CN114494894B - Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment - Google Patents
Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment Download PDFInfo
- Publication number
- CN114494894B CN114494894B CN202210401146.1A CN202210401146A CN114494894B CN 114494894 B CN114494894 B CN 114494894B CN 202210401146 A CN202210401146 A CN 202210401146A CN 114494894 B CN114494894 B CN 114494894B
- Authority
- CN
- China
- Prior art keywords
- vortex
- ocean
- black
- sar image
- automatic identification
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention provides a method and a device for automatically identifying ocean black vortexes and inverting key parameters and electronic equipment, and relates to the technical field of ocean engineering. The method comprises the following steps: acquiring SAR image target data; constructing an SAR image ocean black vortex sample library; establishing a sub-mesoscale ocean black vortex automatic identification model; training the target recognition model by utilizing the SAR image ocean black vortex sample library; carrying out sub-mesoscale ocean black vortex automatic identification by using the trained network model to finally obtain a target identification result; and carrying out vortex center position and vortex edge position key parameter inversion on the identified black vortex. The method realizes the automatic and accurate identification and key parameter inversion of the sub-mesoscale ocean black vortex of the SAR image without manual interference, and has reference value for the underwater research of the mesoscale ocean vortex.
Description
Technical Field
The invention relates to the technical field of ocean engineering, in particular to an ocean black vortex automatic identification and key parameter inversion method, an ocean black vortex automatic identification and key parameter inversion device and electronic equipment.
Background
Ocean vortex is a common ocean phenomenon, has frequent offshore activities in China, plays an important role in exchanging ocean substances and energy, and is an important object of ocean science research. Synthetic Aperture Radar (SAR) has the characteristics of all-time, all-weather, high resolution and large coverage range, and provides a large amount of image data for ocean vortex research. The characteristic of high resolution of the SAR system can observe small and medium-scale ocean vortexes and detailed information of the vortexes, and particularly for the sub-medium-scale vortexes, the small and medium-scale ocean vortexes are difficult to observe through traditional sea surface height data or sea surface temperature data. The vortices that appear due to the presence of the oil film on the surface of the sea are called black vortices. The sea surface oil film can smooth the capillary wave on the sea surface, reduce the backscattering sectional area, is used as a tracer, moves along with ocean vortex, and presents spiral dark curve stripes on the SAR image. Most of the ocean vortices observed in SAR images are "black vortices", particularly in offshore areas. In the past work, the SAR image vortex identification and key parameter extraction mainly adopt a manual visual method, certain subjective judgment difference exists, and the time and labor are wasted by using manual identification from massive data along with the accumulation of vortex SAR images, so that an SAR image ocean black vortex automatic identification and key parameter inversion method is very needed.
Disclosure of Invention
In view of this, the invention aims to provide a method, a device and an electronic device for automatically identifying ocean black vortexes and inverting key parameters, so as to improve the efficiency of automatically identifying SAR ocean vortexes, thereby expanding the application of deep learning in the ocean field.
An ocean black vortex automatic identification and key parameter inversion method comprises the following steps:
s101, acquiring SAR image target data;
s102, constructing an SAR image ocean black vortex sample library;
s103, establishing an ocean black vortex automatic identification model;
s104, inputting the SAR image ocean black vortex sample library in the step S102 into the automatic recognition model in the step S103 for training;
s105, automatically identifying the ocean black vortex by using the trained automatic identification model to finally obtain a target prediction result;
and S106, performing vortex center position and vortex edge position key parameter inversion on the target prediction result data.
Further, in step S101, the SAR image target data includes: firstly, collecting VV polarized ocean vortex images of a plurality of sea areas in a plurality of different time periods of SAR loads C-band of an ENVISAT satellite, an ERS-1/2 satellite, a Sentinel-1 satellite and a GF3 satellite; and then, through a manual visual method, in a global vortex high incidence sea area, SAR image target data is screened and collected.
