CN116147883A - Ocean internal wave detection method based on buoy observation data - Google Patents
Ocean internal wave detection method based on buoy observation data Download PDFInfo
- Publication number
- CN116147883A CN116147883A CN202111349369.XA CN202111349369A CN116147883A CN 116147883 A CN116147883 A CN 116147883A CN 202111349369 A CN202111349369 A CN 202111349369A CN 116147883 A CN116147883 A CN 116147883A
- Authority
- CN
- China
- Prior art keywords
- internal
- wave
- flow velocity
- internal wave
- ocean
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M10/00—Hydrodynamic testing; Arrangements in or on ship-testing tanks or water tunnels
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C13/00—Surveying specially adapted to open water, e.g. sea, lake, river or canal
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Hydrology & Water Resources (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Radar, Positioning & Navigation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Fluid Mechanics (AREA)
- Indicating Or Recording The Presence, Absence, Or Direction Of Movement (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, and discloses a buoy observation data-based ocean internal wave detection method, which is characterized by comprising the following steps of: obtaining buoy profile ocean current observation data, drawing a flow velocity profile time sequence diagram, marking the beginning and ending range of internal waves displayed in the flow velocity profile time sequence diagram, constructing an ocean internal wave flow velocity profile image sample library, training a deep learning neural network model by using the ocean internal wave flow velocity profile image sample library, detecting the flow velocity profile time sequence diagram to be detected by using the trained deep learning neural network model, and calculating the occurrence and ending time of the detected internal waves. The method solves the problems of the traditional internal wave detection method based on remote sensing data through the buoy profile ocean current observation data, is not influenced by atmospheric environmental factors, carries out continuous real-time detection on internal waves in time, and can detect unobvious internal wave phenomena.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence, and particularly relates to a buoy observation data-based ocean internal wave detection method.
Background
The existing detection technology for ocean internal waves mainly depends on remote sensing image data, the internal wave detection technology developed based on remote sensing images is influenced by cloud quantity, precipitation and other atmospheric environment factors, detection can only be carried out under the condition of no cloud or little cloud, a common satellite orbit period is long, the latest remote sensing image can be obtained after waiting for tens of minutes or a few hours, so that great delay exists in detection of the internal waves, meanwhile, the resolution of the remote sensing image data is generally low, a remote sensing satellite can only observe a more obvious internal wave phenomenon, but internal wave activity occurs under water, the water surface phenomenon is not obvious, the detection of the internal waves is forgotten, and further safety threat to offshore operation is formed.
Disclosure of Invention
The invention provides a marine internal wave detection method based on buoy observation data, which takes buoy profile ocean current observation data as a data basis, utilizes a deep learning neural network model to complete prediction of marine internal waves, enables a buoy to be free from the influence of atmospheric environmental factors, can acquire latest data every few seconds or minutes, can acquire hydrological information such as temperature, salinity, depth, flow velocity and the like of local sea area underwater, can accurately observe various obvious or unobvious internal wave phenomena, provides an accurate data basis for prediction of the deep learning neural network model, and improves subsequent prediction precision.
The invention can be realized by the following technical scheme:
the ocean internal wave detection method based on buoy observation data comprises the steps of obtaining buoy profile ocean current observation data, drawing a flow velocity profile time sequence diagram, marking an internal wave starting and ending range displayed in the flow velocity profile time sequence diagram, constructing an ocean internal wave flow velocity profile image sample library, training a deep learning neural network model by using the ocean internal wave flow velocity profile image sample library, detecting a flow velocity profile time sequence diagram to be detected by using the trained deep learning neural network model, and calculating occurrence and ending moments of the detected internal wave.
Further, according to the self-characteristics of the flow velocity profile timing chart, the start and stop time of the flow velocity on the flow velocity profile timing chart is marked as (M, N), the image size thereof is W×H, and the detected internal wave position is marked as (x) i ,y i ,w i ,h i ) Wherein x is i ,y i Indicating the position coordinates, w, of the start of the detected internal wave i ,h i The width and height of a rectangular frame corresponding to the start/end range of the detected internal wave are expressed, and the start time m of the internal wave is calculated by the following equation i And end time n i 。
Further, marking an internal wave starting and ending range displayed by a section time sequence diagram at the flow speed by using a marking tool labelImg, wherein the marking tool comprises obvious internal waves and unobvious internal waves, the obvious internal waves are provided with an upper wave core and a lower wave core, the cores of the obvious internal waves are red, the upper wave core is dark blue, and the lower wave core is dark yellow; the unobvious internal wave is provided with two wave cores, the upper wave core is light blue, and the lower wave core is light yellow.
Further, the deep learning neural network model comprises a Faster RCNN model and a Yolov5 model, and Recall ratio Recall and average accuracy mean mAP are used as evaluation indexes.
The beneficial technical effects of the invention are as follows:
compared with the traditional internal wave detection and prediction method, the method is based on buoy section ocean current observation data, a deep learning target detection technology is utilized to construct a marine internal wave detection model, accurate detection of internal waves is achieved, detection efficiency is improved, the problems of the traditional internal wave detection method based on remote sensing data are solved through the buoy section ocean current observation data, the method is not influenced by atmospheric environment factors, continuous real-time detection of internal waves is carried out in time, and an unobvious internal wave phenomenon can be detected.
