device, computer equipment and storage medium |
-
2019
- 2019-01-22 CN CN201910059294.8A patent/CN109934095A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105809140A (en) * | 2016-03-18 | 2016-07-27 | 华南农业大学 | Method and device for extracting surface water body information based on remote sensing model |
CN108256424A (en) * | 2017-12-11 | 2018-07-06 | 中交信息技术国家工程实验室有限公司 | A kind of high-resolution remote sensing image method for extracting roads based on deep learning |
CN108052940A (en) * | 2017-12-17 | 2018-05-18 | 南京理工大学 | SAR remote sensing images waterborne target detection methods based on deep learning |
CN108229425A (en) * | 2018-01-29 | 2018-06-29 | 浙江大学 | A kind of identifying water boy method based on high-resolution remote sensing image |
CN108710862A (en) * | 2018-05-24 | 2018-10-26 | 河海大学 | A kind of high-resolution remote sensing image Clean water withdraw method |
CN108921225A (en) * | 2018-07-10 | 2018-11-30 | 深圳市商汤科技有限公司 | A kind of image processing method and device, computer equipment and storage medium |
Non-Patent Citations (2)
Title |
---|
SIMON JEGOU´ 等: "The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation", 《2017 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS(CVPRW)》 * |
毛金莲著: "《智能图像检索关键技术研究》", 30 June 2015, 北京理工大学出版社 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111104889A (en) * | 2019-12-04 | 2020-05-05 | 山东科技大学 | Water body remote sensing identification method based on U-net |
CN111104889B (en) * | 2019-12-04 | 2023-09-05 | 山东科技大学 | U-net-based water remote sensing identification method |
CN111127484A (en) * | 2019-12-25 | 2020-05-08 | 北京小白世纪网络科技有限公司 | Liver CT image segmentation method and device based on deep learning neural network |
CN111539965A (en) * | 2020-04-10 | 2020-08-14 | 中国资源卫星应用中心 | Urban surface water domestic high-resolution remote sensing self-adaptive extraction method cooperating with multi-source information |
CN111539965B (en) * | 2020-04-10 | 2024-03-26 | 中国资源卫星应用中心 | Urban surface water domestic high-resolution remote sensing self-adaptive extraction method with cooperation of multi-source information |
CN111914686A (en) * | 2020-07-15 | 2020-11-10 | 云南电网有限责任公司带电作业分公司 | SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition |
CN111914686B (en) * | 2020-07-15 | 2022-10-18 | 云南电网有限责任公司带电作业分公司 | SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition |
CN112560716A (en) * | 2020-12-21 | 2021-03-26 | 浙江万里学院 | High-resolution remote sensing image water body extraction method based on low-level feature fusion |
CN113297965A (en) * | 2021-05-25 | 2021-08-24 | 浙江树人学院(浙江树人大学) | Remote sensing image target extraction method and device based on deep learning |
CN115423829A (en) * | 2022-07-29 | 2022-12-02 | 江苏省水利科学研究院 | Method and system for rapidly extracting water body from single-band remote sensing image |
CN115423829B (en) * | 2022-07-29 | 2024-03-01 | 江苏省水利科学研究院 | Method and system for rapidly extracting water body of single-band remote sensing image |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109934095A (en) | A kind of remote sensing images Clean water withdraw method and system based on deep learning | |
CN110059694B (en) | Intelligent identification method for character data in complex scene of power industry | |
JP6866495B2 (en) | Image quality evaluation method and image quality evaluation system | |
CN110070074B (en) | Method for constructing pedestrian detection model | |
CN107330453B (en) | Pornographic image identification method based on step-by-step identification and fusion key part detection | |
CN106971152B (en) | Method for detecting bird nest in power transmission line based on aerial images | |
CN110929577A (en) | Improved target identification method based on YOLOv3 lightweight framework | |
CN109117876A (en) | A kind of dense small target deteection model building method, model and detection method | |
CN109740676B (en) | Object detection and migration method based on similar targets | |
CN105741267B (en) | The multi-source image change detecting method of cluster guidance deep neural network classification | |
CN105550678A (en) | Human body motion feature extraction method based on global remarkable edge area | |
CN105718879A (en) | Free-scene egocentric-vision finger key point detection method based on depth convolution nerve network | |
CN105160310A (en) | 3D (three-dimensional) convolutional neural network based human body behavior recognition method | |
CN111950453A (en) | Optional-shape text recognition method based on selective attention mechanism | |
CN106204638A (en) | A kind of based on dimension self-adaption with the method for tracking target of taking photo by plane blocking process | |
CN109284779A (en) | Object detecting method based on the full convolutional network of depth | |
CN104599292A (en) | Noise-resistant moving target detection algorithm based on low rank matrix | |
CN106780546B (en) | The personal identification method of motion blur encoded point based on convolutional neural networks | |
CN106408030A (en) | SAR image classification method based on middle lamella semantic attribute and convolution neural network | |
CN107301376B (en) | Pedestrian detection method based on deep learning multi-layer stimulation | |
CN107247952B (en) | Deep supervision-based visual saliency detection method for cyclic convolution neural network | |
CN111353377A (en) | Elevator passenger number detection method based on deep learning | |
CN114511710A (en) | Image target detection method based on convolutional neural network | |
Li et al. | Dept: depth estimation by parameter transfer for single still images | |
CN107133631A (en) | A kind of method and device for recognizing TV station's icon |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190625 |