Coastal sediment feature monitoring method, device and computer equipment

Abstract

The present application relates to a kind of coast sediment feature monitoring method, device and computer equipment.Therein, coast sediment feature monitoring method is guaranteed by gridding processing and digital format conversion method The consistency and accuracy of data, improve the reliability of analysis result.The sediment feature result of the region to be measured of the sediment feature and the quantitative relationship of the unmanned aerial vehicle data of the establishment is predicted, the sediment feature of the area difficult to reach can be monitored quickly, reduces the labor cost and the danger degree of sediment sample collection.At the same time, the work efficiency is improved by automatic data processing and analysis process, reduces manual intervention, while having strong practicality, applicable to a variety of environmental monitoring and analysis scene.

Description

Technical Field

[0001] This invention relates to the field of coastal sediment characteristic monitoring, and more specifically, to a method, apparatus, computer equipment, and storage medium for monitoring coastal sediment characteristics. Background Technology

[0002] The transport processes of coastal sediments determine the geomorphology and habitats of coastal organisms. Sediment characteristics play a crucial role in the study of sedimentary dynamics, geomorphological dynamics, and ecosystem dynamics. Furthermore, coastal erosion is essentially caused by sediment depletion, and the interaction of energy and matter is primarily manifested in the sediments. Typically, sediment characteristics are obtained by collecting sediment samples in the field and performing grain size analysis in the laboratory. Larger gravel particles are analyzed using sieve analysis, while finer silt or clay particles are analyzed using sedimentation methods. Traditional methods for sediment grain size analysis primarily involve collecting sediment samples in the field and then conducting the analysis in the laboratory, and these methods have been applied to areas such as the seabed, rivers, and deserts.

[0003] Currently, the inventors have discovered that traditional methods suffer from poor reliability and low efficiency. Summary of the Invention

[0004] Therefore, it is necessary to provide a highly efficient and reliable method, device, computer equipment, and storage medium for monitoring the characteristics of coastal sediments.

[0005] To achieve the above objectives, this application provides a method for monitoring coastal sediment characteristics, comprising the following steps:

[0006] Acquire drone data of the coastal zone;

[0007] The characteristic parameters of sediment samples from various sampling points along the coastal zone were measured; the characteristic parameters included both continuous and discrete data.

[0008] Discrete data is converted into numerical data using a pre-defined naming rule for numerical formats.

[0009] The continuous and digital data at each sampling point are rasterized to obtain a raster image; the resolution of the raster image is the same as the sampling diameter of the sampling point.

[0010] A sample set of sediment features is constructed based on raster images, and an aggregate dataset is constructed based on UAV data;

[0011] Based on the ensemble dataset and the sample set of sediment characteristics, a quantitative relationship between sediment characteristics and UAV data is established;

[0012] Quantitative relationships are used to predict sediment characteristics in the area to be tested.

[0013] In one embodiment, the step of establishing a quantitative relationship between sediment features and UAV data based on a dataset and a sample set of sediment features includes:

[0014] An initial random forest model is trained using a set of ensemble datasets and a sample set of sediment features to obtain a trained random forest model; wherein the ensemble dataset and the sample set of sediment features have the same number of rows and columns.

[0015] Based on the trained random forest model, the quantitative relationship between sediment characteristics and UAV data was determined.

[0016] In one embodiment, the drone data includes a digital terrain model, slope data, and aspect data;

[0017] The steps for acquiring drone data of the coastal zone include:

[0018] Collect two-dimensional image data of the coastal zone;

[0019] Process two-dimensional image data to obtain a digital land surface model;

[0020] Based on the digital surface model, slope and aspect data are calculated.

[0021] In one embodiment, the drone data includes multispectral data and normalized vegetation data;

[0022] The steps for acquiring drone data of the coastal zone include:

[0023] The project collected true-color images, blue band data, green band data, red band data, red edge band data, and near-infrared band data of the coastal zone; among which, the true-color images, blue band data, green band data, red band data, red edge band data, and near-infrared band data included calibration plates placed on the coastal zone.

[0024] Obtain physical information about calibration plates placed on the coastline;

[0025] Based on physical information, the blue band data, green band data, red band data, red edge band data, and near-infrared band data are corrected to obtain corrected data;

[0026] Atmospheric correction is applied to the correction data to obtain multispectral data;

[0027] Normalized vegetation data were obtained based on red band and near-infrared band data.

[0028] In one embodiment, the drone data also includes point cloud intensity information; wherein the point cloud intensity information is obtained by a laser scanning device located on the flight device.

[0029] In one embodiment, the step of establishing a quantitative relationship between sediment features and UAV data based on a dataset and a sample set of sediment features includes:

[0030] A quantitative relationship was obtained by performing multiple linear regression fitting on the dataset and the sample set of sediment characteristics.