Drawings
FIG. 1 is a schematic general flow diagram of the present invention;
FIG. 2 is a schematic diagram of the apparent internal wave of the present invention;
FIG. 3 is a schematic view of an insignificant internal wave of the present invention;
fig. 4 is a schematic diagram of a network structure of the marine internal wave detection model according to the present invention.
Detailed Description
The following detailed description of the invention refers to the accompanying drawings and preferred embodiments.
As shown in FIG. 1, the invention provides a marine internal wave detection method based on buoy observation data, which comprises the steps of obtaining buoy profile ocean current observation data, drawing a flow velocity profile time sequence diagram, marking an internal wave starting and ending range displayed in the flow velocity profile time sequence diagram, constructing a marine internal wave flow velocity profile image sample library, training a deep learning neural network model by using the marine internal wave flow velocity profile image sample library, detecting the flow velocity profile time sequence diagram to be detected by using the trained deep learning neural network model, and calculating the occurrence and ending time of the detected internal wave. The method comprises the following steps:
s1, acquiring buoy section ocean current observation data, drawing a flow velocity section time sequence diagram, marking the starting and ending range of the internal wave on the image, and constructing an ocean internal wave flow velocity section image sample library. For example, acquiring ocean current observation data (2020.05.10-2020.05.29) of a buoy section in a south China sea area, drawing a total of 8040 flow velocity section time sequence diagrams by using MATLAB, wherein 7083 images of 5 months, 10 days and 5 months, 26 days are used as training data, and 957 images of the remaining 5 months, 27 days and 5 months, 29 days are used as test data for analysis of detection results.
S2, marking all internal wave flow velocity time sequence diagrams by using a LabelImg marking tool according to a target detection Pascal VOC data marking format, and simultaneously marking obvious and unobvious internal waves when marking, wherein the obvious internal waves have an upper wave core and a lower wave core, the cores are red, the upper wave core is dark blue, the lower wave core is dark yellow, layering is obvious, and the color is quite dark as shown in figure 2; the unobvious internal wave also has two wave cores, the upper wave core is light blue, the lower wave core is light yellow, and the color is lighter, as shown in figure 3.
Marking an occurrence interval of the internal wave in the image, generating a marking file, and finally carrying out internal wave time matching according to the starting and ending time of each time sequence diagram and the position of the marked internal wave in the image to obtain a marine internal wave flow velocity profile image sample library.
S3, based on a marine internal wave flow velocity profile image sample library, a marine internal wave detection model is built by using a deep learning target detection fast R-CNN algorithm, a network structure is as shown in figure 4, a backstone network adopts Resnet-50, parameters such as an enhancement mode, a learning rate, iteration times and the like are adjusted to enable model accuracy to be optimal, recall rate Recall and average accuracy mean mAP are used as evaluation indexes, and parameters of the marine internal wave detection model are saved after training ending conditions are met.
After the trained ocean internal wave detection model is obtained, the ocean internal wave detection model is tested by using the residual image of the ocean internal wave flow velocity profile image sample library, the width and height of the area image where the internal wave is located and the starting position of the internal wave are obtained, the starting and ending time of each detected internal wave is calculated after the detection is ended, an operator checks whether the result is accurate or not, and if not, the test image is marked and put into a training set for retraining the model.
S4, when the internal wave is detected, outputting the position of the internal wave, wherein the position is represented by a rectangular frame, and the representation method is B (x i ,y i ,w i ,h i ) Wherein x is i ,y i The upper left corner coordinates, w, of the rectangular frame of the area where the internal wave is located i ,h i For the width and height of the rectangular frame, calculating the occurrence ending time of the internal wave, specifically as follows:
the size of the time sequence diagram of the flow velocity of the internal wave to be detected is W multiplied by H, the time start-stop time of the recorded flow velocity is (M, N), and the position of one internal wave predicted by the marine internal wave detection model is (x) i ,y i ,w i ,h i ) The starting coordinate of the internal wave is x i Ending coordinate is x i +w i ，
Start time m of internal wave i The method comprises the following steps:
end time n of internal wave i The method comprises the following steps:
compared with the traditional internal wave detection method based on the remote sensing image and the artificial naked eye judging method, the marine internal wave deep learning detection model constructed by using the buoy profile ocean current observation data can effectively overcome the defects of the traditional detection method based on the remote sensing image, replaces the artificial naked eye to judge the occurrence of the internal wave, can accurately and continuously detect the internal wave in real time, and can overcome the defects of long time delay, poor environment interference resistance, frequent alarm omission, high labor cost and the like in the traditional method.
While particular embodiments of the present invention have been described above, it will be appreciated by those skilled in the art that these are merely illustrative, and that many changes and modifications may be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims.
