CN114022774A - Radar image-based marine mesoscale vortex monitoring method and device - Google Patents
Radar image-based marine mesoscale vortex monitoring method and device Download PDFInfo
- Publication number
- CN114022774A CN114022774A CN202210019429.XA CN202210019429A CN114022774A CN 114022774 A CN114022774 A CN 114022774A CN 202210019429 A CN202210019429 A CN 202210019429A CN 114022774 A CN114022774 A CN 114022774A
- Authority
- CN
- China
- Prior art keywords
- radar image
- image data
- target
- network model
- detection network
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000012544 monitoring process Methods 0.000 title claims abstract description 36
- 238000001514 detection method Methods 0.000 claims abstract description 97
- 238000012549 training Methods 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 23
- 238000000605 extraction Methods 0.000 claims description 15
- 238000011176 pooling Methods 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 5
- 230000010287 polarization Effects 0.000 claims description 2
- 238000003384 imaging method Methods 0.000 abstract description 3
- 238000004891 communication Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 239000013505 freshwater Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013535 sea water Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a radar image-based method and a radar image-based device for monitoring mesoscale ocean eddy, which relate to the technical field of remote sensing image processing and comprise the following steps: acquiring sample radar image data; determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of the mesoscale vortex in the target radar image data, wherein the target radar image data is radar image data containing the mesoscale vortex; constructing a target detection network model, and training the target detection network model by using a sample data set to obtain a mesoscale vortex detection network model; after the radar image data to be monitored is obtained, the radar image data to be monitored is input into the mesoscale vortex detection network model, the position information of the mesoscale vortex of the radar image data to be monitored is determined, and the technical problem that the existing ocean mesoscale vortex monitoring method is low in monitoring accuracy of the mesoscale vortex position information is solved.
Description
Technical Field
The invention relates to the technical field of remote sensing image processing, in particular to a radar image-based method and device for monitoring mesoscale ocean vortexes.
Background
Mesoscale vortexes, a ubiquitous natural phenomenon in oceans, widely exist in oceans and marginal sea in the world, can wrap and carry a large amount of water bodies in a three-dimensional spiral structure, move at horizontal and vertical movement speeds of meters per second for several weeks or even years, move at a distance of dozens of kilometers to hundreds of kilometers, and carry huge kinetic energy which has important influence on the biogeochemical processes in marine substances, ocean energy, heat, fresh water and seawater. Therefore, the research on the mesoscale vortexes in the ocean has very important scientific significance and application value.
Most research and application results of SAR data on mesoscale vortexes in the prior art are mainly distributed on simple mesoscale vortex classification, so that research on the position of the mesoscale vortexes in accurate monitoring is relatively deficient. Therefore, under the research background of machine vision and deep learning and in a real application environment, how to intelligently and automatically accurately judge the vortex and accurately monitor the vortex position is a problem to be solved currently.
No effective solution has been proposed to the above problems.
Disclosure of Invention
In view of the above, the present invention provides a method and an apparatus for monitoring mesoscale ocean eddy based on radar images, so as to alleviate the technical problem that the existing method for monitoring mesoscale ocean eddy has low accuracy in monitoring mesoscale eddy position information.
In a first aspect, an embodiment of the present invention provides a radar image-based method for monitoring mesoscale ocean eddy, including: obtaining sample radar image data, wherein the sample radar image data comprises: radar image data of different time periods and different sea areas; determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes; constructing a target detection network model, and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model; after the radar image data to be monitored are obtained, inputting the radar image data to be monitored into the mesoscale vortex detection network model, and determining the position information of the mesoscale vortex of the radar image data to be monitored.
Further, based on the target radar image data and the position information of the mesoscale vortexes in the target radar image data, a sample data set is constructed, including: adding a label to the mesoscale vortex in the target radar image data, wherein the label is used for representing the position information of the mesoscale vortex in the target radar image data; performing data enhancement processing on the target radar image data added with the label to obtain the sample data set, wherein the data enhancement processing comprises: horizontal turning treatment, vertical turning treatment and random rotation treatment.
Further, the object detection network model includes: the characteristic extraction layer, regional layer and fine setting pooling layer of selecting, wherein, the characteristic extraction layer comprises residual error network and characteristic pyramid network, the residual error network has 50 layer network structure, includes: the convolution layer, a plurality of residual error layers and full connecting layer, the region selection layer includes: an RPN network, the fine-tuning pooling layer comprising: ROI Align network.