[0031] In one embodiment, the step further includes:

[0032] The Gao-Collins grain size trend analysis method was used to determine the net transport pattern of sediments based on sediment characteristics.

[0033] On one hand, embodiments of the present invention provide a coastal sediment characteristic monitoring device, comprising:

[0034] The acquisition module is used to acquire drone data from the coastal zone;

[0035] The measurement module is used to measure the characteristic parameters of sediment samples from various sampling points on the coastal zone; the characteristic parameters include continuous data and discrete data.

[0036] The data conversion module is used to convert discrete data into digital data using a preset naming rule for digital formats.

[0037] The raster processing module is used to rasterize the continuous and digital data of each sampling point to obtain a raster image; wherein the resolution of the raster image is the same as the sampling diameter of the sampling point.

[0038] The dataset construction module is used to construct a sample set of sediment features based on raster images and to construct a set dataset based on UAV data.

[0039] The quantitative relationship confirmation module is used to establish a quantitative relationship between sediment characteristics and UAV data based on the ensemble dataset and a sample set of sediment characteristics.

[0040] The prediction module is used to predict sediment characteristics of the area to be tested using quantitative relationships.

[0041] On the one hand, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and a computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the above-described method at runtime.

[0042] On the other hand, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0043] One of the above technical solutions has the following advantages and beneficial effects:

[0044] The aforementioned coastal sediment characteristic monitoring method ensures data consistency and accuracy through rasterization and digital format conversion, thereby improving the reliability of the analysis results. By establishing a quantitative relationship between the sediment characteristics and the UAV data to predict the sediment characteristics of the target area, it is possible to quickly monitor sediment characteristics in hard-to-reach areas, reducing labor costs and the risk of collecting sediment samples. Simultaneously, the automated data processing and analysis workflow improves work efficiency, reduces manual intervention, and is highly practical, applicable to various environmental monitoring and analysis scenarios. Attached Figure Description

[0045] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0047] Figure 1 This is a schematic flowchart of a method for monitoring coastal sediment characteristics in one embodiment;

[0048] Figure 2 This is a schematic diagram illustrating the rules for converting sediment types to digital formats in one embodiment;

[0049] Figure 3 This is a first schematic flowchart of the steps for acquiring drone data of the coastal zone in one embodiment;

[0050] Figure 4 This is a second schematic flowchart illustrating the steps for acquiring drone data of the coastal zone in one embodiment;

[0051] Figure 5 This is a third schematic flowchart of the steps for acquiring drone data of the coastal zone in one embodiment. Detailed Implementation

[0052] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0054] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustration and has no specific meaning in itself. Therefore, "module" and "part" may be used interchangeably.

[0055] It is understood that the term "connection" in the following embodiments should be understood as "electrical connection," "communication connection," etc., if the connected circuits, modules, units, etc., have electrical signal or data transmission with each other.

[0056] When used herein, the singular forms of “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” or “having,” etc., specify the presence of the stated feature, whole, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0057] In one embodiment, such as Figure 1 As shown, a method for monitoring coastal sediment characteristics is provided, including the following steps:

[0058] S110, acquires drone data of the coastal zone;

[0059] The drone data refers to data collected from the coastal zone using aerial equipment. This equipment can be drones equipped with the necessary sensors. Drone data can include digital terrain models, slope data, aspect data, multispectral data, point cloud intensity information, and normalized vegetation data. Multispectral data can be corrected blue band, green band, red band, red-edge band, and near-infrared band data.

[0060] S120 measures characteristic parameters of sediment samples from various sampling points along the coastal zone; the characteristic parameters include continuous and discrete data.

[0061] The characteristic parameters may include water content, average grain size, sorting factor, and skewness. Discrete data is the sediment type represented in text format, while continuous data is the specific numerical value. Average grain size represents the average size of particles in the sample and is usually used to describe the overall grain size characteristics of the sample. The sorting factor represents the dispersion of the grain size distribution in the sample, i.e., the uniformity of grain size. Skewness represents the symmetry of the grain size distribution and reflects the degree of skewness in the grain size distribution.