Further, the step S102 specifically includes:
(1) firstly, decoding SAR image target data in the step S101, performing image stretching and contrast ratio limiting adaptive histogram equalization, and outputting to obtain a picture library A;
the SAR image stretching parameters are set as follows:
if it is not
，
Wherein the content of the first and second substances,
represents the 0.75 quantile of the product,
it represents a 0.25 quantile of the number,
which represents the minimum value of the sum of the values,
represents the median value;
if it is not
，
Wherein, in the process,
the standard deviation is expressed in terms of the standard deviation,
which represents the minimum value of the sum of the values,
representing a mean value;
the method for limiting contrast self-adaptive histogram equalization specifically comprises the following steps:
1) dividing the SAR image into 8 multiplied by 8 blocks, calculating a histogram by taking the blocks as units, setting a cutting amplitude limit to 10, and carrying out equalization;
2) the pixel interpolation between blocks is carried out by the following method:
known functionfIn (1)
，
Calculating the value of four points, calculating the unknown functionfAt the point of
A value of (d); interpolation is performed in the X direction to obtain:
interpolation is performed in the y direction to obtain:
interpolated result
The following were used:
(2) secondly, expanding the picture library A to generate a picture library A1 by adopting data expansion methods such as random horizontal turnover, rotation transformation, translation transformation and the like; the picture library A and the picture library A1 jointly form a data set B;
(3) and finally, identifying each picture in the data set B by a manual visual method, manually marking the ocean vortex by adopting an external rectangle, marking whether the marked content contains the ocean black vortex, correctly organizing the format of each data set according to the VOC format, preparing a configuration file, and finally obtaining the SAR image ocean black vortex sample library.
Further, the step S103 specifically includes: based on a convolutional neural network;
(1) establishing a main feature extraction network, performing cross-pixel extraction on each input picture to obtain 4 independent feature layers, and stacking the feature layers, wherein the number of channels is expanded to 4 times of the original number;
(2) performing channel adjustment by using convolution, standardization and a SilU activation function of 3 multiplied by 3;
the SilU has the characteristics of no upper bound, low bound, smoothness and nonmonotone, and the formula is as follows:
(3) performing high and wide compression, expanding the number of channels and performing feature extraction by using a convolution, standardization and SiLU activation function with a convolution kernel of 3 multiplied by 3 and a step length of 2 multiplied by 2;
(4) repeating the step (3) for 2 times to extract the features, and respectively obtaining effective feature layers P1 and P2;
(5) performing maximum pooling feature extraction by utilizing convolution with convolution kernel of 3 × 3 and step length of 2 × 2, pooling by adopting a spatial pyramid, performing maximum pooling feature extraction by utilizing pooling kernels of 5, 9 and 13 respectively, stacking the pooled results, and adjusting the number of channels by utilizing CSPLAyer to obtain an effective feature layer P3;
(6) respectively inputting 3 effective feature layers P1, P2 and P3 into the FPN feature pyramid reinforced feature extraction network for feature fusion; utilizing a deeper characteristic layer to perform up-sampling and fusing a shallow characteristic layer; downsampling the result of the shallow feature fusion, and fusing the result with the deep feature layers to obtain 3 enhanced feature layers PQ1, PQ2 and PQ 3;
(7) judging whether an object corresponds to the feature point or not and judging the category of each feature point by enhancing three feature maps of PQ1, PQ2 and PQ3 output by the feature extraction network.
Further, step S104 is specifically;
randomly dividing the sample into a training set P1, a verification set P2 and a test set P3 according to 80%, 10% and 10% of a sample library; the training set P1 is used for training the network model, and the training set P1 is sent to the network model established in the step S103 to be trained to obtain a plurality of weight models; the verification set P2 is used for adjusting each parameter in the model, when a plurality of weight models are trained in the training set P1, different weight models are respectively used for predicting the P2, and the parameter corresponding to the weight model with the highest score after prediction is selected, so that the optimal weight model is generated.
Further, in the step S105, after the optimal weight model is obtained through the training set P1 and the verification set P2, the test set is used for testing, and the obtained optimal weight model is used for identifying the ocean black vortexes in the SAR picture.
