Coastal Wind Force Estimation Method and System Based on Optical Characteristics of Water Pattern

Abstract

The present invention discloses a method and system for estimating coastal wind force based on the optical characteristics of water ripple patterns. The method includes: obtaining sea surface video data of a preselected coastal area; intercepting continuous videos with stable water ripple areas; parsing the intercepted continuous videos into multiple frames of continuous images and extracting water ripple features; the water ripple features include the amplitude and frequency of sea wave vibrations and the angle of sunlight irradiation; the process of extracting water ripple features includes: identifying specular point pixels in all images, determining the proportion of specular pixel points, and calculating and extracting the vibration frequency and amplitude of sea waves; inputting the extracted water ripple features into a wind force level measurement model constructed based on water ripple patterns to obtain an evaluation result of the coastal site wind force level. This application evaluates the wind force on site through image analysis, without the need to deploy high-cost hardware facilities, does not rely on long-term maintenance of equipment, meets the wind field measurement requirements of a small area, and improves the accuracy of wind field changes.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent wind power assessment, and particularly to a coastal wind power estimation method and system based on the optical characteristics of water ripple patterns. Background Art

[0002] The accurate acquisition of coastal wind power is an important support for ocean operations. Traditional coastal wind power assessment needs to rely on the punctuation data of meteorological stations and is achieved by combining on-site manual measurement methods. It is impossible to accurately measure the small-scale wind power of the entire coastal area. The commonly used existing coastal wind power measurement tools, such as the WRS3000 marine measurement system, which is an integrated multi-layer wind speed and direction sensor, temperature and humidity sensor, and marine physical parameter (wave, current) measurement module, supports long-term continuous monitoring. However, the hardware cost and deployment cost are very high, and this system requires long-term maintenance and regular calibration of sensors. It relies on the patrol and maintenance of professional technicians for a long time, and the maintenance cost is relatively large. When plankton adheres, it will also affect the test accuracy. Therefore, it does not meet the current requirements for coastal wind power testing. In the prior art, satellite remote sensing images are also used for wind power measurement, but due to the problem of spatial resolution, the resolution of microwave scatterometers is usually above 10 kilometers, which is not accurate enough for small-scale wind field changes. Although SAR has high resolution, its revisit period is long and it cannot monitor in real time. Summary of the Invention

[0003] Therefore, the purpose of the present invention is to provide a coastal wind power estimation method and system based on the optical characteristics of water ripple patterns, excluding influencing factors such as tides and ship travel, and realizing fully automated assessment of the wind power level in a designated sea area.

[0004] To achieve the above purpose, a coastal wind power estimation method based on the optical characteristics of water ripple patterns provided by the present invention includes the following steps:

[0005] S1. Obtain the sea surface video data of a preselected coastal area;

[0006] S2. Intercept continuous videos with stable water ripple areas from the sea surface video data;

[0007] S3. Analyze the intercepted continuous videos into multiple frames of continuous images, and extract water ripple features from the multiple frames of continuous images; the water ripple features include the wave vibration amplitude, vibration frequency, and sunlight irradiation angle; extracting the water ripple features from the multiple frames of continuous images includes: identifying the reflective pixel points in all images, determining the proportion of reflective pixel points, and calculating the vibration frequency and vibration amplitude of the waves for extraction;

[0008] S4. Input the extracted water ripple features into the constructed wind power level calculation model based on the water ripple pattern to obtain the coastal on-site wind power level assessment result.

[0009] Further preferably, in S1, the preselected coastal area is selected according to the following rules:

[0010] Select a sea area with relatively less human interference;

[0011] Set up high-definition cameras along the selected sea area, and adjust the camera positions to overlook and observe the sea surface;

[0012] Mark the highest high-tide state and the lowest low-tide state along the coastal section of the selected area to ensure that the selected sea area is an all-weather seawater area during observation;

[0013] The sea surface observation area is selected in the area between 30 meters and 200 meters from the sea surface edge under the lowest low-tide state.

[0014] Further preferably, in S2, intercept continuous videos with stable water ripple areas from the sea surface video data, including the following steps:

[0015] Perform frame clipping on the video stream captured by the camera, delete the aerial area, and only retain the sea surface area;

[0016] Parse the remaining video stream to obtain frame images, and extract them at specific time intervals from the frame images;

[0017] Perform ship recognition on the extracted frame images to determine whether the ships are moving;

[0018] Intercept video clips without ships or with stationary ships, and splice the intercepted video clips to obtain continuous videos with stable water ripple areas.