[0062] Specifically, sediment samples are collected in circular shapes with a diameter of 0.3 m and a thickness not exceeding 4 cm, preferably 2-4 cm. The method for collecting sediment samples is as follows: first, use a thin, hard plastic sheet to gather the sediment samples from the periphery to the center of the entire circular area; then, collect the sediment samples from the center, ensuring the uniformity of the sediment sample characteristics as much as possible, and immediately store them in sealed, numbered sample bags. Each sediment sampling station is recorded three times using GPS-RTK with a fixed solution, and the average value is taken as the location of the sediment sample. In one embodiment, the sampling points are evenly spaced and fully cover the coastal zone of the experimental study area. The number of sediment sampling stations can correspond to the number of flight paths and waypoints. When flying parallel or perpendicular to the coastline, sediment samples should be collected along equally spaced flight paths; in one example, this is at equally spaced waypoints of the UAV. When the number of flight paths is increased, such as when the flight paths are set in a crisscross pattern, sediment samples should be collected at the center of the intersection of the crisscross patterns. The number of sediment sampling stations should be 1 / 8 to 1 / 5 of the number of UAV waypoints, such as 1 / 6. This ensures that the sediment sampling stations are evenly spaced and provide comprehensive coverage, while also ensuring that the sediment sampling stations are centered in the UAV image, reducing image distortion and geometric distortion.

[0063] Furthermore, the moisture content of the sediment samples was measured as follows: First, the sediment samples were uniformly mixed, and approximately 200g of the sample was weighed three times using an electronic balance with a sensitivity of 0.01g. The average weight was taken as the wet weight of the sediment sample. Then, the weighed sediment sample was placed in an oven at 75℃ and dried for at least 3 days until the weight of the sediment sample no longer changed, i.e., dried to constant weight. The dried sediment sample was then weighed again to obtain the dry weight of the sediment sample. The moisture content of the sediment sample was obtained by the following formula:

[0064]

[0065] Wherein, Smoisture content is the moisture content of the sediment sample, Swet weight is the wet weight of the sediment sample, and Sdry weight is the dry weight of the sediment sample.

[0066] The measurement steps for grain size values ​​are as follows: First, the sediment sample is uniformly mixed, and approximately 200g is placed in a beaker. Pure water is added for rinsing, and the mixture is stirred with a glass rod. This process is repeated at least three times to prevent clumping of the dried sediment sample. After rinsing, the sediment sample is allowed to settle, and the supernatant is removed. Then, the sediment sample with the supernatant removed is placed in a 75℃ oven and dried for at least three days. Finally, the dried sediment sample is removed and cooled to room temperature. The pretreated sediment sample is then divided into four equal parts using a sample divider to ensure representativeness. Approximately 50g of the homogeneous sediment sample is taken for testing. Based on the grain size values ​​(average grain size, sorting factor, and skewness) corresponding to the cumulative curve of grain size analysis, the Folk-Ward graphical method is used to calculate the grain size parameters of the sediment sample.

[0067] S130 uses a preset naming rule for digital format to convert discrete data into digital data;

[0068] Specifically, for discrete data, it is necessary to first convert the discrete data into a numerical format. The default naming rules for the numerical format are as follows: a table of m rows and n columns composed of m×n sediment types is called an m×n sediment type matrix, denoted as the sediment type matrix. A table of i rows and j columns composed of i×j sediment type matrices is called an i×j sediment type matrix permutation, denoted as the sediment type matrix permutation. Then, the position of the "sediment type" in the m-th row and n-th column of the "sediment type matrix," and the position of the "sediment type matrix" in the i-th row and j-th column of the "sediment type matrix permutation," are denoted as ijmn, where i, j, m, and n are all positive integers. For example, such as... Figure 2 As shown, firstly, the "sediment types" in each "sediment type matrix arrangement" are named according to their order from top to bottom as rows and from left to right as columns. The two sediment type matrices are denoted as 11 and 12. Then, the "sediment types" in each "sediment type matrix" are named according to their order from top to bottom as rows and from left to right as columns. When the same "sediment type" exists in both "sediment type matrices," a fixed number is used to represent it. For example, if S is located in the 1st row and 1st column of the "sediment type matrix arrangement," the 5th row and 4th column of the "sediment type matrix," and the 1st row and 2nd column of the "sediment type matrix arrangement," then S can be defined as 1154 or 1211. The rules for converting discrete data in sediment features into numerical format provided in this application, which convert discrete data in text format based on the element's position, have strong logic and universality.

[0069] S140, Rasterize the continuous and digital data at each sampling point to obtain a raster image; wherein, the resolution of the raster image is the same as the sampling diameter of the sampling point;

[0070] Specifically, taking a sampling diameter of 0.3 meters as an example, a circular buffer with a radius of 0.15 meters is established for each sediment sample dataset. Each circular buffer is then converted into a raster with a resolution of 0.3 meters, and each raster cell is 0.3 meters × 0.3 meters in size. It is ensured that the center point of the raster is consistent with the location of the sediment sample dataset, so that it can better represent the actual sampling area, ensuring the accurate spatial representation of the sediment sample dataset and making subsequent data analysis and processing more effective.

