CN110956182B - Method for detecting water area shoreline change based on deep learning - Google Patents
Method for detecting water area shoreline change based on deep learning Download PDFInfo
- Publication number
- CN110956182B CN110956182B CN201910898713.7A CN201910898713A CN110956182B CN 110956182 B CN110956182 B CN 110956182B CN 201910898713 A CN201910898713 A CN 201910898713A CN 110956182 B CN110956182 B CN 110956182B
- Authority
- CN
- China
- Prior art keywords
- water area
- feature
- area shoreline
- feature mapping
- water
- 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.)
- Active
Links
Classifications
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method for detecting water area shoreline change based on deep learning, which comprises the following steps of firstly, establishing a water area shoreline change data set; respectively inputting the two images of the water area shoreline change into a ResNet 101 feature extractor to respectively obtain a feature mapping graph A1 and a feature mapping graph A2; then, performing connection operation according to the channel depth, and performing convolution operation to obtain a feature mapping chart B; then do it separately、、Andperforming maximum pooling operation to obtain feature maps C1, C2, C3 and C4 respectively; then do the following stepsPerforming convolution operation to respectively obtain feature maps D1, D2, D3 and D4; finally, respectively carrying out bilinear interpolation up-sampling operation and connection operation according to channel depth to obtain a feature mapping chart F; make the feature map F
Description
Technical Field
The invention relates to the water conservancy field and the image processing field, in particular to a method for detecting water area shoreline change based on deep learning.
Background
The water area shoreline is an important component of the river channel environment and also an important barrier for ensuring flood safety and flood control safety. For a long time, the water area management of China mainly takes manual inspection of a man-boat as a main factor. By manual inspection, illegal occupation or shoreline occupation in the water area management (protection) range is inspected on site. For the area with wide water area range, the work load of monitoring the water area shoreline is large, and if a manual patrol method is used, the comprehensive and timely monitoring is difficult to ensure; in addition, for some water areas with complex water situations, vehicles and ships are difficult to go deep into the water areas, and the inspection difficulty is high; in addition, the inspection cost of the method is high. Aiming at the outstanding problems existing in the method, the water control concept and the policy background under the new situation are combined, the water area shoreline management protection idea is cleared, the comprehensiveness of monitoring the water area shoreline is improved, the change of the water area shoreline is sensed in time, and the water management method is a necessary requirement of water conservancy management work.
The traditional method for manually patrolling the water area shoreline has certain defects in application: for the area with wide water area range, the monitoring workload of the water area shoreline is large, and if a manual inspection method is used, the comprehensive and timely monitoring is difficult to guarantee; in addition, for some water areas with complex water situations, vehicles and ships are difficult to go deep into the water areas, and the inspection difficulty is high; in addition, the inspection cost of the method is high.
Disclosure of Invention
In view of this, the present invention provides a method for detecting a water area shoreline change based on deep learning, so as to implement automatic real-time monitoring of the water area shoreline and greatly improve the comprehensiveness and realizability of river channel supervision.
The invention is realized by adopting the following scheme: a method for detecting water area shoreline change based on deep learning comprises the following steps:
step S1: establishing a water area shoreline change data set;
step S2: inputting a group of picture pairs in the water area shoreline change data set, namely an input image I1 and an input image I2 of two water area shoreline changes into a ResNet 101 feature extractor respectively to obtain a feature mapping graph A1 and a feature mapping graph A2 respectively;
and step S3: connecting the characteristic mapping graph A1 and the characteristic mapping graph A2 obtained in the step S2 according to the channel depth, and then performing convolution operation of 3 multiplied by 3 to obtain a characteristic mapping graph B;
and step S4: performing maximal pooling operations of 3 × 3, 5 × 5, 7 × 7 and 9 × 9 on the feature map B obtained in the step S3 to obtain feature maps C1, C2, C3 and C4 respectively; respectively performing 1 × 1 convolution operation on the feature maps C1, C2, C3 and C4 to respectively obtain feature maps D1, D2, D3 and D4;
step S5: respectively performing bilinear interpolation upsampling operation on the feature maps D1, D2, D3 and D4 obtained in the step S4 to enable the length and width of the feature maps D1, D2, D3 and D4 after upsampling to be consistent with the length and width of the feature map B; connecting the four feature mapping images obtained by up-sampling with the feature mapping image B according to the channel depth to obtain a feature mapping image F;
step S6: and (5) performing 1 × 1 convolution operation on the feature mapping F obtained in the step (S5) to obtain a final water area shoreline change result mask segmentation map for detecting the water area shoreline change.
Further, the specific content of step S1 is: the method comprises the steps of collecting pictures of water area shoreline changes, forming a group of picture pairs by two pictures at different time of the same place, and carrying out manual marking to manufacture a water area shoreline change data set.
Further, the manual labeling is performed on the pictures in the data set, and the labeled contents are as follows: covering the mask of all water areas in the picture; and comparing the masks marked on the two pictures in the group of picture pairs to obtain the change of the water area shoreline of the picture pair.
