Radar sea ice monitoring method and system

Abstract

The invention relates to the technical field of sea ice monitoring, and particularly discloses a radar sea ice monitoring method and system, and the method comprises the steps: transmitting and receiving a radar signal to the sea surface through a radar system, obtaining a sea ice echo signal of a target sea ice sample, carrying out the preprocessing of the sea ice echo signal of the target sea ice sample, extracting the feature information of the sea ice of the target sea ice sample, and carrying out the calculation of the feature information. And analyzing the feature information of the sea ice of the target sea ice sample by adopting a statistical pattern classification method based on the existing sea ice feature database and utilizing a step-by-step discrimination algorithm of a Bayesian criterion, identifying the sea ice type, and obtaining a sea ice classification result and sea ice element information of the target sea ice sample. Real-time, efficient and accurate monitoring of sea ice is achieved, ocean transportation and resource development are supported, the method is applied to multiple fields, sea ice characteristic quantity is extracted, classification accuracy is improved by adopting statistical mode classification, and important data support is provided for the Bohai sea ice area and related ocean environment monitoring, scientific research, engineering design and the like.

Description

Technical Field

[0001] The invention relates to the technical field of sea ice monitoring. Background Art

[0002] Sea ice poses a serious threat to marine transportation, marine resource development and marine environmental safety. Traditional sea ice monitoring methods mainly rely on manual observation and satellite remote sensing, but these methods have problems such as limited monitoring range, untimely data updates and insufficient accuracy. Summary of the invention

[0003] In order to solve the problems of limited monitoring range, untimely data update, insufficient accuracy and the like of existing sea ice monitoring methods, the present invention provides a radar sea ice monitoring method and system.

[0004] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: a radar sea ice monitoring method, comprising the following steps:

[0005] S1. Use a radar system to transmit and receive radar signals to the sea surface to obtain sea ice echo signals of target sea ice samples;

[0006] S2. preprocessing the sea ice echo signal of the target sea ice sample to extract characteristic information of the sea ice of the target sea ice sample;

[0007] S3. Using the statistical pattern classification method, based on the existing sea ice characteristic database and using the Bayesian stepwise discriminant algorithm, the characteristic information of the sea ice of the target sea ice sample is analyzed to identify the sea ice type and obtain the sea ice classification results and sea ice element information of the target sea ice sample.

[0008] Preferably, in step S1, the radar system is an X-band radar system.

[0009] Preferably, in step S1, meteorological data such as wind speed, wind direction, temperature, humidity and air pressure are monitored in real time, and the meteorological data are provided to the radar system as correction parameters. The radar system corrects the signal transmission and reception errors caused by meteorological factors according to the correction parameters.

[0010] Preferably, in step S1, the sea ice echo signal is a signal generated when the electromagnetic wave signal emitted by the radar system to the sea surface is scattered or reflected back by the target sea ice sample after encountering the target sea ice sample. The sea ice echo signal carries characteristic information of the target sea ice sample.

[0011] Preferably, in step S2, data inversion is performed on the extracted sea ice characteristic information to obtain more accurate sea ice characteristic information.

[0012] Preferably, in step S2, the characteristic information of the sea ice of the target sea ice sample is analyzed, and 21 features are selected to form a 21-dimensional feature vector, and the 21-dimensional feature vectors are: power spectrum features of rectangular window, power spectrum features of Parzen window, power spectrum features of Hamming window, power spectrum features of Hamming window, autocorrelation function features, normalized median, mean, variance, standard deviation, range, mean deviation, skewness coefficient, kurtosis coefficient, time correlation coefficient, standard mean deviation, standard skewness coefficient, standard kurtosis coefficient, standard time correlation coefficient, median, median and implicit period; based on the stepwise discriminant algorithm of the Bayesian criterion, the feature quantities with statistical significance in the 21-dimensional feature vector are selected to obtain a group of new feature quantities, and the discriminant function of each group of new feature quantities is calculated by using the training iteration method.

[0013] Preferably, in step S3, according to the discriminant function:

[0014] g i (x) = f(p(w i x);

[0015] Among them, f() is a monotonically increasing function, p(w i |x) indicates that sample x belongs to sea ice type w i The probability of

[0016] The discriminant function is used to quantify the similarity or difference between the target sea ice sample and the sea ice samples in the existing sea ice characteristic database, and then it is determined which sea ice type of the target sea ice sample belongs to in the existing sea ice characteristic database.