[0071] S150, Construct a sample set of sediment features based on raster images, and construct an aggregate dataset based on UAV data;

[0072] Specifically, based on the raster image, data from all sampling points are integrated to construct a sediment feature sample set. Relevant variables are extracted from UAV data to construct an aggregate dataset.

[0073] S160, Based on the ensemble dataset and the sample set of sediment characteristics, establish a quantitative relationship between sediment characteristics and UAV data;

[0074] Specifically, quantitative relationships can be established using multiple linear regression to establish a quantitative relationship between sediment characteristics and UAV data (10 variables), or through mathematical models (random forest model) to establish a quantitative relationship between sediment characteristics and UAV data (10 variables).

[0075] S170 uses quantitative relationships to predict sediment characteristics of the area to be tested.

[0076] Specifically, quantitative relationships are used to predict the UAV data of the area under test, thereby outputting the sediment characteristic classification results of the area under test.

[0077] The aforementioned coastal sediment characteristic monitoring method ensures data consistency and accuracy through rasterization and digital format conversion, thereby improving the reliability of the analysis results. By establishing a quantitative relationship between the sediment characteristics and the UAV data to predict the sediment characteristics of the target area, it is possible to quickly monitor sediment characteristics in hard-to-reach areas, reducing labor costs and the risk of collecting sediment samples. Simultaneously, the automated data processing and analysis workflow improves work efficiency, reduces manual intervention, and is highly practical, applicable to various environmental monitoring and analysis scenarios.

[0078] In one embodiment, such as Figure 3As shown, the steps for establishing a quantitative relationship between sediment features and UAV data based on the dataset and a sample set of sediment features include:

[0079] S310, train an initial random forest model using the ensemble dataset and the sample set of sediment features to obtain a trained random forest model; wherein the ensemble dataset and the sample set of sediment features have the same number of rows and columns;

[0080] Specifically, the sediment feature sample set is divided into a sediment feature training set and a sediment feature validation set, with a ratio of approximately 8:2. The model is trained using the dataset containing the 10 variables and the sediment type training set, and the model parameters are adjusted. The dataset containing the 10 variables and the sediment type training set should have the same number of rows and columns. This ensures a one-to-one spatial correspondence between the independent and dependent variables, avoiding spatial attribute mismatches; otherwise, the established quantitative relationship will be inaccurate.

[0081] When tuning the parameters of the Random Forest algorithm, the first parameter is the number of decision trees, which depends on the complexity of the data. A relatively stable number of 500 trees is acceptable, with 400-600 being preferable. This ensures stable model recognition accuracy, avoids overfitting, and reduces computational complexity and hardware requirements. However, given the spatial variation of sediment features and the 10 variables for UAVs, setting the number to 100, 300, 500, and 1000 was also tested to stabilize the Random Forest algorithm's accuracy and achieve optimal results. The second parameter is the number of features randomly selected at each node, chosen as the square root of the total number of features. This selection ensures that the decision trees in the Random Forest algorithm are distinct and independent, increasing diversity and allowing for accurate classification of factors with linear or non-linear relationships. It also avoids overfitting and improves classification performance. The third parameter is the stopping criterion (used for node splitting), choosing a default value of 1 for the minimum number of samples in a node and calculating the minimum impurity based on a Gini coefficient of 0. This choice allows the model to select the optimal feature split points to generate a decision tree, thereby making the classification results more accurate.

[0082] S320 determines the quantitative relationship between sediment characteristics and UAV data based on a trained random forest model.

[0083] Furthermore, the sediment type validation set is linked with the quantitative relationship to verify the accuracy of the sediment type classification results. If the overall model accuracy is greater than 90% and the Kappa coefficient is greater than 0.8, the model accuracy is considered reasonable and reliable. It is also important to ensure that the raster images in the dataset and the sediment type validation set have the same number of rows and columns. This ensures a one-to-one spatial correspondence between the validation and result data, avoiding spatial attribute mismatches; otherwise, the obtained validation accuracy will be inaccurate. The above method possesses strong geospatial attributes and offers excellent visualization and interactive performance.

[0084] In the sediment feature results, some small patches inevitably appear in the sediment type classification results. This application removes or reclassifies these small patches, and uses a method similar to convolution filtering to assign false pixels in larger categories to the classification results of sediment types. The specific steps are as follows: First, define a transformation kernel size, and replace the category of the central pixel with the category of the dominant pixel (the one with the most pixels) in the transformation kernel to obtain the sediment type classification result 1.