Further, the step S106 specifically includes:
(1) and (3) inverting vortex center position parameters: scanning the sub-image line by line and column by column, and judging the number of intersection points of a scanning line and a medium-speed spiral line in the image; the closer to the center of the constant-speed spiral line, the more the number of the intersection points is; counting the number of intersection points in the horizontal direction and the vertical direction respectively, obtaining an interval with larger number of the intersection points, namely the general range of the vortex center point, by setting a certain error range, then obtaining an optimal center point position for each vortex edge by a traversal method, comparing the obtained optimal vortex center position with the vortex position manually and visually interpreted, and calculating the absolute deviation of the two positions for evaluating the inversion error of the vortex center position;
(2) inverting vortex edge position parameters:
1) calculating all connected regions in the image to obtain a coordinate index of each connected region;
2) calculating the length of each communication area, and screening out the longest 2 main communication area arc sections;
3) performing curve fitting on the longest 2 main communication areas, performing conversion from a Cartesian coordinate system to a polar coordinate system on points in each arc section, and calculating a spiral line parameter which is optimal to approximate to a vortex edge by combining a least square judgment error;
with vortex edgeLThe polar equation of an archimedes' spiral is:
wherein, the first and the second end of the pipe are connected with each other,
and
are all real numbers, and are all real numbers,
the diameter of the electrode is the diameter of the electrode,
is a polar angle;
the archimedes spiral coordinate, the transformation equation from cartesian to polar coordinate system is as follows:
wherein, the first and the second end of the pipe are connected with each other,
and
representing the coordinates of points on the vortex edge L,
and
representing the coordinates of the location of the vortex core,
the diameter of the electrode is the diameter of the electrode,
is a polar angle; for any point on the edge L
Corresponding to
And distance
The following were used:
substituting the formulas (4) and (5) into the formula (1), and calculating by least square method to obtain constant velocity spiral parameters optimally approximating to the vortex edge
And
；
4) measuring the fitting effect of the fitting curve using the maximum absolute error
Maximum absolute error, i.e. the distance the worst case edge point is measured from the curve:
wherein, the first and the second end of the pipe are connected with each other,
the distance of the edge points from the fitted curve.
An ocean black vortex automatic identification and key parameter inversion device comprises:
the SAR image acquisition module is used for acquiring target data of an SAR target image to be detected, wherein the target data comprise VV polarized ocean vortex images of an ENVISAT satellite, an ERS-1/2 satellite, a Sentinel-1 satellite and a GF3 satellite SAR load C-band, a plurality of different time periods and a plurality of sea areas
The pre-training module of the ocean black vortex automatic identification model is used for training to obtain an SAR image ocean black vortex automatic identification optimal weight model;
the automatic identification module of ocean black vortexes is used for realizing the automatic identification of the sub-mesoscale ocean black vortexes in the SAR image and determining the sub-mesoscale ocean black vortexes in the SAR image;
and the ocean black vortex key parameter inversion module is used for inverting vortex center position and vortex edge position parameters.
An electronic device comprising a memory for storing a program that enables a processor to perform the above method and a processor configured to execute the program stored in the memory.
A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program performs the steps of the above-mentioned method when being executed by a processor.
The invention has the advantages and beneficial effects that:
the SAR image target data is obtained; constructing an SAR image ocean black vortex sample library by using the target data; establishing a sub-mesoscale ocean black vortex automatic identification model; training an automatic identification model by utilizing the SAR image ocean black vortex sample library; carrying out sub-mesoscale ocean black vortex automatic identification by using the trained network model to finally obtain a target prediction result; and performing vortex center position and vortex edge position key parameter inversion on the target prediction result data. The invention solves the problems of time and labor consumption in manual visual identification and interpretation, realizes automatic accurate and efficient identification of the sub-mesoscale ocean black vortex of the SAR image, and provides the key parameter inversion method with high accuracy.
Drawings
Fig. 1 is a flowchart of automatic identification and key parameter inversion of ocean black vortexes according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of an apparatus for automatically identifying ocean black vortexes and inverting key parameters according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of an electronic device according to an embodiment of the present invention.