[0019] Further preferably, in S2, the performing ship recognition on the extracted frame images to determine whether the ships are moving; includes the following process:

[0020] Convert the extracted frame images to the gray space;

[0021] Use the Canny operator to extract the gray images of 10 frame images to obtain the edge information images;

[0022] Adopt the method of straight line pattern matching to process the edge information images: calculate the pixel values pixel by pixel, the pixel value of 0 represents black, and 1 represents white. If there is a continuous pixel area with a value of 1 and the continuous length is greater than the threshold, then mark this area as the straight line edge information;

[0023] In the entire image, if there is a high-density straight line edge information within an actual distance of 50 meters, then identify the entity of this straight line edge information as a ship.

[0024] Further preferably, when intercepting video segments where there are no vessels or the vessels are stationary, it further includes:

[0025] Parsing the intercepted video segment into a second-frame image;

[0026] Performing hierarchical processing on all the second-frame images; obtaining multiple image segments, and numbering the image segments from top to bottom as: P1, P2, …, P n；

[0027] Using a second Canny operator to extract edge information from the second-frame image, and selecting 1 image with the least proportion of edge information as an alternative stable water ripple region image;

[0028] Based on the edge information density extracted again from the alternative stable water ripple region image according to the hierarchical image segments, eliminating abnormal wave regions as the finally selected stable water ripple region.

[0029] Further preferably, in S3, the extraction of the vibration frequency of the sea waves includes the following steps:

[0030] Identifying specular point pixels in all images, determining the proportion of specular pixel points and performing curve fitting, and using the average value of the peak interval and trough interval of the fitting curve as the vibration frequency of the sea waves;

[0031] For the extraction of the vibration amplitude of the sea waves, in all images, calculate the proportion of specular pixel points in each image, identify the partial image with the highest proportion of specular point pixels, and calculate the average value of the proportion of specular point pixels as the vibration amplitude of the sea waves.

[0032] Further preferably, it further includes constructing a mapping relationship between the extracted water ripple features and the wind force level, including:

[0033] Dividing the daily sunshine period into multiple sub-periods in units of 10 minutes;

[0034] Based on the sampled date and sub-period, determining the irradiation angle of sunlight in three-dimensional space relative to the coastal area;

[0035] Constructing a database and storing the sea wave vibration amplitude, vibration frequency, and three-dimensional sunlight irradiation angle as a data item;

[0036] Completing the corresponding wind force level for each data item in the database through historical weather data.

[0037] Further preferably, in S4, the wind force level measurement model uses Tensorflow as the machine learning framework;

[0038] Take the daily sea wave vibration amplitude, vibration frequency, and sunlight irradiation angle in the constructed database as the input samples, and take the corresponding wind force level as the sample label;

[0039] Use the sample library formed by samples and labels to weighted cycle train the neural network until the model converges.

[0040] Further preferably, the sample library is made in the following form:

[0041] WaveRange n = aWaveRange n_0 + bWaveRange n_1 + cWaveRange n_2 + dWaveRange n_3；

[0042] WaveFreq n = aWaveFreq n_0 + bWaveFreq n_1 + cWaveFreq n_2 + dWaveFreq n_3；

[0043] WindLevel n = aWindLevel n_0+ bWindLevel n_1 +cWindLevel n_2 +dWindLevel n_3；

[0044] Among them, WaveRange n represents the sea wave vibration amplitude sample; WaveFreq n represents the sea wave vibration frequency sample; WindLevel n represents the wind force level sample; the sea wave vibration amplitude data at the current moment is recorded as WaveRange n_0 , the sea wave vibration frequency is recorded as WaveFreq n_0 , the wind force level is recorded as WindLevel n_0 ; the data of the previous 3 moments are respectively recorded as WaveRange n_1 、WaveRange n_2 、Waverange n_3 ,WaveFreq n_1 、WaveFreq n_2 、WaveFreq n_3 ,WindLevel n_1 、WindLeveln_2 、WindLevel n_3 ; a, b, c, and d are weighted empirical coefficients respectively.

[0045] The present invention also provides a coastal wind force estimation system based on the optical characteristics of water ripple patterns for implementing the above-mentioned coastal wind force estimation method based on the optical characteristics of water ripple patterns, including:

[0046] Coastal sensor module: used to obtain the sea surface video data of a preselected coastal area;

[0047] Sea surface stable water ripple area video segment intercepting module, used to intercept continuous videos with stable water ripple areas in the sea surface video data;

[0048] Water ripple pattern optical feature extraction module; parsing the intercepted continuous videos into multiple frames of continuous images, and extracting water ripple features from the multiple frames of continuous images;

[0049] Water ripple pattern optical feature corresponding wind force level database construction module, used to form a sample library with the extracted water ripple features and the corresponding wind force levels, and training the constructed wind force level calculation model based on water ripple patterns using deep learning algorithms;

[0050] Coastal real-time wind force estimation module, inputting the extracted water ripple features into the constructed wind force level calculation model based on water ripple patterns to obtain the coastal site wind force level evaluation result.