Further, training a target detection network model by using the sample data set to obtain a mesoscale vortex detection network model, including: dividing the sample data set into a training set and a test set according to a preset proportion; training the target detection network model by using the training set to obtain an intermediate target detection network model; and testing the intermediate target detection network model by using the test set, and determining the tested intermediate target detection network model as the mesoscale vortex detection network.
Further, the training set contains a plurality of sample data; training the target detection network model by using the training set to obtain an intermediate target detection network model, comprising: extracting, namely extracting a convolution feature map of the sample data by using the feature extraction layer; a first calculation step of inputting the convolution characteristic graph into the region selection layer to obtain a first loss of the target candidate region; processing, namely performing pooling processing and bilinear difference processing on the convolution feature map and the target candidate region by using the fine-tuning pooling layer to obtain a processed target candidate region; a second calculation step of obtaining a second loss of the processed target candidate region by using the full connection layer and the processed target candidate region; a determining step of determining a target loss of the network model based on the first loss and the second loss; if the target loss is not in the preset range, performing reverse callback on the parameters of the target detection network model based on the target loss, determining the target detection network model with the parameter reverse callback completed as the target detection network model, repeatedly executing the extracting step, the first calculating step, the processing step, the second calculating step and the determining step until the target loss is in the preset range or the repeated execution times reach the preset times, and determining the target detection network model with the target loss in the preset range or the target detection network model obtained when the repeated execution times reach the preset times as the intermediate target detection network model.
Further, inputting the convolution feature map into the region selection layer to obtain the target candidate region and the first loss of the target candidate region, including: inputting the convolution characteristic graph into the region selection layer to generate the target candidate region; adding a binary label to the target candidate region to obtain a type of the target candidate region, wherein the binary label is used for characterizing the type of the candidate region, and the type of the candidate region includes: a foreground candidate region and a background candidate region; and calculating to obtain a first loss of the target candidate region based on the type of the target candidate region and the offset corresponding to the target candidate region.
Further, the sample radar image data is radar image data polarized by VV.
In a second aspect, an embodiment of the present invention provides a device for monitoring mesoscale vortices in the ocean, including: the device comprises an acquisition unit, a construction unit, a training unit and a determination unit, wherein the acquisition unit is used for acquiring sample radar image data, and the sample radar image data comprises: radar image data of different time periods and different sea areas; the construction unit is used for determining target radar image data in the sample radar image data and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes; the training unit is used for constructing a target detection network model and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model; the determining unit is used for inputting the radar image data to be monitored into the mesoscale vortex detection network model after the radar image data to be monitored are obtained, and determining the position information of the mesoscale vortex of the radar image data to be monitored.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, where the memory is used to store a program that supports the processor to execute the method in the first aspect, and the processor is configured to execute the program stored in the memory.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored.
In an embodiment of the present invention, sample radar image data is obtained, where the sample radar image data includes: radar image data of different time periods and different sea areas; determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes; constructing a target detection network model, and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model; after radar image data to be monitored are acquired, the radar image data to be monitored are input into the mesoscale vortex detection network model, and the position information of the mesoscale vortex of the radar image data to be monitored is determined, so that the aim of accurately monitoring the position information of the mesoscale vortex and the mesoscale vortex is fulfilled, the technical problem that the monitoring accuracy of the mesoscale vortex position information is low in the existing marine mesoscale vortex monitoring method is further solved, and the technical effects of intelligently and automatically accurately judging the vortex and accurately monitoring the vortex position are achieved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for monitoring mesoscale ocean eddy based on radar images according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a feature extraction layer provided in an embodiment of the present invention;
FIG. 3 is a diagram of a model test monitoring result provided by an embodiment of the present invention;
FIG. 4 is a schematic diagram of a device for monitoring mesoscale vortices in the ocean according to an embodiment of the present invention;
fig. 5 is a schematic diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The first embodiment is as follows:
in accordance with an embodiment of the present invention, there is provided an embodiment of a method for monitoring mesoscale vortices in the ocean, it being noted that the steps illustrated in the flowchart of the drawings may be performed in a computer system such as a set of computer-executable instructions, and that while a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.