Fig. 4 is a result intention of automatic marine black vortex identification provided in the embodiment of the present invention.
Fig. 5 is a schematic diagram of a vortex-center inversion result provided in an embodiment of the present invention.
FIG. 6 is a schematic illustration of the location of the vortex edge provided by 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. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
The first embodiment is as follows:
the embodiment provides an automatic ocean black vortex identification and key parameter inversion method, as shown in fig. 1, the automatic sub-mesoscale ocean black vortex identification method comprises the following steps:
step S101, SAR image target data is obtained
The target data comprise VV polarized ocean vortex images of a plurality of sea areas of a plurality of different time periods of ENVISAT satellites, ERS-1/2 satellites, Sentinel-1 satellites and GF3 satellite SAR loads C-band;
step S102, an SAR image ocean black vortex sample library is constructed
(1) Firstly, decoding SAR image target data in the step S101, performing image stretching and contrast ratio limiting adaptive histogram equalization, and outputting to obtain a picture library A;
the SAR image stretching parameters are set as follows:
if it is not
，
Wherein the content of the first and second substances,
represents the 0.75 quantile of the product,
it represents a 0.25 quantile of the molecule,
which represents the minimum value of the sum of the values,
represents the median value;
if it is used
，
Wherein, in the process,
the standard deviation is expressed in terms of the standard deviation,
the minimum value is indicated and is,
represents the mean value;
the method for limiting contrast self-adaptive histogram equalization specifically comprises the following steps:
1) dividing the SAR image into 8 multiplied by 8 blocks, calculating a histogram by taking the blocks as units, setting a cutting amplitude limit to 10, and carrying out equalization;
2) the pixel interpolation between blocks is carried out by the following method:
known functionfIn
，
Calculating the value of four points, calculating the unknown function fAt the point of
A value of (d);
interpolation is performed in the X direction to obtain:
interpolation in the y direction yields:
interpolated result
The following:
(2) secondly, expanding the picture library A to generate a picture library A1 by adopting data expansion methods such as random horizontal turnover, rotation transformation, translation transformation and the like; the picture library A and the picture library A1 jointly form a data set B;
(3) finally, identifying each picture in the data set B by a manual visual method, manually marking the ocean vortex by adopting an external rectangle, marking whether the marked content contains the ocean black vortex, correctly organizing the format of each data set according to the VOC format, and preparing a configuration file to finally obtain an SAR image ocean black vortex sample library;
step S103, establishing an ocean black vortex automatic identification model, wherein the embodiment aims at the sub-mesoscale ocean black vortex:
1) firstly, establishing a main feature extraction network, firstly, carrying out cross-pixel point extraction on each input picture to obtain 4 independent feature layers, and then stacking the feature layers, wherein the number of channels is expanded to 4 times of the original number;
2) secondly, channel adjustment is carried out by utilizing convolution, standardization and a SilU activation function of 3 multiplied by 3;
the SilU has the characteristics of no upper bound and lower bound, smoothness and nonmonotony, and the formula is as follows:
where x represents an image pixel value.
3) Secondly, performing high and wide compression and channel number expansion by using convolution, standardization and a SilU activation function with a convolution kernel of 3 multiplied by 3 and a step length of 2 multiplied by 2, and performing feature extraction by using CSPLAyer;
4) then, repeating the step 3) for 2 times to perform feature extraction, and respectively obtaining effective feature layers P1 and P2;
5) then, convolution is carried out by using a convolution kernel of 3 × 3 and the step length of 2 × 2, spatial pyramid pooling is adopted, the pooling kernels of 5, 9 and 13 are respectively used for carrying out maximum pooling for feature extraction, the pooled results are stacked, and then the number of channels is adjusted by using a CSPLAyer to obtain an effective feature layer P3.