[0051] The disclosed coastal wind force estimation method and system based on the optical characteristics of water ripple patterns of the present application evaluate the oscillation amplitude and oscillation frequency of sea waves through image analysis methods, so as to evaluate the on-site wind force. There is no need to deploy high-cost hardware facilities, and the detection results do not depend on the long-term maintenance of equipment, meeting the wind field measurement requirements of short distances and small ranges, and improving the accuracy of wind field changes. Description of the Drawings

[0052] Figure 1 is a schematic flow chart of the coastal wind force estimation method based on the optical characteristics of water ripple patterns of the present invention;

[0053] Figure 2 is a block diagram of the modules of the coastal wind force estimation method based on the optical characteristics of water ripple patterns of the present invention;

[0054] Figure 3 is a schematic diagram of sea surface area extraction of the coastal wind force estimation method based on the optical characteristics of water ripple patterns of the present invention;

[0055] Figure 4 is a schematic diagram of "stable water ripple area" extraction of the coastal wind force estimation method based on the optical characteristics of water ripple patterns of the present invention. Detailed Embodiments

[0056] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] As Figure 1 shown, an embodiment of the present invention provides a coastal wind force estimation method based on the optical characteristics of water ripple patterns, including the following steps:

[0058] S1. Obtain the sea surface video data of the preselected coastal area; it should be noted that: the camera is installed at a suitable position in the coastal area so that the nearshore sea surface can be observed during both high tide and low tide.

[0059] The preselected coastal area is selected according to the following rules:

[0060] Select a sea surface area with relatively less human interference;

[0061] Install a high-definition camera along the selected sea surface area and adjust the camera position to observe the sea surface from above;

[0062] Mark the highest high tide state and the lowest low tide state along the coastal section of the selected area to ensure that the selected sea surface area is an all-sea area during all-weather observations;

[0063] The sea surface observation area is selected between 30 meters and 200 meters from the sea surface edge under the lowest low tide state.

[0064] The specific implementation steps are as follows:

[0065] Step 1.1. In the coastal area where wind force needs to be measured, try to select a coastal section with fewer sea vessels and fewer people (water sports) on the sea surface to ensure that the acquisition of the optical characteristics of the water ripple patterns of subsequent sea waves is not affected by human interference;

[0066] Step 1.2. Mark the highest high tide state and the lowest low tide state on the selected coastal section;

[0067] Step 1.3. Install a high-definition camera on the shore, and the camera is installed 4 - 6 meters above the ground;

[0068] Step 1.4. Adjust the observation angle of the camera. The camera should observe the sea surface from above, and the sea surface observation follows two principles: 1) It is an all-sea area during all-weather (under high tide and low tide states); 2) Select an appropriate sea surface for observation (the greater the observation distance, the greater the error; if the observation distance is too close, the water surface is prone to turbidity, affecting the subsequent extraction of water ripple optical characteristics);

[0069] Step 1.5. According to the above principles, the sea surface observation area should be selected between 30 meters and 200 meters from the shore (sea surface edge) under the lowest low tide state.

[0070] S2, intercepting a continuous video with a stable water ripple area in the sea surface video data;

[0071] The specific steps include:

[0072] a) Edit the video stream captured by the camera, delete the aerial area, and keep only the sea surface area. This is mainly to avoid misjudgment caused by non-sea surface targets (such as clouds, airplanes, etc.); Figure 3 shown.

[0073] b) The sea surface line vector concentration recognition method is used for the reserved area to identify the ship. The sea surface line vector concentration recognition method includes analyzing the characteristics of the sea surface image. Only in the area where the ship is located in the image, there will be a large number of straight line vectors and regular surface vectors (unchanged rectangles, arcs, etc.). Therefore, the ship in the sea surface can be quickly identified based on this. Note: Other sea surface images, such as water ripples (including dark surfaces and reflective light surfaces), will not have highly concentrated straight line and regular surface elements. The specific steps include the following:

[0074] b1) Parse the retained video stream to obtain frame images, and extract them from the frame images at specific time intervals; for example, if the video segment length is 10 minutes and the camera sampling frequency is 15 frames, then there are One frame image is extracted every 1 minute, for a total of 10 frames, which are used for subsequent ship identification.