Compared with the prior art, the invention has the following beneficial effects:
the invention realizes the automatic real-time monitoring of the water area shoreline and greatly improves the comprehensiveness and the realizability of river channel supervision.
Detailed Description
The present invention will be further described with reference to the following examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiment provides a method for detecting water area shoreline change based on deep learning, which comprises the following steps:
step S1: establishing a water area shoreline change data set;
step S2: inputting a group of picture pairs in the water area shoreline change data set, namely two image input images I1 and I2 of the water area shoreline change into a ResNet 101 feature extractor respectively to obtain a feature mapping graph A1 and a feature mapping graph A2 respectively;
and step S3: connecting the characteristic mapping graph A1 and the characteristic mapping graph A2 obtained in the step S2 according to the channel depth, and then performing convolution operation of 3 multiplied by 3 to obtain a characteristic mapping graph B; (the length, width, and depth of the feature map B are the same as those of the feature maps A1 and A2)
And step S4: performing maximal pooling operations of 3 × 3, 5 × 5, 7 × 7 and 9 × 9 on the feature map B obtained in the step S3 to obtain feature maps C1, C2, C3 and C4 respectively; respectively performing 1 × 1 convolution operation on the feature maps C1, C2, C3 and C4 to respectively obtain feature maps D1, D2, D3 and D4;
step S5: respectively performing bilinear interpolation upsampling operation on the feature maps D1, D2, D3 and D4 obtained in the step S4 to enable the length and width of the feature maps D1, D2, D3 and D4 after upsampling to be consistent with the length and width of the feature map B; connecting the four feature mapping images obtained by up-sampling with the feature mapping image B according to the channel depth to obtain a feature mapping image F;
step S6: and (5) performing 1 × 1 convolution operation on the feature mapping F obtained in the step (S5) to obtain a final water area shoreline change result mask segmentation map for detecting the water area shoreline change.
In this embodiment, the specific content of step S1 is: collecting pictures of water area shoreline changes, forming a group of picture pairs by two pictures at different time of the same place, and manually marking the picture pairs to manufacture a water area shoreline change data set.
In this embodiment, the manual labeling is performed on the pictures in the data set, and the labeled contents are as follows: covering the mask of all water areas in the picture; and obtaining the change of the water area shoreline of the picture pair by comparing the masks marked on the two pictures in the group of picture pairs.
Preferably, in this embodiment, the relevant pictures of the water area shoreline change are collected, two pictures at different times at the same place form a group of picture pairs, and manual labeling is performed to manufacture the water area shoreline change data set.
Based on diversification in natural environment, the pictures in the data set take the following elements into consideration in the collection process: (1) two pictures of a group of picture pairs are shot at different times in the same place, and the shooting angles are consistent; (2) the water area bank line is changed; (3) interference of different illumination intensities; (4) interference of the background; (5) interference from different weather conditions, etc.
Preferably, the present embodiment can overcome the defects of the conventional manual detection method, and the detection of the water area shoreline is completed fully automatically by using the latest artificial intelligence technology. By applying the embodiment, the automatic real-time monitoring of the water area shoreline can be realized only by using a common camera and matching with a background algorithm; the method can be directly used for replacing background algorithm service on the existing water gauge monitoring system, and is convenient to install and simple in configuration. The comprehensiveness and the realizability of river channel supervision are greatly improved.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.
Claims (3)
1. A method for detecting water area shoreline change based on deep learning is characterized in that: the method comprises the following steps:
step S1: establishing a water area shoreline change data set;
step S2: inputting a group of picture pairs in the water area shoreline change data set, namely an input image I1 and an input image I2 of two water area shoreline changes into a ResNet 101 feature extractor respectively to obtain a feature mapping graph A1 and a feature mapping graph A2 respectively;
and step S3: connecting the feature mapping A1 and the feature mapping A2 obtained in the step S2 according to the channel depth, and then performingObtaining a feature map B by convolution operation;
and step S4: respectively making the feature mapping maps B obtained in the step S3、/>、/>And &>Performing maximum pooling operation to obtain feature maps C1, C2, C3 and C4 respectively; the feature maps C1, C2, C3 and C4 are each evaluated>Performing convolution operation to respectively obtain feature maps D1, D2, D3 and D4;
step S5: respectively performing bilinear interpolation upsampling operation on the feature maps D1, D2, D3 and D4 obtained in the step S4 to enable the length and the width of the feature maps after the feature maps D1, D2, D3 and D4 are upsampled to be consistent with those of the feature map B; connecting the four feature mapping images obtained by up-sampling with the feature mapping image B according to the channel depth to obtain a feature mapping image F;
2. The method for detecting the water bank line change based on the deep learning of claim 1, wherein: the specific content of the step S1 is as follows: the method comprises the steps of collecting pictures of water area shoreline changes, forming a group of picture pairs by two pictures at different time of the same place, and carrying out manual marking to manufacture a water area shoreline change data set.