6) Then, the 3 effective feature layers P1, P2 and P3 are respectively input into the FPN feature pyramid enhanced feature extraction network for feature fusion. Utilizing a deeper characteristic layer to perform up-sampling and fusing a shallow characteristic layer; performing downsampling on the result of the shallow feature fusion, and fusing the result with the deep feature layer to obtain 3 enhanced feature layers PQ1, PQ2 and PQ 3;
7) finally, whether an object corresponds to the feature point or not is judged by enhancing three feature maps of PQ1, PQ2 and PQ3 output by the feature extraction network, and the category of each feature point is judged. Inhibiting and eliminating repeated boundary frames by using a non-maximum value to obtain a more accurate detection result;
step S104, inputting the SAR image ocean black vortex sample library in the step S102 into the automatic identification model in the step S103 for training;
step S105, performing sub-mesoscale ocean black vortex automatic identification by using the trained network model to finally obtain a target prediction result, as shown in FIG. 4;
and S106, performing vortex center position and vortex edge position key parameter inversion on the target prediction result data.
(1) And (3) inverting parameters of the vortex center position: and (3) scanning the sub-image row by row and column by column to judge the number of the intersection points of the scanning line and the medium-speed spiral line in the image. The closer to the center of the isovelocity helix, the greater the number of intersections. The number of the intersection points is counted in the horizontal direction and the vertical direction, an interval with a larger number of the intersection points, namely the general range of the vortex center point, can be obtained by setting a certain error range, and then an optimal center point position is obtained for each vortex edge by a traversal method, as shown in fig. 5. And comparing the obtained optimal vortex center position with the vortex position manually and visually interpreted, and calculating the absolute deviation of the optimal vortex center position and the vortex position for evaluating the inversion error of the vortex center position.
(2) Inverting vortex edge location parameters
1) Calculating all connected regions in the image to obtain a coordinate index of each connected region;
2) calculating the length of each communication area, and screening out the longest 2 main communication area arc sections;
3) curve fitting is carried out on the longest 2 main communication areas, conversion from a Cartesian coordinate system to a polar coordinate system is carried out on points in each arc section, conversion from the Cartesian coordinate system to the polar coordinate system is carried out on the points in each arc section, and spiral line parameters which are best approximate to the vortex edge are calculated by combining with least square judgment errors, as shown in figure 6.
With a swirl edgeLFor example, the polar equation for an archimedes' spiral is:
wherein the content of the first and second substances,
and
are all real numbers, and are all real numbers,
the diameter of the electrode is the same as the diameter of the electrode,
is a polar angle;
the archimedes spiral coordinate, the transformation equation from cartesian to polar coordinate system is as follows:
wherein the content of the first and second substances,
and
representing the vortex edgeLThe coordinates of the point(s) of (c),
and
represents the position coordinates of the vortex center,
the diameter of the electrode is the same as the diameter of the electrode,
is a polar angle; for edgesLAt any point on
Corresponding to
And distance
The following:
substituting the formulas (4) and (5) into the formula (1), and calculating by least square method to obtain the constant-speed spiral line parameter which best approximates to the vortex edge
And
；
4) measuring the fitting effect of the fitting curve using the maximum absolute error
Maximum absolute error, i.e. the distance the worst case edge point is measured from the curve:
wherein the content of the first and second substances,
the distance of the edge points from the fitted curve.
Example two:
the embodiment provides an apparatus for automatically identifying ocean black vortexes and inverting key parameters, a schematic structural diagram of which is shown in fig. 2, wherein the apparatus includes:
the SAR image acquisition module 710 is used for acquiring SAR target image target data to be detected, wherein the target data comprise an ENVISAT, an ERS-1/2 satellite, a Sentinel-1 satellite and a GF3 satellite SAR load C-band, VV polarized ocean vortex images of a plurality of sea areas in different time periods
The pre-training module 720 of the ocean black vortex automatic identification model is used for training to obtain an SAR image ocean black vortex automatic identification optimal weight model;
and the ocean black vortex automatic identification module 730 is used for realizing the automatic identification of the sub-mesoscale ocean black vortex in the SAR image and determining the sub-mesoscale vortex in the SAR image.