[0075] b2) Perform ship recognition on the extracted frame image to determine whether the ship is moving, including:

[0076] Convert 10 frames of images to grayscale space (assuming they were originally in RGB color space);

[0077] Use the Canny operator to extract the grayscale images of 10 frames and obtain the edge information image;

[0078] The edge information image is processed by "straight line mode" matching: the pixel value is calculated pixel by pixel (0 is black, 1 is white). If there is a continuous pixel area with a value of 1, and the continuous length is greater than the threshold (to avoid interference from water marks), it indicates that the area is "straight line edge information";

[0079] In the entire image, if a high density of "straight edge information" appears within the actual distance of 50 meters, the entity of the straight edge information is identified as a ship; it can be explained that since the camera is fixed and the position of the sea surface of the image taken is fixed, the actual length can be converted from the length of the target object in the image;

[0080] Determine whether the ship is moving: If the high-density area of "straight line edge information" is the same in 10 frames of images, it is determined that the ship has not moved in the corresponding time period and is in a stationary state; otherwise, it is determined that the ship is moving.

[0081] b3) intercepting video clips without ships or with stationary ships, and splicing the intercepted video clips to obtain a continuous video with a stable water ripple area.

[0082] Further preferably, when capturing a video clip without a ship or a stationary ship, the method further includes:

[0083] Parsing the captured video clip into a second frame of image;

[0084] All the second frame images are processed in layers to obtain multiple image segments, which are numbered from top to bottom as: P1, P2, ..., P n；

[0085] A second Canny operator is used to extract edge information from the second frame image, and an image with the least edge information is selected as a candidate stable water ripple area image;

[0086] The edge information density of the candidate stable water ripple area image is extracted again according to the layered image segments, and the abnormal wave area is eliminated as the final selected stable water ripple area.

[0087] S3, using the intercepted continuous video to analyze multiple frames of continuous images, and extracting water ripple features from the multiple frames of continuous images; the water ripple features include wave vibration amplitude, vibration frequency and sunlight irradiation angle; extracting water ripple features from the multiple frames of continuous images includes: identifying reflective point pixels in all images, determining the proportion of reflective pixel points, calculating the vibration frequency and vibration amplitude of the waves;

[0088] Further preferably, in S3, the extraction of the vibration frequency of the ocean waves comprises the following steps:

[0089] Identify the reflective pixels in all images, determine the proportion of reflective pixels and perform curve fitting, and use the average value of the peak interval and trough interval of the fitting curve as the vibration frequency of the waves;

[0090] The vibration amplitude of the waves is extracted by calculating the proportion of reflective pixels in each image among all the images, identifying some images with the highest proportion of reflective pixels, and calculating the average proportion of reflective pixels as the vibration amplitude of the waves.

[0091] Further preferably, the method further includes constructing a mapping relationship between the extracted hydrological characteristics and the wind force level, including:

[0092] Divide the daily sunshine period into multiple sub-periods in units of 10 minutes;

[0093] Based on the sampling date and sub-period, determine the irradiation angle of sunlight in three-dimensional space relative to the coastal area;

[0094] Construct a database and store the sea wave vibration amplitude, vibration frequency, and three-dimensional irradiation angle of sunlight as a data item;

[0095] Based on historical weather data, supplement the corresponding wind force level for each data item in the database.

[0096] Judgment of sea wave vibration frequency. Since within a short period (such as 10 minutes), the azimuth angle of sunlight (or other natural light) is relatively fixed, the vibration frequency of sea waves can be evaluated through the periodic light reflection characteristics of sea waves. The implementation method is as follows:

[0097] 301) Extract the video stream of the "stable water pattern area" within a time period (such as 10 minutes);

[0098] 302) Parse the video stream into 9000 frame images;

[0099] 303) Convert all frame images to the grayscale space;

[0100] 304) For each frame image, calculate the pixel value (0 - 255) pixel by pixel;

[0101] 305) Set a threshold to identify the light reflection characteristics of the pixel, such as 200, that is, pixel points with a pixel value greater than 200 are identified as reflecting light (showing an optical characteristic close to white) at that moment;

[0102] 306) Calculate the proportion of light-reflecting pixel points in each of the 9000 images;

[0103] 307) Plot a quadratic curve for the 9000 values, calculate the intervals between the curve peaks (maximum values) and the intervals between the troughs (minimum values); considering that local sea waves on the sea surface may produce non-regular vibrations, therefore, exclude the 500 largest and 500 smallest intervals;

[0104] 308) Take the average of all peak intervals and trough intervals to finally obtain the vibration frequency of the waves: Through conversion based on the sampling rate of frame images, if it is 15 frames / second, the time difference between two images is 1 / 15 second;

[0105] Judgment of sea wave vibration amplitude. Since judging the size of sea waves through the reflection of sunlight on the sea surface is related to the direct sunlight angle, the moment of the video segment needs to be considered. The implementation method is as follows:

[0106] 311) Similar to step 301, parse the video segment into 9000 frame images and calculate the proportion of reflective pixel points in each image;

[0107] 312) To reduce the influence caused by the irregular changes of the waves, eliminate the 1000 frame images with the highest reflection ratio;

[0108] 313) Among the remaining 8000 images, select the 500 images with the highest reflection ratio and take their average value;

[0109] 314) Record the average value of the highest reflection ratio together with the corresponding moment (date and time point).