[0017] Preferably, in step S3, the sea ice types include primary ice, ice skin, Nile ice, lotus leaf ice, gray ice, gray-white ice and white ice; the sea ice element information includes ice surface state, ice thickness, flow rate, ice volume, floating ice density and maximum floating ice block size; the ice surface state includes flat ice, overlapping ice and snow-covered ice.

[0018] A radar sea ice monitoring system, comprising:

[0019] A radar module is used to transmit and receive radar signals to the sea surface to obtain sea ice echo signals of target sea ice samples;

[0020] A data acquisition and processing module is used to connect with the radar module to collect sea ice echo signal data of the target sea ice sample and perform preprocessing to extract characteristic information of the sea ice of the target sea ice sample;

[0021] The sea ice recognition module is used to analyze the characteristic information of the sea ice of the target sea ice sample, identify the sea ice type, and obtain the sea ice classification results and sea ice element information of the target sea ice sample.

[0022] Preferably, it also includes:

[0023] A data inversion module, which is used to perform data inversion on the sea ice echo signal data of the preprocessed target sea ice sample, so as to obtain more accurate characteristic information of the sea ice of the target sea ice sample;

[0024] A display and output module, which is used to display the sea ice classification result and sea ice element information of the target sea ice sample on the screen of the application terminal;

[0025] A video linkage module, which is used to obtain the impact data of the sea ice sample, and combine spectral and color temperature analysis to monitor the change of the sea ice sample in real time;

[0026] A weather station module, which is used to monitor wind speed, wind direction, air temperature, humidity and air pressure in real time, and provide calibration parameters for the radar system;

[0027] A storage system, which is used to store data;

[0028] A network module, which is used for data transmission.

[0029] The beneficial effects of the present invention are as follows:

[0030] The present invention realizes real-time, efficient and accurate monitoring of sea ice, provides important data support for ocean transportation and ocean resource development, and can be applied to oil and gas field exploration in the Bohai ice area, ice prevention and production of oil and gas pipelines in the ice area, ice tanker external transportation operation in the ice area, sea ice early warning and emergency response, ports, waterways, anchorages in the Bohai ice area, sea ice monitoring and forecasting, Bohai sea ice monitoring and engineering sea ice, marine environment monitoring, sea ice scientific research and engineering sea ice design and other fields. A variety of characteristic quantities representing sea ice characteristics are extracted, and a statistical pattern classification method is used for sea ice identification and classification, improving the accuracy and reliability of classification. At the same time, combined with the weather station system and the video linkage system, it provides more comprehensive all-weather, continuous and real-time monitoring of sea ice information. Description of the Drawings

[0031] Figure 1 It is the sea ice thickness inversion data of the embodiment of the present invention;

[0032] Figure 2 It is a schematic diagram of the radar sea ice monitoring process of the embodiment of the present invention. Detailed Embodiments

[0033] As Figure 2 shown, this embodiment provides a radar sea ice monitoring method, including the following steps:

[0034] S1. Use an X-band radar system to transmit and receive radar signals to the sea surface, monitor meteorological data such as wind speed, wind direction, temperature, humidity and air pressure in real time, and provide the meteorological data to the radar system as correction parameters. The radar system corrects the signal transmission and reception errors caused by meteorological factors according to the correction parameters to obtain the sea ice echo signal of the target sea ice sample. The sea ice echo signal is the electromagnetic wave signal emitted by the radar system to the sea surface. After encountering the sea ice, it is scattered or reflected by the sea ice. The echo signal carries the characteristic information of the sea ice, and performs A / D conversion processing on the sea ice echo signal of the target sea ice sample, digitizes and images the converted sea ice echo signal of the target sea ice sample, and obtains a picture of the target sea ice sample;

[0035] The X-band radar system includes a radar antenna unit, a host unit, a radar sea ice image recorder and a sea ice collection unit; the performance indicators of the radar antenna unit are: X-band, wavelength 3cm, antenna length 3.6 meters, this antenna length is more suitable for sea ice measurement dedicated to marine remote sensing, vertical polarization mode, more suitable for marine measurement, beam width 0.65°×23°, high azimuth resolution, 45ns-1.2us narrow pulse, distance accuracy 30m, azimuth accuracy: 1°, sea ice scanning radius ≥6-8 nautical miles, transmission power 25kw; the performance indicators of the host unit are: display resolution 1280×1024, image color 32-level grayscale, support 0.1-24 It can automatically track 200 targets within the nautical mile range, support unmanned remote control of radar via the network, and support remote control of the chart version radar. The radar sea ice image recorder used is the HB / HT20-1 radar ice measuring integrated machine, which has a built-in USB3.0 interface and a digital radar echo high-speed collector, and can collect, display, and store raw data or images of sea ice samples in real time. The performance indicators of the sea ice collection unit are: sampling frequency 50-100mhz, amplitude resolution 8-12bit, radar map resolution 3m, 6m, 9m, map size 4000x4000 (pixels), sampling speed 24-28 frames / minute, and ice type and ice thickness recognition rate ≥90%.