3. The method for detecting the water bank line change based on the deep learning of claim 2, wherein: the collected pictures of the water area shoreline change are marked manually, and the marked contents are as follows: covering the mask of all water areas in the picture; and comparing the masks marked on the two pictures in the group of picture pairs to obtain the change of the water area shoreline of the picture pair.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910898713.7A CN110956182B (en) | 2019-09-23 | 2019-09-23 | Method for detecting water area shoreline change based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910898713.7A CN110956182B (en) | 2019-09-23 | 2019-09-23 | Method for detecting water area shoreline change based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110956182A CN110956182A (en) | 2020-04-03 |
CN110956182B true CN110956182B (en) | 2023-04-07 |
Family
ID=69976310
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910898713.7A Active CN110956182B (en) | 2019-09-23 | 2019-09-23 | Method for detecting water area shoreline change based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110956182B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115439748B (en) * | 2022-09-13 | 2023-09-26 | 中山大学 | Coastline erosion degree monitoring method and device and electronic equipment |
CN115661438B (en) * | 2022-10-21 | 2024-03-08 | 广州地铁集团有限公司 | Off-site water level monitoring and flood prevention early warning method based on machine vision |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102867187A (en) * | 2012-07-04 | 2013-01-09 | 西安电子科技大学 | NSST (NonsubsampledShearlet Transform) domain MRF (Markov Random Field) and adaptive threshold fused remote sensing image change detection method |
CN103106658A (en) * | 2013-01-23 | 2013-05-15 | 中国人民解放军信息工程大学 | Island or reef coastline rapid obtaining method |
KR101488214B1 (en) * | 2014-07-11 | 2015-01-30 | (주)지오시스템리서치 | Apparatus for monitoring geographical features change of intertidal zone using image pictured by camera and the method thereof |
CN106327455A (en) * | 2016-08-18 | 2017-01-11 | 中国科学院遥感与数字地球研究所 | Improved method for fusing remote-sensing multispectrum with full-color image |
CN108776772A (en) * | 2018-05-02 | 2018-11-09 | 北京佳格天地科技有限公司 | Across the time building variation detection modeling method of one kind and detection device, method and storage medium |
-
2019
- 2019-09-23 CN CN201910898713.7A patent/CN110956182B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102867187A (en) * | 2012-07-04 | 2013-01-09 | 西安电子科技大学 | NSST (NonsubsampledShearlet Transform) domain MRF (Markov Random Field) and adaptive threshold fused remote sensing image change detection method |
CN103106658A (en) * | 2013-01-23 | 2013-05-15 | 中国人民解放军信息工程大学 | Island or reef coastline rapid obtaining method |
KR101488214B1 (en) * | 2014-07-11 | 2015-01-30 | (주)지오시스템리서치 | Apparatus for monitoring geographical features change of intertidal zone using image pictured by camera and the method thereof |
CN106327455A (en) * | 2016-08-18 | 2017-01-11 | 中国科学院遥感与数字地球研究所 | Improved method for fusing remote-sensing multispectrum with full-color image |
CN108776772A (en) * | 2018-05-02 | 2018-11-09 | 北京佳格天地科技有限公司 | Across the time building variation detection modeling method of one kind and detection device, method and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN110956182A (en) | 2020-04-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110956182B (en) | Method for detecting water area shoreline change based on deep learning | |
CN105023014B (en) | A kind of shaft tower target extraction method in unmanned plane inspection transmission line of electricity image | |
CN105184779B (en) | One kind is based on the pyramidal vehicle multiscale tracing method of swift nature | |
CN104361314B (en) | Based on infrared and transformer localization method and device of visual image fusion | |
CN109377480A (en) | Arable land use change detection method based on deep learning | |
CN107679495B (en) | Detection method for movable engineering vehicles around power transmission line | |
CN112100301A (en) | Method for realizing dynamic monitoring of water area shoreline by using high-resolution remote sensing technology | |
CN101226637B (en) | Method for detecting automatically contact point of vehicle wheel and ground | |
CN110866494A (en) | Optical remote sensing image-based town group extraction method and system | |
CN102831393A (en) | Rapid image recognizing method of power tower pole outline | |
CN113989510B (en) | River drainage outlet overflow detection method and device and related equipment | |
CN105869164A (en) | Method and system for detecting on/off state of switch | |
CN111209923A (en) | Deep learning technology-based muck truck cover or uncover identification method | |
CN104318240A (en) | Flower bud discriminating method based on computer vision | |
CN114049347A (en) | Crude oil leakage detection method based on feature enhancement | |
CN109579934A (en) | Water logging multiple monitoring method and system | |
CN116884216A (en) | Road congestion judging method based on satellite remote sensing image | |
CN111985436A (en) | Workshop ground mark line identification fitting method based on LSD | |
CN108764287A (en) | Object detection method and system based on deep learning and grouping convolution | |
CN112489018A (en) | Intelligent power line inspection method and inspection line | |
CN111062926A (en) | Video data processing method and device and storage medium | |
CN116309370A (en) | Real-time detection method and system for apparent diseases of concrete dam | |
CN116229362A (en) | Intelligent ponding detection and depth estimation method based on real-time monitoring video | |
CN108510450A (en) | A kind of photo-irradiation treatment method and device of crop leaf image | |
CN111951166A (en) | Road green belt pruning task automatic generation method based on neural network and CIM |
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 |