[0110] S4. Input the extracted water pattern features into the constructed wind force level measurement model based on the water pattern form to obtain the evaluation result of the wind force level at the coastal site. It should be noted that a database of the optical characteristics of the water pattern form corresponding to the wind force level also needs to be established, and the database is used to train the wind force level measurement model;

[0111] The steps for constructing the database of the optical characteristics of the water pattern form corresponding to the wind force level include:

[0112] Step 4.1. Divide the daily sunshine period into multiple "sub-periods" in units of 10 minutes;

[0113] Step 4.2. According to the sampling date and "sub-period", the irradiation angle of sunlight in the three-dimensional space relative to the coastal area can be judged;

[0114] Step 4.3. Construct a database and store the wave vibration amplitude, vibration frequency and three-dimensional irradiation angle of sunlight as a data item;

[0115] Step 4.4. By means of manual on-site determination, complete the corresponding wind force level for each data item in the database.

[0116] In S4, the wind force level measurement model uses TensorFlow as the machine learning framework; the daily wave vibration amplitude, vibration frequency and sunlight irradiation angle in the constructed database are used as input samples, and the corresponding wind force level is used as the sample label;

[0117] Use the sample library formed by the samples and labels to weighted cycle train the neural network until the model converges.

[0118] Further preferably, the sample library is made in the following form:

[0119] WaveRange n = aWaveRange n_0 + bWaveRange n_1+ cWaveRange n_2 + dWaveRange n_3；

[0120] WaveFreq n = aWaveFreq n_0 + bWaveFreq n_1 + cWaveFreq n_2 + dWaveFreq n_3；

[0121] WindLevel n = aWindLevel n_0+ bWindLevel n_1 +cWindLevel n_2 +dWindLevel n_3；

[0122] Among them, WaveRange n represents the sea wave vibration amplitude sample; WaveFreq n represents the sea wave vibration frequency sample; WindLevel n represents the wind force level sample; The sea wave vibration amplitude data at the current moment is denoted as WaveRange n_0 , the sea wave vibration frequency is denoted as WaveFreq n_0 , and the wind force level is denoted as WindLevel n_0 ; The data of the previous 3 moments are respectively denoted as WaveRange n_1 , WaveRange n_2 , Waverange n_3 , WaveFreq n_1 , WaveFreq n_2 , WaveFreq n_3 , WindLevel n_1 , WindLevel n_2 , WindLevel n_3 ; a, b, c, and d are respectively weighted empirical coefficients.

[0123] The present invention also provides a coastal wind force estimation system based on the optical characteristics of water pattern morphology for implementing the above-mentioned coastal wind force estimation method based on the optical characteristics of water pattern morphology, including:

[0124] Coastal Sensor Module 10: It is used to obtain the sea surface video data of a preselected coastal area; sensors are deployed at appropriate positions in the coastal area so that the nearshore sea surface can be observed during both high tide and low tide. First, in the coastal area where wind force needs to be measured, try to select coastal sections with fewer sea vessels and fewer people (water sports) to ensure that the subsequent acquisition of the optical characteristics of the water pattern of the waves is not affected by human interference; then, mark the highest high tide state and the lowest low tide state in the selected coastal section; then, set up a high-definition camera on the shore, and the camera is set up 4 - 6 meters above the ground; then, adjust the observation angle of the camera, and the camera should look down at the sea surface, and the sea surface should be observed following two principles: 1) The area should be seawater all day long (under high tide and low tide states); 2) Select an appropriate sea surface for observation (the greater the observation distance, the greater the error; if the observation distance is too close, the water surface is likely to be turbid, affecting the subsequent extraction of water pattern optical characteristics); finally, according to the above principles, the sea surface observation area should be selected between 30 meters and 200 meters from the shore (sea surface edge) under the lowest low tide state.