[0036] S2. Preprocess and invert the sea ice echo signal of the target sea ice sample, such as Figure 1 Each pixel point shown is output as a text in the form of longitude and latitude, and the characteristic information of the sea ice of the target sea ice sample is extracted;

[0037] According to the radio wave scattering mechanism, radio waves have a certain penetrating effect on ice. The interference between the upper surface radio waves and the lower surface radio waves can affect the echo amplitude and waveform. The roughness of the ice surface directly determines the echo intensity. Generally speaking, the increase in roughness increases the scattering. Under the dynamic action of wind, current and waves, sea ice squeezes, breaks and overlaps each other, making the surface roughness of ice increase. The surface roughness of accumulated ice is even greater, and the average scattering coefficient σ 0It can be used as an important basis for ice thickness classification. The characteristic information of sea ice of the target sea ice sample is statistically analyzed to obtain the average scattering coefficient σ 0 As shown in the following two tables, according to the GB / T14914-2006 specification, the thickness of gray ice, gray-white ice, and white ice decreases with the surface roughness of ice. The scattering is stronger at each level. The crushed ice is mixed with water, and the wave pattern fluctuates severely, making it easier to identify. The echo in the water area is very weak, and the echo intensity is very different from that of gray ice, gray-white ice, and white ice. Compared with lotus leaf ice, the ridged edge of lotus leaf ice in the water area has obvious differences in statistical characteristics in terms of echo.

[0038]

[0039]

[0040] At the same time, 21 features are selected from the characteristic information of sea ice in the target sea ice sample to form a 21-dimensional feature vector. The 21-dimensional feature vectors are: power spectrum characteristics of rectangular window, power spectrum characteristics of Parzen window, power spectrum characteristics of Hamming window, power spectrum characteristics of Hamming window, autocorrelation function characteristics, normalized median, mean, variance, standard deviation, range, mean deviation, skewness coefficient, kurtosis coefficient, time correlation coefficient, standard mean deviation, standard skewness coefficient, standard kurtosis coefficient, standard time correlation coefficient, median, median and implicit period; based on the stepwise discriminant algorithm of the Bayesian criterion, the feature vectors with statistical significance in the 21-dimensional feature vectors are selected to obtain a new set of feature vectors, and the discriminant function of the new feature vector is calculated using the training iteration method.

[0041] S3. Using the statistical pattern classification method, based on the existing sea ice feature database, the discriminant function of the stepwise discriminant algorithm of the Bayesian criterion is used to analyze the characteristic information of the sea ice of the target sea ice sample. The discriminant function is:

[0042] g i (x) = f(p(w i |x);

[0043] Among them, f() is a monotonically increasing function, p(w i |x) indicates that sample x belongs to sea ice type w i The probability of

[0044] The similarity or difference between the target sea ice sample and the sea ice samples in the existing sea ice characteristic database is quantified by the discriminant function, and then the sea ice type of the target sea ice sample is determined to which ice type in the existing sea ice characteristic database the sea ice sample belongs, and the sea ice classification result and sea ice element information of the target sea ice sample are obtained;

[0045] Sea ice types include primary ice, ice skin, Nile ice, lotus leaf ice, gray ice, gray-white ice and white ice; sea ice element information includes ice surface state, ice thickness, flow rate, ice volume, floating ice density and maximum floating ice block size; ice surface state includes flat ice, overlapping ice and snow-covered ice;

[0046] The flow velocity calculation steps are as follows: obtain more than three consecutive images of the target sea ice sample, automatically identify the typical ice area on the first image through cross-correlation analysis of multiple images, then track the drift data of the ice through multiple images, and finally divide the number of drift pixels of the ice by the interval time of multiple images to obtain the flow velocity of the floating ice, and obtain the flow direction of the floating ice according to the pixel positions of the starting and ending points of the ice;

[0047] The calculation steps of ice volume are as follows: grayscale identification is performed on each pixel of the target sea ice sample image, and then the value of the threshold range <11 is located as sea water, and the value of the threshold range 12-240 is defined as sea ice. The ice and water are separated first, and then the proportion of floating ice in the image is calculated;