And the ocean black vortex key parameter inversion module 740 is used for inverting vortex center position and vortex edge position parameters.
The device for automatically identifying ocean black vortexes and inverting the key parameters provided by the embodiment of the invention has the same technical characteristics as the method for automatically identifying ocean black vortexes and inverting the key parameters provided by the embodiment, so that the same technical problems can be solved, and the same technical effect can be achieved. For the sake of brevity, where not mentioned in the examples section, reference may be made to the corresponding matter in the preceding method examples.
Example three:
this embodiment also provides an electronic device comprising a memory for storing a program that enables a processor to perform the method of the first embodiment, and a processor configured to execute the program stored in the memory.
Referring to fig. 3, the present embodiment further provides an electronic device 100, including: the system 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 used to execute executable modules, such as computer programs, stored in the memory 51.
The Memory 51 may include a Random Access Memory (RAM) and 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, a PCI bus, an 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. 3, 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 instructions in the form of hardware integrated logic circuits or 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 also provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to perform the steps of the method in the first embodiment.
Although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the following descriptions are only illustrative and not restrictive, and that the scope of the present invention is not limited to the above embodiments: 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 (7)
1. An ocean black vortex automatic identification and key parameter inversion method is characterized by comprising the following steps:
s101, acquiring SAR image target data;
s102, constructing an SAR image ocean black vortex sample library:
(1) firstly, decoding SAR image target data in the step S101, stretching the image, limiting contrast ratio self-adaptive histogram equalization, and outputting to obtain a picture library A;
(2) secondly, expanding the picture library A to generate a picture library A1 by adopting a random horizontal turning, rotation transformation and translation transformation data expansion method; the picture library A and the picture library A1 jointly form a data set B;
(3) finally, identifying each picture in the data set B by a manual visual method, manually marking the ocean vortex by adopting an external rectangle, marking whether the marked content contains the ocean black vortex, correctly organizing the format of each data set according to the VOC format, and preparing a configuration file to finally obtain an SAR image ocean black vortex sample library;
s103, establishing an ocean black vortex automatic identification model;
s104, inputting the SAR image ocean black vortex sample library in the step S102 into the automatic recognition model in the step S103 for training;
s105, automatically identifying the sub-mesoscale ocean black vortexes by using the trained network model to finally obtain a target prediction result;
s106, performing vortex center position and vortex edge position key parameter inversion on the target prediction result data: (1) and (3) obtaining vortex center position parameters by inversion: scanning the sub-image line by line and column by column, and judging the number of intersection points of a scanning line and a medium-speed spiral line in the image; then, an optimal central point position is obtained for each vortex edge through a traversal method, the obtained optimal vortex central position is compared with a vortex position which is manually and visually interpreted, and the absolute deviation of the two is calculated and used for evaluating the inversion error of the vortex central position;
(2) and (3) inverting vortex edge position parameters:
1) calculating all connected areas in the image, and obtaining a coordinate index of each connected area;
2) calculating the length of each communication area, and screening out the longest 2 main communication area arc sections;
3) performing curve fitting on the longest 2 main communication areas, performing conversion from a Cartesian coordinate system to a polar coordinate system on points in each arc section, and calculating a spiral line parameter which is optimal to approximate to a vortex edge by combining a least square judgment error;
with vortex edgeLThe polar equation of an archimedes' spiral is:
wherein the content of the first and second substances,
and
are all real numbers, and are all real numbers,
the diameter of the electrode is the diameter of the electrode,
is a polar angle;
the archimedes spiral coordinate, the transformation equation from cartesian to polar coordinate system is as follows:
wherein, the first and the second end of the pipe are connected with each other,
and
representing the coordinates of points on the vortex edge L,
and
representing the coordinates of the location of the vortex core,
the diameter of the electrode is the same as the diameter of the electrode,
is a polar angle; to any point on the edge L
Corresponding to
And distance
The following:
substituting the formulas (4) and (5) into the formula (1), and calculating by least square method to obtain constant velocity spiral parameters optimally approximating to the vortex edge
And
；
4) measuring the fitting effect of the fitting curve using the maximum absolute error
Maximum absolute error, i.e. the distance the worst case edge point is measured from the curve:
wherein, the first and the second end of the pipe are connected with each other,
the distance of the edge points from the fitted curve is taken.