[0125] Video Segment Interception Module 20 for Stable Water Pattern Area on the Sea Surface, which is used to intercept continuous videos with stable water pattern areas in the sea surface video data. Through the intelligent processing of the video camera, the influencing factors of tides and ship travel on the waves are excluded, and the "stable water pattern area" is selected. First, it is necessary to identify whether there are ships in the video segment captured by the camera in real time: First, perform frame cropping on the video stream captured by the camera to avoid misjudgment caused by non-sea surface targets (such as clouds and airplanes in the sky); as Figure 3 shown; analyze the sea surface image characteristics. Only in the area where the ship is located in the image, there will be a large number of straight vectors and regular surface vectors (unchanged rectangles, arcs, etc.). Therefore, ships on the sea surface can be quickly identified based on this; note: For other sea surface images, such as water patterns (including dark surfaces and reflected light surfaces), there will be no highly concentrated straight lines and regular surface elements; then, parse the video stream into frame images; if the video segment length is 10 minutes and the camera sampling frequency is 15 frames, then there are Ten frame images are extracted, one frame image every 1 minute, for a total of 10 frame images, which are used for subsequent ship recognition. Then, the 10 frame images are converted to the grayscale space (assuming the original is the RGB color space), and the Canny operator is used to extract the grayscale images of the 10 frame images to obtain the edge information image. Then, the edge information image is processed by the "linear mode" matching method: calculate the pixel value pixel by pixel (0 is black, 1 is white). If there is a continuous pixel region with a value of 1 and the continuous length is greater than the threshold (to avoid interference from water ripple parts), it indicates that this region is "linear edge information". Finally, in the entire image, if there is a high density of "linear edge information" within an actual distance of 50 meters, the entity with this linear edge information is recognized as a ship. It should be noted that since the camera is fixed and the sea surface position of the captured image is fixed, the actual length can be calculated by converting the length of the target object in the image. Then, it is necessary to determine whether the ship is moving: if the high-density regions of "linear edge information" are the same in the 10 frame images, it is determined that the ship has not moved and is in a stationary state during the corresponding time period; otherwise, it is determined that the ship is in motion. Then, in the video segment, a sub-video segment without ships or with ships in a stationary state is intercepted. Further, the influence of regular sea waves caused by natural tides needs to be excluded: the intercepted sub-video segment is also parsed into continuous frame images; analyze the wave characteristics of natural tides: usually, the overall waves on the coastline surge towards the coastline periodically; in the sub-video segment, 10 frame images are equally selected. If the sub-video segment is not continuous, only the longest sub-video segment is considered for equal division; the 10 frame images are respectively subjected to hierarchical processing, as shown in Figure 4 shown, divided into multiple image segments, and the image segments are numbered from top to bottom as: P1, P2, …, P n , where the value of n can be manually set by the operator; for the 10 frame images, after being converted to grayscale images (to avoid the influence of color and light and only judge from morphological features), the Canny operator is used to extract edge information; it should be noted that the parameters of the Canny operator need to be adjusted to try to filter out "fine edges" (normal water ripples) during the edge information extraction process; among the 10 frame images, the image with the least proportion of edge information is selected as the "alternative stable water ripple area" image; the proportion of edge information: in the edge information image, the ratio of the number of edge information pixel points to the number of pixel points in the entire image range. Finally, for the "alternative stable water ripple area" (such as the topmost video segment in Figure 4 ), the abnormal wave area in this video segment needs to be removed, such as the circled part in the topmost video segment in Figure 4 . The Canny operator is used to extract edge information, and it is judged by the density of the edge information; it should be noted that the parameters of the Canny operator in this step are different from those of the Canny operator in step 2.4; as shown in Figure 4 the "finally selected area" is the "stable water ripple area" adopted by the present invention.

[0126] Optical feature extraction module 30 for water ripple patterns; parsing the intercepted continuous video into multiple frames of continuous images, and extracting water ripple features from the multiple frames of continuous images;

[0127] Form a sample library with the extracted water ripple features and the corresponding wind force levels, and use deep learning algorithms to train the wind force level measurement model based on water ripple patterns. Through continuous video frame images, calculate the regular features of the water ripples to evaluate the vibration frequency and size of the sea waves affected by the wind force, and thus provide data support for subsequent wind force assessment. First is the judgment of the sea wave vibration frequency. Since within a short period (such as 10 minutes), the azimuth angle of sunlight (or other natural light) is relatively fixed, the vibration frequency of the sea waves can be evaluated through the periodic light reflection characteristics of the sea waves: within a time period (such as 10 minutes), extract the video stream of the "stable water ripple area"; parse the video stream into 9000 frame images; convert all frame images to the grayscale space; for each frame image, calculate the pixel value (0~255) of each pixel; set a threshold to identify the light reflection characteristics of the pixel, such as 200, that is, the pixel points with pixel values greater than 200 are identified as the pixel points reflecting light (showing optical characteristics close to white) at that moment; calculate the proportion of the light-reflecting pixel points in each of the 9000 images; plot a quadratic curve for the 9000 values, and calculate the interval between the peaks (maximum values) and the interval between the valleys (minimum values) of the curve; considering that non-regular vibrations may occur in local sea waves on the sea surface, 500 values with the largest and smallest intervals are excluded; take the average of all peak intervals and valley intervals to finally obtain the vibration frequency of the waves: through the conversion of the frame image sampling rate, such as 15 frames / second, the time difference between two images is 1 / 15 second. Next is the judgment of the sea wave vibration amplitude. Since judging the size of the sea waves through the sea surface light reflection is related to the direct sunlight angle, the time of the video segment needs to be considered: in the same way as the sea wave vibration frequency judgment method, parse the video segment into 9000 frame images and calculate the proportion of the light-reflecting pixel points in each image; to reduce the influence caused by the irregular changes of the sea waves, exclude the 1000 frame images with the highest light reflection ratio; among the remaining 8000 images, take the 500 images with the highest light reflection ratio and take their average value; record the time (date and time point) and the average of the highest light reflection ratio at the same time.