[0085] To further improve the accuracy of sediment classification results, this application can also employ clustering methods as an alternative. Clustering uses morphological operators to cluster and merge neighboring similar classification regions. Classification images often lack spatial continuity (the presence of spots or holes in the classification regions). While low-pass filtering can be used to smooth these images, the category information is often interfered with by the encoding of neighboring categories; clustering can solve this problem. First, the selected categories are merged together using an enlargement operation, and then an erosion operation is performed on the classification image using a transformation kernel of a specified size.

[0086] To further improve the accuracy of sediment type classification, this application can also employ a filtering method as an alternative. Filtering addresses the issue of isolated pixels in the classification image. It uses a speckle grouping method to eliminate these isolated pixels. The category selection method analyzes four or eight surrounding pixels to determine if a pixel belongs to the same group. If the number of analyzed pixels in a class is less than the input threshold, these pixels are removed from that class and reclassified as unclassified pixels. Both parallel schemes can effectively remove a small number of misclassified rasters and reclassify them correctly, thus achieving higher accuracy results.

[0087] The classification result 1 of sediment types obtained through the above steps has higher classification accuracy than the previous sediment type classification result, and removes the influence of a small amount of raster information being misclassified, resulting in a smoother sediment type classification result.

[0088] Furthermore, the probability value of each cell in the sediment type classification results can be obtained. These probability values ​​are then reclassified based on different thresholds to obtain sediment type classification result 2. Specifically, ensure you have a probability map for each cell corresponding to each sediment type. These probability maps typically have the same spatial resolution and coordinate system as the original classification result map. Then determine how to reclassify the cells based on the probability values. For example, set a fixed probability threshold (e.g., 0.5) for each sediment type; a cell is only classified as a certain type if its probability exceeds this threshold. Alternatively, consider the probability distribution of all sediment types in each cell and select the type with the highest probability as the classification result, or set a relative threshold (e.g., the difference between the highest and second-highest probabilities). Or, set different thresholds for different sediment types to reflect the confidence requirements of different classifications. Finally, iterate through the probability values ​​of each cell and reassign sediment type labels according to the selected thresholds to obtain classification result 2. Compared to the classification results of sediment types and sediment type classification result 1, sediment type classification result 2 can be applied to different production needs and achieve the highest accuracy requirements.

[0089] In the sediment feature predictions obtained using the random forest algorithm, each pixel in the experimental study area of ​​this application has a value, thus each pixel is equivalent to a sediment sampling station. This allows for the acquisition of sediment feature information that is (number of pixels in the experimental study area / number of actual sampling stations) times more detailed than that of actual sediment sampling stations. Furthermore, this avoids the problem of discrete sediment sampling stations resulting in insufficient coverage of the study area. Overly sparse distribution of sediment sampling locations leads to low resolution, poor representativeness, excessive reliance on the subjectivity of investigators, and the analysis of the entire study area with a limited number of sediment samples, which may deviate from reality and cause cumulative errors in later studies.

[0090] The above method can change the conventional field sediment sampling method, greatly save costs, reduce the danger of artificial sediment sampling in the field, and obtain sediment feature information with higher resolution and more detailed information than the original sediment sampling.

[0091] In one embodiment, such as Figure 4 As shown, the drone data includes digital terrain models, slope data, and aspect data;

[0092] The steps for acquiring drone data of the coastal zone include:

[0093] S410, collects two-dimensional image data of the coastal zone;

[0094] S420 processes two-dimensional image data to obtain a digital land surface model;

[0095] S430 calculates slope and aspect data based on the digital terrain model.

[0096] Specifically, a drone equipped with a high-resolution camera flies along a predetermined route to acquire true-color JPG images covering the target area. These images should have as much overlap as possible to ensure the accuracy of subsequent processing. Generally, the overlap between adjacent images should be between 60% and 80%. Multiple images are then stitched together and corrected to generate a high-precision orthophoto and digital terrain model. Slope indicates the degree of inclination of the terrain surface relative to the horizontal plane, usually expressed as a percentage or degrees. Aspect indicates the direction in which the terrain surface faces, usually expressed in degrees (0° is north, increasing clockwise). The digital terrain model can be processed using GIS software to obtain slope and aspect data.