Fig. 1 is a flow chart of a method for monitoring mesoscale vortices in the ocean according to an embodiment of the present invention, as shown in fig. 1, the method including the steps of:
step S102, sample radar image data is obtained, wherein the sample radar image data comprises: radar image data of different time periods and different sea areas;
the sample radar image data is derived from VV polarization images of multiple sea areas at different time periods of an ENVISAT satellite, an ERS-12 satellite, a Sentinel-1 satellite and a GF3 satellite SAR load C-band.
Step S104, determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes;
step S106, constructing a target detection network model, and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model;
step S108, after the radar image data to be monitored are obtained, inputting the radar image data to be monitored into the mesoscale vortex detection network model, and determining the position information of the mesoscale vortex of the radar image data to be monitored.
In an embodiment of the present invention, sample radar image data is obtained, where the sample radar image data includes: radar image data of different time periods and different sea areas; determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes; constructing a target detection network model, and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model; after radar image data to be monitored are acquired, the radar image data to be monitored are input into the mesoscale vortex detection network model, and the position information of the mesoscale vortex of the radar image data to be monitored is determined, so that the aim of accurately monitoring the position information of the mesoscale vortex and the mesoscale vortex is fulfilled, the technical problem that the monitoring accuracy of the mesoscale vortex position information is low in the existing marine mesoscale vortex monitoring method is further solved, and the technical effects of intelligently and automatically accurately judging the vortex and accurately monitoring the vortex position are achieved.
In the embodiment of the present invention, step S102 includes the following steps:
step S11, adding a label to the mesoscale vortex in the target radar image data, wherein the label is used for representing the position information of the mesoscale vortex in the target radar image data;
step S12, performing data enhancement processing on the target radar image data to which the tag is added, to obtain the sample data set, where the data enhancement processing includes: horizontal turning treatment, vertical turning treatment and random rotation treatment.
In the embodiment of the invention, the labels are added to the mesoscale vortexes in the target radar image data, and can be manually marked by experts.
The target radar image data are small in quantity and not enough for deep learning training, so that data enhancement is needed, horizontal turning, vertical turning and random rotation operations are carried out by adopting image binary affine invariance, and the same operation is carried out on labeled labels, so that a sample data set is obtained.
In an embodiment of the present invention, the target detection network model includes: the device comprises a feature extraction layer, a region selection layer and a fine tuning pooling layer, wherein the feature extraction layer is composed of a residual error network and a feature pyramid network, and the residual error network comprises: the convolution layer, a plurality of residual error layers and full connecting layer, the region selection layer includes: an RPN network, the fine-tuning pooling layer comprising: ROI Align network.
In the embodiment of the present invention, step S106 includes the following steps:
step S21, dividing the sample data set into a training set and a test set according to a preset proportion;
step S22, training the target detection network model by using the training set to obtain an intermediate target detection network model;
and step S23, testing the intermediate target detection network model by using the test set, and determining the tested intermediate target detection network model as the mesoscale vortex detection network.
Specifically, the training set includes a plurality of sample data, and step S22 further includes the following steps:
extracting, namely extracting a convolution feature map of the sample data by using the feature extraction layer;
a first calculation step of inputting the convolution characteristic graph into the region selection layer to obtain a first loss of the target candidate region;
processing, namely performing pooling processing and bilinear difference processing on the convolution feature map and the target candidate region by using the fine-tuning pooling layer to obtain a processed target candidate region;
a second calculation step of obtaining a second loss of the processed target candidate region by using the full connection layer and the processed target candidate region;
a determining step of determining a target loss of the detection model based on the first loss and the second loss;
if the target loss is not in the preset range, performing reverse callback on the parameters of the target detection network model based on the target loss, determining the target detection network model with the parameter reverse callback completed as the target detection network model, repeatedly executing the extracting step, the first calculating step, the processing step, the second calculating step and the determining step until the target loss is in the preset range or the repeated execution times reach the preset times, and determining the target detection network model with the target loss in the preset range or the target detection network model obtained when the repeated execution times reach the preset times as the intermediate target detection network model.