[0128] The morphological features of water patterns mainly include the vibration frequency and amplitude of ocean waves. Using the corresponding database of the morphological features of water patterns and wind force levels, the date and time (the direct sunlight angle) need to be considered. First, divide the daily sunshine period into multiple "sub-periods" in units of 10 minutes; according to the sampling date and "sub-period", the irradiation angle of sunlight in three-dimensional space relative to the coastal area can be determined; then, construct a database and store the wave vibration amplitude, vibration frequency, and three-dimensional irradiation angle of sunlight as a data item; finally, through manual on-site determination, complete the corresponding wind force level for each data item in the database.

[0129] The coastal real-time wind force estimation module 40 inputs the extracted water pattern features into the constructed wind force level calculation model based on water pattern morphology to obtain the evaluation result of the wind force level at the coastal site.

[0130] The wind force level calculation model based on water pattern morphology uses a constructed deep learning model to evaluate the wind force level at the coastal site. First, based on the TensorFlow 2.0 machine learning framework, construct a recurrent neural network model; then, based on the previously constructed database, make a sample library: the samples are the wave vibration amplitude, vibration frequency, and sunlight irradiation angle, and the sample labels are the manually calibrated on-site wind force levels, and train the recurrent neural network until the model converges; finally, use the trained model at the coastal real-time site, and through the optical characteristics of the sea surface water pattern collected by the camera, the wind force level can be evaluated fully automatically.

[0131] Obviously, the above embodiments are only examples for clear illustration and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A coastal wind force estimation method based on the optical characteristics of water ripple patterns, characterized in that, It includes the following steps: S1. Obtain the sea surface video data of a preselected coastal area; S2. Intercept continuous videos with stable water ripple areas from the sea surface video data; S3. Analyze the intercepted continuous videos into multiple frames of continuous images, and extract water ripple features from the multiple frames of continuous images; The water ripple features include the wave vibration amplitude, vibration frequency, and sunlight irradiation angle; The process of extracting water ripple features from multiple frames of continuous images includes: identifying specular point pixels in all images, determining the proportion of specular pixel points, and calculating and extracting the vibration frequency and vibration amplitude of the waves; S4. Input the extracted water ripple features into the constructed wind force level measurement model based on water ripple patterns to obtain the evaluation result of the wind force level at the coastal site.

2. The coastal wind power estimation method based on the optical characteristics of water ripple patterns according to claim 1, characterized in that In S1, the preselected coastal area is selected according to the following rules: Select a sea surface area with relatively less human interference; Set up a high-definition camera along the selected sea surface area, and adjust the camera position to observe the sea surface from above; Mark the highest high tide state and the lowest low tide state along the coastal section of the selected area to ensure that the selected sea surface area is all seawater area during all-weather observation; The sea surface observation area is selected in the area between 30 meters and 200 meters from the sea surface edge under the lowest low tide state.

3. The coastal wind force estimation method based on the optical characteristics of water ripple patterns according to claim 1, characterized in that, In S2, intercepting continuous videos with stable water ripple areas from the sea surface video data includes the following steps: Perform frame cropping on the video stream captured by the camera, delete the aerial area, and only retain the sea surface area; Analyze the remaining video stream to obtain frame images, and extract them at specific time intervals from the frame images; Perform ship recognition on the extracted frame images to determine whether the ship is moving; Intercept video segments without ships or with stationary ships, and splice the intercepted video segments to obtain continuous videos with stable water ripple areas.

4. The coastal wind force estimation method based on the optical characteristics of water ripple patterns according to claim 3, characterized in that, In S2, the process of performing ship recognition on the extracted frame images to determine whether the ship is moving includes the following process: Convert the extracted frame images to the grayscale space; Use the Canny operator to extract the grayscale images of 10 frame images to obtain the edge information images; Process the edge information images in the way of straight line pattern matching: calculate the pixel values pixel by pixel, where the pixel value of 0 represents black and 1 represents white. If there is a continuous pixel area with a value of 1 and the continuous length is greater than the threshold, then mark this area as the straight line edge information; In the whole image, if there is a high density of straight line edge information within an actual distance of 50 meters, then identify the entity of this straight line edge information as a ship.