[0097] Furthermore, all UAVs require compass and inertial navigation system calibration, continuous RTK network access during flight, and data collection under a fixed solution state. A fixed solution represents the final solution obtained by receiving differential data from a base station and calculating it using carrier phase differential data; it has high accuracy, generally within 1-3 cm. Reaching a fixed solution state ensures stable aircraft attitude and accurate position. 2) All flights use the WGS-84 coordinate system to ensure data has a unified coordinate system. 3) LiDAR and multispectral sensors are calibrated using ground calibration plates. 4) All measurements are set to the same flight altitude, preferably 30m-80m, with a forward overlap of no less than 80%, preferably 80%-85%, and a lateral overlap of no less than 70%, preferably 70%-80%. 5) Obvious markers are placed in the experimental research area as ground control points to calibrate and check the accuracy of UAV data. These ground control point markers are made of wooden boards no smaller than 40cm x 40cm, with clearly visible cross-shaped reflective strips. In areas with irregular coastal shapes, the number of flight paths should be increased as much as possible at the edges and corners, for example, by setting the paths in a crisscross pattern. Where there are man-made breakwaters, flight paths should be set perpendicular to the breakwaters whenever possible to avoid data loss caused by the breakwaters blocking light or creating large areas of shadow. In the case of crisscrossing tidal channels, flight paths should be increased as much as possible at the edges of the tidal channels to reduce splicing anomalies during subsequent data processing.

[0098] In one embodiment, such as Figure 5 As shown, the drone data includes multispectral data and normalized vegetation data;

[0099] The steps for acquiring drone data of the coastal zone include:

[0100] S510 collects true-color images, blue band data, green band data, red band data, red edge band data, and near-infrared band data of the coastal zone; among which, the true-color images, blue band data, green band data, red band data, red edge band data, and near-infrared band data include calibration plates placed on the coastal zone;

[0101] S520, acquires physical information of a calibration plate placed on the coastline;

[0102] S530, based on physical information, corrects blue band data, green band data, red band data, red edge band data and near-infrared band data to obtain corrected data;

[0103] S540 performs atmospheric correction on the calibration data to obtain multispectral data;

[0104] S550, based on red band data and near-infrared band data, yields normalized vegetation data.

[0105] Specifically, during the acquisition of true-color images, blue band data, green band data, red band data, red-edge band data, and near-infrared band data, the drone should fly when the solar altitude angle is higher than 30 degrees. The calibration plate is placed on the coastline; therefore, the true-color images, blue band data, green band data, red band data, red-edge band data, and near-infrared band data all contain calibration plate image information. When acquiring radiometric calibration plate data, avoid having shadows covering the reflector. Multiple plates can be photographed together or separately. If the weather changes drastically, it is recommended to place the plates together and photograph them simultaneously. The image quality depends on the distance between the camera and the calibration plate, as well as the side length of the calibration plate. It is recommended that the optimal camera distance be 7-8 times the side length of the calibration plate. The raw data collected in the second measurement was in sets; that is, each shot taken during the flight was recorded as a set. Each set of data included one true-color image and five TIF files (blue band data, green band data, red band data, red-edge band data, and near-infrared band data). The five TIF files, in order, are: blue band (450nm±16nm), green band (560nm±16nm), red band (650nm±16nm), red-edge band (730nm±16nm), and near-infrared band (840nm±16nm).

[0106] The reflectance of the calibration plate is known and used to correct for radiometric inhomogeneities. The known reflectance values ​​of the calibration plate are input into the Pix 4D Mapper software. These values ​​are used to adjust the radiance values ​​of the image to reflect the reflectance characteristics of actual ground features. Using the software's radiometric correction function, based on the reflectance values ​​of the calibration plate (i.e., the aforementioned physical information), the radiometric inhomogeneities of the image are corrected to obtain corrected data. Further, using an atmospheric correction model, with flight altitude and weather conditions input, atmospheric correction is performed on the image to remove the influence of the atmosphere, resulting in the aforementioned multispectral data. The normalized difference in vegetation index (NDVI) is equal to the difference between near-infrared and red light reflectance divided by the sum of near-infrared and red light reflectance.

[0107] In one embodiment, the drone data also includes point cloud intensity information; wherein the point cloud intensity information is obtained by a laser scanning device located on the flight device. The laser scanning device can be a LiDAR sensor.

[0108] In one embodiment, the step of establishing a quantitative relationship between sediment features and UAV data based on a dataset and a sample set of sediment features includes:

[0109] A quantitative relationship was obtained by performing multiple linear regression fitting on the dataset and the sample set of sediment characteristics.

[0110] The data for the 10 variables were normalized to eliminate the influence of different dimensions on the model. Since sediment type data is discrete, this invention uses continuous data (water content, average grain size, sorting coefficient, skewness) as an example and performs a multiple linear regression fitting. This invention treats the data for the 10 variables as independent entities for multiple linear regression fitting.

[0111] The fitting formula for the multiple linear regression method is as follows:

[0112]

[0113] In the formula, y1, y2, y3, and y4 represent moisture content, average particle size, sorting coefficient, and skewness, respectively; x1, x2, x3, x4, x5, x6, x7, x8, x9, and x 10 These are digital land surface model, slope, aspect, blue band, green band, red band, red edge band, near-infrared band, normalized vegetation index, and point cloud intensity.