[0048] The calculation steps of floating ice density are as follows: identify the target sea ice sample image, remove the large water areas with connected pixels of 300*300 and less than gray value 11 according to the threshold, and the remaining sea areas are ice areas. Extract all ice areas on the map, calculate the water area in each ice area, and then divide the water area by the ice area. The percentage obtained is the floating ice density;

[0049] The calculation steps of ice surface features are as follows: pre-process the target sea ice sample image, grayscale the image, and then perform filtering and azimuth correction. Finally, the grayscale value thickness macro and deep learning shape judgment are introduced to obtain the thickness value of each pixel and block, and then approximate to form connected pixels for connection judgment. Then, the graphic boundary calculation module is introduced to identify the entire radar image as a radar sea ice image composed of multiple ice blocks. Each complete ice block is counted as an independent ice block. The thickness within a certain range is considered to be the same thickness value. Then, the floating ice with the same thickness value is compared. The median value of each ice block is calculated. The pixel value difference that deviates from the median value is greater than 2, which is an uneven area. If the number of pixel points with a difference value greater than 2 for each ice block exceeds 60%, it is overlapping ice, and if it is less than 40%, it is flat ice. If the proportion is equal, there are both.

[0050] The steps for calculating the maximum size of floating ice are as follows: find the largest piece of connected sea ice pixels based on the target sea ice sample image, and then calculate the maximum diagonal length of the floating ice as the maximum size of the floating ice.

[0051] S4. The sea ice classification results and sea ice element information of the target sea ice samples are transmitted to the application station in real time and stored.

[0052] This embodiment also provides a radar sea ice monitoring system, including:

[0053] The radar module is used to transmit and receive radar signals to the sea surface through an X-band radar to obtain sea ice echo signals of the target sea ice samples. The radar sea ice image recorder performs A / D conversion on the sea ice echo signals of the target sea ice samples, digitizes and images the converted sea ice echo signals of the target sea ice samples, and automatically saves them.

[0054] A data acquisition and processing module is used to connect with the radar module to collect sea ice echo signal data of the target sea ice sample and perform preprocessing to extract characteristic information of the sea ice of the target sea ice sample;

[0055] The sea ice recognition module is used to analyze the characteristic information of the sea ice of the target sea ice sample, and identify the sea ice type by combining image analysis technologies such as digital image processing, automatic contour recognition and image threshold, so as to obtain the sea ice classification results and sea ice element information of the target sea ice sample;

[0056] A data inversion module is used to perform data inversion on the pre-processed sea ice echo signal data of the target sea ice sample to obtain more accurate sea ice characteristic information of the target sea ice sample;

[0057] A display and output module is used to display the sea ice classification results and sea ice element information of the target sea ice sample in real time on the screen of the application end;

[0058] The video linkage module is used to obtain the impact data of sea ice samples and monitor the changes of sea ice samples in real time by combining spectrum and color temperature analysis;

[0059] The weather station module is used to monitor wind speed, wind direction, temperature, humidity and air pressure in real time, provide correction parameters for the radar system and transmit them to the station for use in marine forecasting, disaster prevention and mitigation;

[0060] The storage module is used to store the sea ice echo signal data of the target sea ice samples. The stored data includes original data, echo data, result data, collector fault data and other log data. When the radar module is working normally, three radar original images are automatically saved every hour. Each original image is about 4M, and the amount of data saved every day is 24*3*4M=288M. The radar original image analysis is performed every six hours, and the result data is about 130M. The amount of data saved every day is 4*13M=520M. The total daily data storage volume is ≈850M;

[0061] Network module, used for data transmission, real-time transmission of sea ice classification results and sea ice element information of target sea ice samples to the station;

[0062] The supporting structure includes a tower or a lifting mast, which can be raised to a height of 10-15m and is used to house the hardware equipment in the radar module.

[0063] The radar sea ice monitoring system of this embodiment can continuously and in real time monitor and obtain sea ice element information, and can transmit it to the station in real time. It can work in severe cold and should adapt to the working environment of -30℃ ~ 60℃, especially in the severe sea conditions in winter. It can effectively ensure the effective acquisition and transmission of sea ice element information, and has the performance of wind resistance, corrosion resistance and other adaptability to the harsh marine environment. The system has an average trouble-free working time of ≥10,000 hours; the average fault recovery time is ≤1 hour, all data storage cycles meet the full ice period, and the complete sea ice data of multi-year ice periods can be stored for a long time. The equipment runs stably, the main system maintenance time interval is greater than half a year, it has remote monitoring, diagnosis and partial maintenance functions, and has power-off protection and restart capabilities for equipment and data.

[0064] The present invention is described by way of embodiments, and those skilled in the art will appreciate that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application belong to the protection scope of the present invention.