In the embodiment of the invention, the target detection network model comprises a feature extraction layer (ResNet 50+ FPN), a region selection layer (RPN) and a fine tuning pooling layer (ROI Align). The characteristic extraction layer mainly extracts target characteristics in the SAR image by using a residual error network (ResNet 50) and a characteristic pyramid network (FPN), wherein ResNet50 is composed of a 50-layer network structure, the network structure comprises a convolutional layer, a plurality of residual error layers and a full-connection layer, the residual error layers comprise residual error blocks and 1 × 1 residual error blocks, and one residual error block can be expressed as:
As shown in fig. 2, FPN mainly functions to provide multi-scale detection for feature extraction, and performs "top-down" upsampling on 5-layer residual results output "bottom-up" by ResNet50, thereby implementing feature fusion between a lower layer and a higher layer, and achieving the purposes of rich feature semantic information and accurate target position. ResNet50 and FPN thus form a "bottom-up" and "top-down" structure for extracting features of radar image data of arbitrary size.
Inputting sample data into a feature extraction layer to obtain a 4 m n convolution feature map, wherein the convolution feature map comprises 4 layers of fusion features C1, C2, C3 and C4, m and n of the 4 layers of C1, C2, C3 and C4 are different in size, and m and n represent sizes of the convolution feature map.
Next, the RPN network generates candidate regions (anchors) using 4 × m × n convolved feature maps, and the anchors uses 32 × 32, 64 × 64, 128 × 128, 256 × 256, 512 × 512 pixels and a 1:1, 1:2, 2:1 ratio to obtain 15 candidate regions, indicating that the convolved feature image considers 4 × m × n 15 possible candidate windows. Each anchor is then assigned a binary label (1: foreground, 0: background), with two types of positive label foreground:
1) the highest intersection-to-union ratio with the mesoscale vortex group route (GT) label box ("ratio of intersection and union of predicted bounding box" and "true bounding box": IoU) overlapping anchors (less than 0.7);
2) anchor with IoU overlap of greater than 0.7 with any GT label box, negative label background: the IoU ratio for all GT boxes is below 0.3 anchor;
from this, 15 × 2 dimensional scores (scores) can be obtained for classification (cls = 4 × m × n × 15 × 2). Each anchor has 4 offsets for x, y, w, h for the positive label position candidate box (reg = 4 m n 15 4). This classification and regression process yields the RPN penalty as the first penalty, which is calculated as follows:
wherein the content of the first and second substances,
representing the probability that the ith anchor is predicted to be a true tag,
1 when a positive sample, 0 when a negative sample,
representing the bounding box regression parameters that predict the ith anchor,
denotes the GT Box corresponding to the ith anchor,
representing the total number of samples in cls counted at one time,
indicating the number of regs calculated at one time.
Then, the ROI Align extracts the candidate frame mapping from the convolution feature map by using pooling and bilinear interpolation method to generate a feature map with fixed size, the pooling is used to scale the feature by fixed size, the bilinear interpolation is used to fine-tune the edge with fixed size, and then the full connection layer is passed through to obtain the second loss, the calculation formula of the second loss is as follows:
pis the softmax probability predicted by the classifier,ucorresponding to the real label of the target, only one target,
corresponding classes predicted by corresponding bounding box regressorsuThe regression parameters of (a) are determined,vand (4) bounding box regression parameters corresponding to the real target.
Finally, the first loss is addedL 1And the second lossL 2The two losses are combined to train the target detection network model. Parameters are adjusted for many times during the training process, suitable parameters are searched (operations =0.0025, Batchsize =2, Epoch = 10000), and Epoch is stored every 4000 times. And when the last Epoch training is completed, obtaining the mesoscale vortex detection network.
After the mesoscale vortex detection network is obtained, performing mesoscale vortex detection verification by using a test set and a mesoscale vortex detection network model, wherein the test standard is as follows:
TP is IoU > the number of detection frames (same GT is calculated once) of the set threshold (threshold = 0.5), FP is IoU ≦ the number of detection frames of the set threshold (threshold = 0.5), or the number of redundant detection frames of the same GT is detected, FN is the number of GT not detected. Test results as shown in fig. 3, the model can accurately monitor the mesoscale vortex locations (grey border).
And finally, inputting real-time radar image data to be monitored, which is sent by the satellite, into the target detection network model for completing the test, and monitoring the marine mesoscale vortex in real time, so as to identify the position information of the mesoscale vortex and the mesoscale vortex.
This application can not restrict radar image data size through constituting the feature extraction layer by residual error network and feature pyramid network, under the circumstances that the machine memory allows, can be with the radar image data of arbitrary size, input mesoscale vortex detection network model in, carry out real-time supervision marine mesoscale vortex phenomenon to radar image data intelligent automatic monitoring marine mesoscale vortex problem has been solved.