[0114] In one embodiment, the step further includes:

[0115] The Gao-Collins grain size trend analysis method was used to determine the net transport pattern of sediments based on sediment characteristics.

[0116] Specifically, the first step is to compare each adjacent sampling point on the sampled grid to find all particle size trend vectors. Whether two sampling points are adjacent can be measured by the feature distance Lcr (Lcr is usually the maximum sampling interval). If the actual distance between two sampling points is less than Lcr, they are considered adjacent; otherwise, they are considered not adjacent.

[0117] The second step is to calculate the sum of the trend vectors for each sampling point using the following formula:

[0118]

[0119] In the formula, n is the total number of trend vectors of the sampled points considered, and d(x,y) i Let D(x,y) be the trend vector, and let D(x,y) be the sum of the individual trend vectors.

[0120] The third step is to smooth the composite vector D(x,y), the purpose of which is to eliminate the "noise" (i.e., high-frequency variations in D(x,y) in the image) of D(x,y). The mathematical transformation for smoothing is as follows:

[0121]

[0122] In the formula, D j The resultant vector is obtained from equation (7), where k is the total number of adjacent sampling points (whether they are adjacent or not is still determined by the feature distance L). cr To determine), D m (x,y) is the smoothed trend vector. D m The planar distribution image of (x,y) represents the net transport pattern of sediments.

[0123] In one embodiment, a coastal sediment characteristic monitoring device is provided, comprising:

[0124] The acquisition module is used to acquire drone data from the coastal zone;

[0125] The measurement module is used to measure the characteristic parameters of sediment samples from various sampling points on the coastal zone; the characteristic parameters include continuous data and discrete data.

[0126] The data conversion module is used to convert discrete data into digital data using a preset naming rule for digital formats.

[0127] The raster processing module is used to rasterize the continuous and digital data of each sampling point to obtain a raster image; wherein the resolution of the raster image is the same as the sampling diameter of the sampling point.

[0128] The dataset construction module is used to construct a sample set of sediment features based on raster images and to construct a set dataset based on UAV data.

[0129] The quantitative relationship confirmation module is used to establish a quantitative relationship between sediment characteristics and UAV data based on the ensemble dataset and a sample set of sediment characteristics.

[0130] The prediction module is used to predict sediment characteristics of the area to be tested using quantitative relationships.

[0131] Specific limitations regarding the coastal sediment characteristic monitoring device can be found in the limitations of the coastal sediment characteristic monitoring method described above, and will not be repeated here. Each module in the aforementioned coastal sediment characteristic monitoring device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. It should be noted that the module division in this embodiment is illustrative and only represents a logical functional division; other division methods may be used in actual implementation.

[0132] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0133] Acquire drone data of the coastal zone;

[0134] The characteristic parameters of sediment samples from various sampling points along the coastal zone were measured; the characteristic parameters included both continuous and discrete data.

[0135] Discrete data is converted into numerical data using a pre-defined naming rule for numerical formats.

[0136] The continuous and digital data at each sampling point are rasterized to obtain a raster image; the resolution of the raster image is the same as the sampling diameter of the sampling point.

[0137] A sample set of sediment features is constructed based on raster images, and an aggregate dataset is constructed based on UAV data;

[0138] Based on the ensemble dataset and the sample set of sediment characteristics, a quantitative relationship between sediment characteristics and UAV data is established;

[0139] Quantitative relationships are used to predict sediment characteristics in the area to be tested.

[0140] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, performs the following steps:

[0141] Acquire drone data of the coastal zone;

[0142] The characteristic parameters of sediment samples from various sampling points along the coastal zone were measured; the characteristic parameters included both continuous and discrete data.

[0143] Discrete data is converted into numerical data using a pre-defined naming rule for numerical formats.

[0144] The continuous and digital data at each sampling point are rasterized to obtain a raster image; the resolution of the raster image is the same as the sampling diameter of the sampling point.

[0145] A sample set of sediment features is constructed based on raster images, and an aggregate dataset is constructed based on UAV data;

[0146] Based on the ensemble dataset and the sample set of sediment characteristics, a quantitative relationship between sediment characteristics and UAV data is established;

[0147] Quantitative relationships are used to predict sediment characteristics in the area to be tested.

[0148] In specific implementation, the embodiments of this application can be referred to the above embodiments and have corresponding technical effects.

[0149] It is understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described herein, or combinations thereof. For software implementation, the techniques described herein can be implemented by units that perform the functions described herein. Software code can be stored in memory and executed by a processor. The memory can be implemented in the processor or external to the processor.