5. The coastal wind power estimation method based on the optical characteristics of water ripple patterns according to claim 3, wherein When intercepting video segments without ships or with stationary ships, it also includes: Analyze the intercepted video segments into second frame images; Perform a hierarchical processing on all the second frame images; obtain a plurality of image segments, and number the image segments from top to bottom as: P1, P2, …, P n ; Use the second Canny operator to extract the edge information of the second frame images, and select 1 image with the least proportion of edge information as the alternative stable water ripple area image; Based on the edge information density extracted again from the image segments after stratifying the alternative stable water ripple area image, eliminate the abnormal wave areas as the finally selected stable water ripple area.

6. The coastal wind power estimation method based on the optical characteristics of water ripple patterns according to claim 1, wherein, In S3, the extraction of the vibration frequency of the waves includes the following steps: Identify the reflective point pixels in all images, determine the proportion of the reflective pixel points and perform curve fitting, and use the average value of the peak interval and the trough interval of the fitted curve as the vibration frequency of the sea wave; The extraction of the vibration amplitude of the sea wave is carried out by calculating the proportion of the reflective pixel points in each image in all images, identifying the partial image with the highest proportion of the reflective point pixels, and calculating the average value of the proportion of the reflective point pixels as the vibration amplitude of the sea wave.

7. The coastal wind force estimation method based on the optical characteristics of water ripple patterns according to claim 1, wherein It also includes constructing a mapping relationship between the extracted water pattern features and the wind force level, including: Divide the daily sunshine period into multiple sub-periods in units of 10 minutes; According to the sampling date and sub-period, judge the irradiation angle of sunlight in the three-dimensional space relative to the coastal area; Construct a database and store the sea wave vibration amplitude, vibration frequency, and three-dimensional irradiation angle of sunlight as a data item; Complete the corresponding wind force level for each data item in the database through historical weather data.

8. The coastal wind force estimation method based on the optical characteristics of water ripple patterns according to claim 7, wherein In S4, the wind force level measurement model uses TensorFlow as the machine learning framework; Use the daily sea wave vibration amplitude, vibration frequency, and sunlight irradiation angle in the constructed database as the input samples, and use the corresponding wind force level as the sample label; Use the sample library formed by the samples and labels to weightedly and cyclically train the neural network until the model converges.

9. The coastal wind force estimation method based on the optical characteristics of water ripple patterns according to claim 8, wherein, The sample library is made in the following form: WaveRange n = aWaveRange n_0 + bWaveRange n_1 + cWaveRange n_2 + dWaveRange n_3； WaveFreq n = aWaveFreq n_0 + bWaveFreq n_1 + cWaveFreq n_2 + dWaveFreq n_3； WindLevel n = aWindLevel n_0+ bWindLevel n_1 +cWindLevel n_2 +dWindLevel n_3； Among them, WaveRange n represents the sample of the sea wave vibration amplitude; WaveFreq n represents the sample of the sea wave vibration frequency; WindLevel n represents the sample of the wind force level; the sea wave vibration amplitude data at the current moment is denoted as WaveRange n_0 , the sea wave vibration frequency is denoted as WaveFreq n_0 , and the wind force level is denoted as WindLevel n_0 ; the data of the previous 3 moments are respectively denoted as WaveRange n_1 , WaveRange n_2 , Waverange n_3 , WaveFreq n_1 , WaveFreq n_2 , WaveFreq n_3 , WindLevel n_1 , WindLevel n_2 , WindLevel n_3 ; a, b, c, and d are respectively weighted empirical coefficients.

10. A coastal wind force estimation system based on the optical characteristics of water ripple patterns, characterized in that, The coastal wind force estimation method based on the optical characteristics of the water pattern shape described in any one of the above claims 1-9 is implemented, including: Coastal sensor module: used to obtain the sea surface video data of the preselected coastal area; Sea surface stable water pattern area video segment intercepting module, used to intercept continuous videos with stable water pattern areas in the sea surface video data; Water pattern shape optical feature extraction module; use the intercepted continuous video to parse into multiple frames of continuous images, and extract the water pattern features from the multiple frames of continuous images; Water pattern shape optical feature corresponding wind force level database construction module, used to form a sample library of the extracted water pattern features and the corresponding wind force levels, and use deep learning algorithms to train the constructed wind force level measurement model based on the water pattern shape; Coastal real-time wind force estimation module, input the extracted water pattern features into the constructed wind force level measurement model based on the water pattern shape to obtain the coastal site wind force level evaluation result.

Images