Example two:
the embodiment of the invention also provides a device for monitoring the mesoscale ocean vortexes, which is used for executing the method for monitoring the mesoscale ocean vortexes provided by the embodiment of the invention.
As shown in fig. 4, fig. 4 is a schematic view of the device for monitoring mesoscale ocean vortices, which includes: an acquisition unit 10, a construction unit 20, a training unit 30 and a determination unit 40.
The acquiring unit 10 is configured to acquire sample radar image data, where the sample radar image data includes: radar image data of different time periods and different sea areas;
the constructing unit 20 is configured to determine target radar image data in the sample radar image data, and construct a sample data set based on the target radar image data and position information of a mesoscale vortex in the target radar image data, where the target radar image data is radar image data including the mesoscale vortex;
the training unit 30 is configured to construct a target detection network model, and train the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model;
the determining unit 40 is configured to, after the radar image data to be monitored is acquired, input the radar image data to be monitored into the mesoscale vortex detection network model, and determine position information of the mesoscale vortex of the radar image data to be monitored.
In an embodiment of the present invention, sample radar image data is obtained, where the sample radar image data includes: radar image data of different time periods and different sea areas; determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes; training a target detection network model by using the sample data set to obtain a mesoscale vortex detection network model; after radar image data to be monitored are acquired, the radar image data to be monitored are input into the mesoscale vortex detection network model, and the position information of the mesoscale vortex of the radar image data to be monitored is determined, so that the aim of accurately monitoring the position information of the mesoscale vortex and the mesoscale vortex is fulfilled, the technical problem that the monitoring accuracy of the mesoscale vortex position information is low in the existing marine mesoscale vortex monitoring method is further solved, and the technical effects of intelligently and automatically accurately judging the vortex and accurately monitoring the vortex position are achieved.
Example three:
an embodiment of the present invention further provides an electronic device, including a memory and a processor, where the memory is used to store a program that supports the processor to execute the method described in the first embodiment, and the processor is configured to execute the program stored in the memory.
Referring to fig. 5, an embodiment of the present invention further provides an electronic device 100, including: the device comprises a processor 50, a memory 51, a bus 52 and a communication interface 53, wherein the processor 50, the communication interface 53 and the memory 51 are connected through the bus 52; the processor 50 is arranged to execute executable modules, such as computer programs, stored in the memory 51.
The Memory 51 may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 53 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used.
The bus 52 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 5, but this does not indicate only one bus or one type of bus.
The memory 51 is used for storing a program, the processor 50 executes the program after receiving an execution instruction, and the method executed by the apparatus defined by the flow process disclosed in any of the foregoing embodiments of the present invention may be applied to the processor 50, or implemented by the processor 50.
The processor 50 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 50. The Processor 50 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 51, and the processor 50 reads the information in the memory 51 and completes the steps of the method in combination with the hardware thereof.
Example four:
the embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program performs the steps of the method in the first embodiment.
In addition, in the description of the embodiments of the present invention, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A radar image-based marine mesoscale vortex monitoring method is characterized by comprising the following steps:
obtaining sample radar image data, wherein the sample radar image data comprises: radar image data of different time periods and different sea areas;
determining target radar image data in the sample radar image data, and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes;
constructing a target detection network model, and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model;
after the radar image data to be monitored are obtained, inputting the radar image data to be monitored into the mesoscale vortex detection network model, and determining the position information of the mesoscale vortex of the radar image data to be monitored.
2. The method of claim 1, wherein constructing a sample data set based on the target radar image data and location information of mesoscale vortices in the target radar image data comprises:
adding a label to the mesoscale vortex in the target radar image data, wherein the label is used for representing the position information of the mesoscale vortex in the target radar image data;
performing data enhancement processing on the target radar image data added with the label to obtain the sample data set, wherein the data enhancement processing comprises: horizontal turning treatment, vertical turning treatment and random rotation treatment.
3. The method of claim 1, wherein the target detection network model comprises: the characteristic extraction layer, regional layer and fine setting pooling layer of selecting, wherein, the characteristic extraction layer comprises residual error network and characteristic pyramid network, the residual error network has 50 layer network structure, includes: the convolution layer, a plurality of residual error layers and full connecting layer, the region selection layer includes: an RPN network, the fine-tuning pooling layer comprising: ROI Align network.