2. The method according to claim 1, wherein in step S101, the SAR image target data comprises: VV polarized ocean vortex images of a plurality of sea areas at a plurality of different time periods of SAR loads C-band of an ENVISAT satellite, an ERS-1/2 satellite, a Sentinel-1 satellite and a GF3 satellite; and then, screening and collecting SAR image target data in the global vortex high-incidence sea area by a manual visual method.
3. The method of claim 1, wherein the SAR image stretching parameters are set as follows:
if it is not
，
Wherein the content of the first and second substances,
represents the 0.75 quantile of the product,
it represents a 0.25 quantile of the number,
the minimum value is indicated and is,
represents the median value;
if it is not
，
Wherein the content of the first and second substances,
the standard deviation is expressed in terms of the standard deviation,
the minimum value is indicated and is,
represents the mean value;
the contrast-limited adaptive histogram equalization method specifically comprises the following steps:
1) 8 multiplied by 8 blocking the SAR image, calculating a histogram by taking the block as a unit, setting a cutting amplitude limit to 10, and carrying out equalization;
2) and (3) interpolating pixels among blocks, wherein the pixel interpolation method comprises the following steps:
known functionfIn (1)
，
Calculating the value of four points, calculating the unknown function f At the point of
A value of (d);
interpolation is performed in the x direction to obtain:
interpolation is performed in the y direction to obtain:
interpolated result
The following:
4. the method according to claim 1, wherein the step S104 is embodied as;
randomly dividing the sample library into a training set P1, a verification set P2 and a test set P3 according to 80%, 10% and 10% of the sample library, wherein the training set P1 is used for training a network model, and the training set P1 is sent into the network model established in the step S103 to be trained to obtain a plurality of weight models; the verification set P2 is used for adjusting each parameter in the model, when a plurality of weight models are trained in the training set P1, different weight models are respectively used for predicting the P2, and the parameter corresponding to the weight model with the highest score after prediction is selected, so that the optimal weight model is generated; after the optimal weight model is obtained through the training set P1 and the verification set P2, the test set is used for testing, and the obtained optimal weight model is used for identifying the ocean black vortexes in the SAR picture.
5. An apparatus for automatic identification of ocean black vortexes and key parameter inversion using the method of claim 1, wherein the apparatus comprises:
the SAR image acquisition module is used for acquiring target data of an SAR target image to be detected;
the pre-training module of the ocean black vortex automatic identification model is used for training to obtain an SAR image ocean black vortex automatic identification optimal weight model;
the automatic identification module of ocean black vortexes is used for realizing the automatic identification of the sub-mesoscale ocean black vortexes in the SAR image and determining the sub-mesoscale ocean black vortexes in the SAR image;
and the ocean black vortex key parameter inversion module is used for inverting vortex center position and vortex edge position parameters.
6. An electronic device comprising a memory for storing a program that enables a processor to perform the method of any of claims 1 to 4 and a processor configured to execute the program stored in the memory.