[0150] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0151] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0152] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for monitoring the characteristics of coastal sediments, characterized in that, include: Acquiring drone data of the coastal zone; wherein the drone data includes a digital land model, slope data, and aspect data; the steps for acquiring drone data of the coastal zone include: acquiring two-dimensional image data of the coastal zone; processing the two-dimensional image data to obtain the digital land model; calculating the slope data and aspect data based on the digital land model; wherein the drone data also includes multispectral data and normalized vegetation data; the steps for acquiring drone data of the coastal zone include: acquiring true-color images, blue band data, green band data, red band data, red-edge band data, and near-infrared data of the coastal zone. The data includes band data; wherein the true-color image, blue band data, green band data, red band data, red-edge band data, and near-infrared band data include a calibration plate placed on the coastal zone; the physical information of the calibration plate placed on the coastal zone is acquired; based on the physical information, the blue band data, green band data, red band data, red-edge band data, and near-infrared band data are corrected to obtain corrected data; atmospheric correction is performed on the corrected data to obtain the multispectral data; and the normalized vegetation data is obtained based on the red band data and near-infrared band data. The characteristic parameters of sediment samples from various sampling points along the coastal zone were measured; these characteristic parameters included both continuous and discrete data; and included water content, average grain size, sorting factor, and skewness. The discrete data is converted into digital data using a pre-defined naming rule for a digital format. The continuous data and digital data of each sampling point are rasterized to obtain a raster image; wherein the resolution of the raster image is the same as the sampling diameter of the sampling point. A sample set of sediment features is constructed based on the raster images, and a dataset is constructed based on the UAV data; Based on the dataset and the sample set of sediment features, establish a quantitative relationship between the sediment features and the UAV data; Using the quantitative relationship, the sediment characteristics of the area to be tested are predicted.

2. The method for monitoring coastal sediment characteristics according to claim 1, characterized in that, The steps for establishing a quantitative relationship between sediment features and UAV data based on the dataset and the sample set of sediment features include: An initial random forest model is trained using the dataset and the sample set of sediment features to obtain a trained random forest model; wherein the dataset and the sample set of sediment features have the same number of rows and columns. Based on the trained random forest model, the quantitative relationship between the sediment characteristics and the UAV data is determined.

3. The method for monitoring coastal sediment characteristics according to claim 1, characterized in that, The UAV data also includes point cloud intensity information; wherein, the point cloud intensity information is obtained by detection using a laser scanning device installed on the flight device.

4. The method for monitoring coastal sediment characteristics according to claim 1, characterized in that, The steps for establishing a quantitative relationship between sediment features and UAV data based on the dataset and the sample set of sediment features include: The quantitative relationship is obtained by performing multiple linear regression fitting on the dataset and the sample set of sediment characteristics.

5. The method for monitoring coastal sediment characteristics according to claim 1, characterized in that, It also includes the following steps: The net transport pattern of sediments was determined based on the sediment characteristics using the Gao-Collins grain size trend analysis method.

6. A monitoring device for coastal sediment characteristics, characterized in that, include: The acquisition module is used to acquire drone data of the coastal zone; wherein the drone data includes a digital surface model, slope data, and aspect data; it is also used to acquire two-dimensional image data of the coastal zone; process the two-dimensional image data to obtain the digital surface model; and calculate the slope data and aspect data based on the digital surface model; wherein the drone data also includes multispectral data and normalized vegetation data; it is also used to acquire true-color images, blue band data, green band data, red band data, red-edge band data, and near-infrared band data of the coastal zone; wherein the true... The color image, the blue band data, the green band data, the red band data, the red-edge band data, and the near-infrared band data include a calibration plate placed on the coastal zone; the physical information of the calibration plate placed on the coastal zone is acquired; based on the physical information, the blue band data, the green band data, the red band data, the red-edge band data, and the near-infrared band data are corrected to obtain corrected data; atmospheric correction is performed on the corrected data to obtain the multispectral data; based on the red band data and the near-infrared band data, the normalized vegetation data is obtained; The measurement module is used to measure characteristic parameters of sediment samples from various sampling points on the coastal zone; the characteristic parameters include continuous data and discrete data; the characteristic parameters include water content, average grain size, sorting factor, and skewness. The data conversion module is used to convert the discrete data into digital data using a preset digital format naming rule; A raster processing module is used to rasterize the continuous data and digital data of each sampling point to obtain a raster image; wherein the resolution value of the raster image is the same as the sampling diameter value of the sampling point; A dataset construction module is used to construct a sample set of sediment features based on the raster image, and to construct a set dataset based on the UAV data; The quantitative relationship confirmation module is used to establish a quantitative relationship between the sediment features and the UAV data based on the dataset and the sample set of sediment features. The prediction module is used to predict the sediment characteristics of the area to be tested using the quantitative relationship.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images