4. The method of claim 3, wherein training a target detection network model using the sample data set to obtain a mesoscale vortex detection network model comprises:
dividing the sample data set into a training set and a test set according to a preset proportion;
training the target detection network model by using the training set to obtain an intermediate target detection network model;
and testing the intermediate target detection network model by using the test set, and determining the tested intermediate target detection network model as the mesoscale vortex detection network.
5. The method of claim 4, wherein the training set contains a plurality of sample data;
training the target detection network model by using the training set to obtain an intermediate target detection network model, comprising:
extracting, namely extracting a convolution feature map of the sample data by using the feature extraction layer;
a first calculation step of inputting the convolution characteristic graph into the region selection layer to obtain a target candidate region and a first loss of the target candidate region;
processing, namely performing pooling processing and bilinear difference processing on the convolution feature map and the target candidate region by using the fine-tuning pooling layer to obtain a processed target candidate region;
a second calculation step of obtaining a second loss of the processed target candidate region by using the full connection layer and the processed target candidate region;
a determining step of determining a target loss of the target detection network model based on the first loss and the second loss;
if the target loss is not in the preset range, performing reverse callback on the parameters of the target detection network model based on the target loss, determining the target detection network model with the parameter reverse callback completed as the target detection network model, repeatedly executing the extracting step, the first calculating step, the processing step, the second calculating step and the determining step until the target loss is in the preset range or the repeated execution times reach the preset times, and determining the target detection network model with the target loss in the preset range or the target detection network model obtained when the repeated execution times reach the preset times as the intermediate target detection network model.
6. The method of claim 5, wherein inputting the convolution signature into the region selection layer to obtain the target candidate region and the first loss of the target candidate region comprises:
inputting the convolution characteristic graph into the region selection layer to generate the target candidate region;
adding a binary label to the target candidate region, wherein the binary label is used for characterizing the type of the target candidate region, and the type of the target candidate region includes: a foreground candidate region and a background candidate region;
calculating a first loss of the target candidate region based on the type of the target candidate region and the offset corresponding to the target candidate region.
7. The method of claim 1,
the sample radar image data is radar image data subjected to VV polarization.
8. A device for monitoring mesoscale vortices in the ocean, comprising: an acquisition unit, a construction unit, a training unit and a determination unit, wherein,
the acquiring unit is configured to acquire sample radar image data, where the sample radar image data includes: radar image data of different time periods and different sea areas;
the construction unit is used for determining target radar image data in the sample radar image data and constructing a sample data set based on the target radar image data and position information of mesoscale vortexes in the target radar image data, wherein the target radar image data is radar image data containing mesoscale vortexes;
the training unit is used for constructing a target detection network model and training the target detection network model by using the sample data set to obtain a mesoscale vortex detection network model;
the determining unit is used for inputting the radar image data to be monitored into the mesoscale vortex detection network model after the radar image data to be monitored are obtained, and determining the position information of the mesoscale vortex of the radar image data to be monitored.
9. An electronic device comprising a memory for storing a program that enables a processor to perform the method of any of claims 1 to 7 and a processor configured to execute the program stored in the memory.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of the claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210019429.XA CN114022774A (en) | 2022-01-10 | 2022-01-10 | Radar image-based marine mesoscale vortex monitoring method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210019429.XA CN114022774A (en) | 2022-01-10 | 2022-01-10 | Radar image-based marine mesoscale vortex monitoring method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114022774A true CN114022774A (en) | 2022-02-08 |
Family
ID=80069749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210019429.XA Pending CN114022774A (en) | 2022-01-10 | 2022-01-10 | Radar image-based marine mesoscale vortex monitoring method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114022774A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114494894A (en) * | 2022-04-18 | 2022-05-13 | 中国海洋大学 | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment |
| CN115953394A (en) * | 2023-03-10 | 2023-04-11 | 中国石油大学(华东) | Target segmentation-based detection method and system for mesoscale ocean vortexes |
| CN116719031A (en) * | 2023-08-08 | 2023-09-08 | 中国科学院空天信息创新研究院 | Ocean vortex detection method and system for synthetic aperture radar SAR image |
| CN117011575A (en) * | 2022-10-27 | 2023-11-07 | 腾讯科技（深圳）有限公司 | Training method and related device for small sample target detection model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163287A (en) * | 2019-05-24 | 2019-08-23 | 三亚中科遥感研究所 | A kind of mesoscale eddy detection method and device |
| CN112766108A (en) * | 2021-01-08 | 2021-05-07 | 西安电子科技大学 | SAR image target detection method based on context information |
| CN112766194A (en) * | 2021-01-26 | 2021-05-07 | 上海海洋大学 | Detection method for mesoscale ocean eddy |
| CN113592786A (en) * | 2021-07-12 | 2021-11-02 | 浙江大学 | Deep learning-based detection method for mesoscale vortices in ocean |
-
2022
- 2022-01-10 CN CN202210019429.XA patent/CN114022774A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163287A (en) * | 2019-05-24 | 2019-08-23 | 三亚中科遥感研究所 | A kind of mesoscale eddy detection method and device |
| CN112766108A (en) * | 2021-01-08 | 2021-05-07 | 西安电子科技大学 | SAR image target detection method based on context information |
| CN112766194A (en) * | 2021-01-26 | 2021-05-07 | 上海海洋大学 | Detection method for mesoscale ocean eddy |
| CN113592786A (en) * | 2021-07-12 | 2021-11-02 | 浙江大学 | Deep learning-based detection method for mesoscale vortices in ocean |
Non-Patent Citations (2)
| Title |
|---|
| DI ZHANG 等: "SAR Eddy Detection Using Mask-RCNN and Edge Enhancement", 《IGARSS 2020-2020 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM》 * |
| 谢鹏飞 等: "基于深度学习的中尺度涡检测技术及其在声场中的应用", 《海洋信息》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114494894A (en) * | 2022-04-18 | 2022-05-13 | 中国海洋大学 | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment |
| CN117011575A (en) * | 2022-10-27 | 2023-11-07 | 腾讯科技（深圳）有限公司 | Training method and related device for small sample target detection model |
| CN115953394A (en) * | 2023-03-10 | 2023-04-11 | 中国石油大学(华东) | Target segmentation-based detection method and system for mesoscale ocean vortexes |
| CN116719031A (en) * | 2023-08-08 | 2023-09-08 | 中国科学院空天信息创新研究院 | Ocean vortex detection method and system for synthetic aperture radar SAR image |
| CN116719031B (en) * | 2023-08-08 | 2023-10-13 | 中国科学院空天信息创新研究院 | Ocean vortex detection method and system for synthetic aperture radar SAR image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114022774A (en) | Radar image-based marine mesoscale vortex monitoring method and device | |
| CN111523459B (en) | Remote sensing image bare area identification method and device, electronic equipment and storage medium | |
| CN113139543B (en) | Training method of target object detection model, target object detection method and equipment | |
| CN112329881B (en) | License plate recognition model training method, license plate recognition method and device | |
| CN112001374B (en) | Cloud detection method and device for hyperspectral image | |
| CN114494894B (en) | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment | |
| CN114972191A (en) | Method and device for detecting farmland change | |
| CN114140683A (en) | Aerial image target detection method, equipment and medium | |
| CN113487610B (en) | Herpes image recognition method and device, computer equipment and storage medium | |
| CN112669323A (en) | Image processing method and related equipment | |
| CN112132131A (en) | Measuring cylinder liquid level identification method and device | |
| CN114037907A (en) | Detection method and device for power transmission line, computer equipment and storage medium | |
| CN112052907A (en) | Target detection method and device based on image edge information and storage medium | |
| CN114565824A (en) | Single-stage rotating ship detection method based on full convolution network | |
| CN113095316B (en) | Image rotation target detection method based on multilevel fusion and angular point offset | |
| CN114549970A (en) | Night small target fruit detection method and system fusing global fine-grained information | |
| CN111931721B (en) | Method and device for detecting color and number of annual inspection label and electronic equipment | |
| CN116778176B (en) | SAR image ship trail detection method based on frequency domain attention | |
| CN117523573A (en) | Water meter reading method and system based on multi-stage visual detection fusion fault early warning | |
| CN116363583A (en) | Human body identification method, device, equipment and medium for top view angle | |
| CN111582012A (en) | Method and device for detecting small target ship | |
| CN116704366A (en) | Storm drop zone identification early warning method and device based on transform deep learning model | |
| CN112149536B (en) | Squall line wind speed prediction method | |
| CN114488070B (en) | Radar echo extrapolation method and device based on deep learning model | |
| CN113011415A (en) | Improved target detection method and system based on Grid R-CNN model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220208 |
|
| RJ01 | Rejection of invention patent application after publication |