Claims

1. A radar sea ice monitoring method, characterized in that: The following steps are involved: S1. Use a radar system to transmit and receive radar signals to the sea surface to obtain sea ice echo signals of target sea ice samples; S2. preprocessing the sea ice echo signal of the target sea ice sample to extract characteristic information of the sea ice of the target sea ice sample; S3. Using the statistical pattern classification method, based on the existing sea ice characteristic database and using the Bayesian stepwise discriminant algorithm, the characteristic information of the sea ice of the target sea ice sample is analyzed to identify the sea ice type and obtain the sea ice classification results and sea ice element information of the target sea ice sample.

2. The radar sea ice monitoring method according to claim 1, characterized in that: In step S1, the radar system is an X-band radar system.

3. The radar sea ice monitoring method according to claim 1, characterized in that: In step S1, meteorological data such as wind speed, wind direction, temperature, humidity and air pressure are monitored in real time, and the meteorological data are provided to the radar system as correction parameters. The radar system corrects the signal transmission and reception errors caused by meteorological factors according to the correction parameters.

4. The radar sea ice monitoring method according to claim 1, characterized in that: In step S1, the sea ice echo signal is a signal that is scattered or reflected back by the target sea ice sample after the electromagnetic wave signal emitted by the radar system to the sea surface encounters the target sea ice sample. The sea ice echo signal carries characteristic information of the target sea ice sample.

5. The radar sea ice monitoring method according to claim 1, characterized in that: In step S2, data inversion is performed on the extracted sea ice characteristic information to obtain more accurate sea ice characteristic information.

6. The radar sea ice monitoring method according to claim 1, characterized in that: In the step S2, the characteristic information of the sea ice of the target sea ice sample is analyzed, and 21 features are selected to form a 21-dimensional feature vector, and the 21-dimensional feature vectors are: power spectrum features of rectangular window, power spectrum features of Parzen window, power spectrum features of Hamming window, power spectrum features of Hamming window, autocorrelation function features, normalized median, mean, variance, standard deviation, range, mean deviation, skewness coefficient, kurtosis coefficient, time correlation coefficient, standard mean deviation, standard skewness coefficient, standard kurtosis coefficient, standard time correlation coefficient, median, median and implicit period; based on the stepwise discriminant algorithm of the Bayesian criterion, the feature quantities with statistical significance in the 21-dimensional feature vector are selected to obtain a group of new feature quantities, and the discriminant function of each group of new feature quantities is calculated by using the training iteration method.

7. The radar sea ice monitoring method according to claim 6, characterized in that: In step S3, according to the discriminant function: g i (x)=f(p(w i | x); Among them, f() is a monotonically increasing function, p(w i| x) indicates that sample x belongs to sea ice type w i The probability of The discriminant function is used to quantify the similarity or difference between the target sea ice sample and the sea ice samples in the existing sea ice characteristic database, and then it is determined which sea ice type of the target sea ice sample belongs to in the existing sea ice characteristic database.

8. The radar sea ice monitoring method according to claim 1, characterized in that: In step S3, the sea ice types include primary ice, ice skin, Nile ice, lotus leaf ice, gray ice, gray-white ice and white ice; the sea ice element information includes ice surface state, ice thickness, flow rate, ice volume, floating ice density and maximum floating ice block size; the ice surface state includes flat ice, overlapping ice and snow-covered ice.

9. A radar sea ice monitoring system, characterized in that: include: A radar module is used to transmit and receive radar signals to the sea surface to obtain sea ice echo signals of target sea ice samples; A data acquisition and processing module is used to connect with the radar module to collect sea ice echo signal data of the target sea ice sample and perform preprocessing to extract characteristic information of the sea ice of the target sea ice sample; The sea ice recognition module is used to analyze the characteristic information of the sea ice of the target sea ice sample, identify the sea ice type, and obtain the sea ice classification results and sea ice element information of the target sea ice sample.

10. The radar sea ice monitoring system according to claim 9, characterized in that: Also includes: A data inversion module is used to perform data inversion on the pre-processed sea ice echo signal data of the target sea ice sample to obtain more accurate sea ice characteristic information of the target sea ice sample; A display and output module is used to display the sea ice classification results and sea ice element information of the target sea ice sample on the screen of the application end; The video linkage module is used to obtain the impact data of sea ice samples and monitor the changes of sea ice samples in real time by combining spectrum and color temperature analysis; Weather station module, used to monitor wind speed, wind direction, temperature, humidity and air pressure in real time, and provide correction parameters for the radar system; Storage system for storing data; Network module, used for data transmission.

Images