Claims (4)
1. A marine internal wave detection method based on buoy observation data is characterized by comprising the following steps of: obtaining buoy profile ocean current observation data, drawing a flow velocity profile time sequence diagram, marking the beginning and ending range of internal waves displayed in the flow velocity profile time sequence diagram, constructing an ocean internal wave flow velocity profile image sample library, training a deep learning neural network model by using the ocean internal wave flow velocity profile image sample library, detecting the flow velocity profile time sequence diagram to be detected by using the trained deep learning neural network model, and calculating the occurrence and ending time of the detected internal waves.
2. The buoy observation data-based ocean internal wave detection method of claim 1, wherein: based on the characteristics of the flow velocity profile timing chart, the start and stop time of the flow velocity is marked as (M, N), the image size is W×H, and the detected internal wave position is marked as (x) i ,y i ,w i ,h i ) Wherein x is i ,y i Indicating the position coordinates, w, of the start of the detected internal wave i ,h i The width and height of a rectangular frame corresponding to the start/end range of the detected internal wave are expressed, and the start time m of the internal wave is calculated by the following equation i And end time n i 。
3. The buoy observation data-based ocean internal wave detection method of claim 2, wherein: marking an internal wave starting and ending range displayed by a section time sequence diagram at a flow speed by using a marking tool labelImg, wherein the marking tool comprises obvious internal waves and unobvious internal waves, the obvious internal waves are provided with an upper wave core and a lower wave core, the cores of the obvious internal waves are red, the upper wave core is dark blue, and the lower wave core is dark yellow; the unobvious internal wave is provided with two wave cores, the upper wave core is light blue, and the lower wave core is light yellow.
4. The buoy observation data-based ocean internal wave detection method of claim 1, wherein: the deep learning neural network model comprises a Faster RCNN model and a YOLOv5 model, and Recall ratio Recall and average precision mean mAP are used as evaluation indexes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111349369.XA CN116147883A (en) | 2021-11-15 | 2021-11-15 | Ocean internal wave detection method based on buoy observation data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111349369.XA CN116147883A (en) | 2021-11-15 | 2021-11-15 | Ocean internal wave detection method based on buoy observation data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116147883A true CN116147883A (en) | 2023-05-23 |
Family
ID=86358631
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111349369.XA Pending CN116147883A (en) | 2021-11-15 | 2021-11-15 | Ocean internal wave detection method based on buoy observation data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116147883A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117168545A (en) * | 2023-10-30 | 2023-12-05 | 自然资源部第一海洋研究所 | Ocean phenomenon observation method and system based on buoy end local identification |
-
2021
- 2021-11-15 CN CN202111349369.XA patent/CN116147883A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117168545A (en) * | 2023-10-30 | 2023-12-05 | 自然资源部第一海洋研究所 | Ocean phenomenon observation method and system based on buoy end local identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021243743A1 (en) | Deep convolutional neural network-based submerged oil sonar detection image recognition method | |
| CN112130132B (en) | Underground pipeline detection method and system based on ground penetrating radar and deep learning | |
| CN113628227B (en) | Coastline change analysis method based on deep learning | |
| Pasqualini et al. | Contribution of side scan sonar to the management of Mediterranean littoral ecosystems | |
| CN111637874B (en) | Multi-AUV layered detection system and detection method for red tide sea area | |
| CN107247927B (en) | Method and system for extracting coastline information of remote sensing image based on tassel cap transformation | |
| CN106443593B (en) | Sweep the adaptive oil spilling information extracting method of enhancing slowly based on coherent radar | |
| CN112215125A (en) | Water level identification method based on YOLOv3 | |
| CN113688948A (en) | Method for identifying attached marine organism species based on YOLO v5 | |
| CN116147883A (en) | Ocean internal wave detection method based on buoy observation data | |
| CN116306203A (en) | Intelligent simulation generation method for marine target track | |
| CN114018317B (en) | Data acquisition device and method for marine environment | |
| CN113469097B (en) | Multi-camera real-time detection method for water surface floaters based on SSD network | |
| CN116562467B (en) | Marine fish target species identification and distribution prediction method and system | |
| CN113393027A (en) | Navigation mark drift intelligent prediction method based on deep learning | |
| CN112633256A (en) | Hydrology test system based on laser radar point cloud technology | |
| CN117372854A (en) | Real-time detection method for hidden danger diseases of deep water structure of dam | |
| CN116912672A (en) | Unmanned survey vessel-based biological integrity evaluation method for large benthonic invertebrates | |
| CN115755068A (en) | Online intelligent diagnosis system for submarine pipeline state | |
| CN113673155B (en) | Water area sand content inversion method based on support vector machine | |
| CN115408938A (en) | Method for inverting propagation speed of solitary wave in ocean by single-scene optical remote sensing image | |
| CN114659503A (en) | Marine ecological environment monitoring method based on artificial intelligence | |
| Lee et al. | Artificial Neural Network for Glider Detection in a Marine Environment by Improving a CNN Vision Encoder | |
| CN115983141B (en) | Deep learning-based method, medium and system for inverting wave height of ocean waves | |
| Suursaar et al. | Hydrodynamically induced or modified patterns derived from satellite images in the coastal waters of Estonia |
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 |