7. 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 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210401146.1A CN114494894B (en) | 2022-04-18 | 2022-04-18 | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210401146.1A CN114494894B (en) | 2022-04-18 | 2022-04-18 | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114494894A CN114494894A (en) | 2022-05-13 |
| CN114494894B true CN114494894B (en) | 2022-07-22 |
Family
ID=81489566
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210401146.1A Active CN114494894B (en) | 2022-04-18 | 2022-04-18 | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114494894B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114859352B (en) * | 2022-07-07 | 2022-10-28 | 自然资源部第二海洋研究所 | Adaptive stretching method and device for SAR satellite marine observation image |
| CN116451465A (en) * | 2023-04-17 | 2023-07-18 | 中国人民解放军61540部队 | Satellite-borne SAR mesoscale vortex imaging simulation method and system |
| 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 |
| CN117576553B (en) * | 2024-01-15 | 2024-04-02 | 中国海洋大学 | Dual-polarized SAR image ocean ice vortex identification method and device and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107505616A (en) * | 2017-09-15 | 2017-12-22 | 浙江大学 | A kind of decision method of the ocean surface wind retrieving optimal resolution based on SAR |
| CN111693999A (en) * | 2020-05-27 | 2020-09-22 | 哈尔滨工程大学 | Multi-sensor fusion wind speed and direction estimation method based on radar wind measurement combination strategy |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11455758B2 (en) * | 2018-11-16 | 2022-09-27 | The Government of the United States of America, as represented by the Secretarv of the Navy | Oceanic eddy detection from two-dimensional sea level topography gradients |
| CN110298280B (en) * | 2019-06-20 | 2023-04-07 | 上海海洋大学 | Ocean vortex identification method based on MKL multi-feature fusion |
| CN113821986B (en) * | 2021-11-22 | 2022-02-22 | 中国海洋大学 | Vortex sea surface signal and underwater key parameter inversion method based on random forest |
| CN114022774A (en) * | 2022-01-10 | 2022-02-08 | 航天宏图信息技术股份有限公司 | Radar image-based marine mesoscale vortex monitoring method and device |
-
2022
- 2022-04-18 CN CN202210401146.1A patent/CN114494894B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107505616A (en) * | 2017-09-15 | 2017-12-22 | 浙江大学 | A kind of decision method of the ocean surface wind retrieving optimal resolution based on SAR |
| CN111693999A (en) * | 2020-05-27 | 2020-09-22 | 哈尔滨工程大学 | Multi-sensor fusion wind speed and direction estimation method based on radar wind measurement combination strategy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114494894A (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114494894B (en) | Ocean black vortex automatic identification and key parameter inversion method and device and electronic equipment | |
| CN112084869B (en) | Compact quadrilateral representation-based building target detection method | |
| CN109033950B (en) | Vehicle illegal parking detection method based on multi-feature fusion cascade depth model | |
| CN101923711B (en) | SAR (Synthetic Aperture Radar) image change detection method based on neighborhood similarity and mask enhancement | |
| CN109697441B (en) | Target detection method and device and computer equipment | |
| CN111860336A (en) | High-resolution remote sensing image inclined ship target detection method based on position sensing | |
| CN109360190B (en) | Building damage detection method and device based on image superpixel fusion | |
| CN110796048A (en) | Ship target real-time detection method based on deep neural network | |
| WO2018000252A1 (en) | Oceanic background modelling and restraining method and system for high-resolution remote sensing oceanic image | |
| CN110675374B (en) | Two-dimensional image sewage flow detection method based on generation countermeasure network | |
| CN109165603B (en) | Ship detection method and device | |
| CN113033315A (en) | Rare earth mining high-resolution image identification and positioning method | |
| CN114022774A (en) | Radar image-based marine mesoscale vortex monitoring method and device | |
| CN112580542A (en) | Steel bar counting method based on target detection | |
| CN111815708B (en) | Service robot grabbing detection method based on dual-channel convolutional neural network | |
| CN111738071B (en) | Inverse perspective transformation method based on motion change of monocular camera | |
| CN113095316B (en) | Image rotation target detection method based on multilevel fusion and angular point offset | |
| CN112906689B (en) | Image detection method based on defect detection and segmentation depth convolutional neural network | |
| CN114565824A (en) | Single-stage rotating ship detection method based on full convolution network | |
| CN116778176B (en) | SAR image ship trail detection method based on frequency domain attention | |
| CN112329677A (en) | Remote sensing image river target detection method and device based on feature fusion | |
| CN112308040A (en) | River sewage outlet detection method and system based on high-definition images | |
| CN114240940B (en) | Cloud and cloud shadow detection method and device based on remote sensing image | |
| CN116612097A (en) | Method and system for predicting internal section morphology of wood based on surface defect image | |
| CN114758123A (en) | Remote sensing image target sample enhancement method |
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 |