Method and system for monitoring water area of small reservoir

Abstract

The invention discloses a small reservoir water area monitoring method and system. The method comprises the following steps: acquiring water area data images from different sources; after wave band synthesis is carried out on the water area data images, an ISODATA classification algorithm is utilized to classify the water area data images, and a water area extraction model is established; dividing the water area data image into water body, building and ground raster data by using the water area extraction model, converting the water body raster data into vector surface data, and calculating the area of the vector surface data to obtain the water area of the small reservoir; the water area extraction model is optimized by constructing a verification data set and evaluating the water body extraction precision by using a confusion matrix. Through ISODATA classification algorithm optimization, accurate monitoring of the water area of the small reservoir is achieved, the accuracy and efficiency of water extraction are improved, and the limitation of a traditional monitoring means in large-area and remote small reservoir monitoring is overcome.

Description

Technical Field

[0001] The invention relates to the field of water conservancy science and technology, and in particular to a method and system for monitoring the water area of ​​a small reservoir. Background Art

[0002] As an important water conservancy facility in human society, reservoirs play an important role in water source regulation and distribution, flood control, irrigation, regional water supply, etc. Reservoir area is a characteristic parameter of reservoirs and an important basis for hydrological departments to judge the status of reservoirs. Changes in water area will not only affect land use types and local climate changes, but may also cause environmental problems such as soil structure changes, soil erosion, and pollution spread.

[0003] Traditionally, the monitoring of reservoir water area is mainly carried out by setting up hydrological monitoring stations and using professional instruments to measure water volume, water environment and other parameters. The measurement results are relatively accurate, but there are disadvantages such as complex monitoring process in large areas, unstable information transmission, and large equipment loss, which makes it difficult to meet the requirements for rapid acquisition of reservoir parameters. In addition, many small reservoirs are scattered and remote, with inconvenient transportation, resulting in inadequate management, low social attention, and a single source of monitoring data. It is also difficult to achieve the purpose by manual monitoring. In recent years, the emergence of satellite remote sensing has brought new methods for monitoring reservoir water area. Remote sensing monitoring has the advantages of instantaneous imaging, low cost, convenient transmission, and all-day and all-weather. It can also reach places that cannot be reached by humans, greatly improving monitoring efficiency. However, it also faces new challenges. Satellite remote sensing data alone cannot improve the monitoring accuracy very well. Effective extraction of water bodies is the key to improving monitoring accuracy. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for monitoring the water area of ​​a small reservoir to solve the problems of the complexity, high cost and insufficient monitoring accuracy caused by the single data source of the traditional water area monitoring method.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for monitoring the water area of ​​a small reservoir, comprising:

[0008] Obtain water data images from different sources;

[0009] After band synthesis of the water area data image, the water area data image is classified using the ISODATA classification algorithm to establish a water area extraction model;

[0010] The water area extraction model is used to divide the water area data image into water body, building and ground raster data, the water body raster data is converted into vector surface data, the area of ​​the vector surface data is calculated, and the water area of ​​the small reservoir is obtained;

[0011] The water body extraction model was optimized by constructing a validation dataset and using the confusion matrix to evaluate the water body extraction accuracy.

[0012] As a preferred solution of the method for monitoring the water area of ​​a small reservoir described in the present invention, wherein: performing band synthesis on the water area data image includes:

[0013] Multiple water area data image bands with specific physical characteristics are selected, and water area data images of different bands are fused into a new multi-band image according to rules.

[0014] As a preferred solution of the method for monitoring the water area of ​​a small reservoir described in the present invention, the water area data image is classified using the ISODATA classification algorithm, including:

[0015] Preset an initial clustering center K, classify the pixel points of the water area data image into the corresponding initial clustering center according to the spectral characteristics, and obtain an initial clustering result;

[0016] According to the initial clustering result, the cluster center value is updated, which is expressed as:

[0017]

[0018] Where M is the spectral feature, c im represents the m-th dimension feature of the i-th category, x i Indicates that it belongs to category S i Pixels of

[0019] The average distance between cluster centers and the total average distance from all pixels to the cluster centers are calculated, and the initial cluster centers are split and merged according to the average distance and the total average distance.

[0020] As a preferred solution of the method for monitoring the water area of ​​a small reservoir described in the present invention, the splitting and merging of the initial cluster centers includes:

[0021] If the current cluster center value k <= K / 2, the current cluster center is split;

[0022] If the current cluster center value k>=2K, the current cluster center is merged;

[0023] If K / 2 < k < 2K, further judgment is made according to the number of iterations of the ISODATA classification algorithm. When the number of iterations is odd, a splitting operation is performed, and when the number of iterations is even, a merging operation is performed.

[0024] As a preferred solution of the small reservoir water area monitoring method described in the present invention, wherein: the splitting operation includes,

[0025] Calculate the standard deviation of the pixel points in the spectral bands of all clusters in the water area data image, and obtain a standard deviation vector, expressed as:

[0026]

[0027] wherein, X lm is the m-th spectral feature of the I-th pixel point in the cluster S i , C im is the mean value of the m-th spectral feature in the cluster S i , σ im is the standard deviation of the m-th spectral feature of the cluster S i , m is the dimension of the spectral feature, and N i is the pixel point belonging to the cluster category.

[0028] According to the standard deviation vector, obtain the maximum component in each standard deviation vector to obtain a component set, and judge whether the splitting calculation is completed for the component set.

[0029] As a preferred solution of the small reservoir water area monitoring method described in the present invention, wherein: the merging operation includes,

[0030] Calculate the mutual distance between all cluster centers in the water area data image, expressed as,

[0031] D ij =|z i -z j |, i = 1, 2,..., N c -1, j = i + 1,..., N c

[0032] wherein, D ij is the mutual distance, z i is the i-th cluster center, z j is the j-th cluster center, and N c is the total number of cluster centers;

[0033] Compare the mutual distance D ij with the set minimum distance θ c , and compare D ij < θ cThe values ​​of are arranged, and the cluster centers corresponding to the mutual distances are merged. The new cluster center is the weighted average of the two original cluster centers.

[0034] If the distances between them are the same, the two cluster centers are merged, and the new cluster center is the weighted average of the two original cluster centers.

[0035] As a preferred solution of the method for monitoring the water area of ​​a small reservoir described in the present invention, wherein: constructing a validation data set and using a confusion matrix to evaluate the water body extraction accuracy includes:

[0036] Verification points are randomly generated within the management scope of the reservoir, and the types of features at the randomly generated verification points are interpreted using visual interpretation methods to calculate the confusion matrix;

[0037] The overall classification accuracy and Kappa coefficient generated by the confusion matrix were used to evaluate the extraction results of the water area extraction model in different images and different types of small reservoirs.

[0038] In a second aspect, the present invention provides a system for monitoring the water area of ​​a small reservoir, comprising:

[0039] A data acquisition module is used to acquire water area data images from different sources;

[0040] An image classification and extraction module is used to classify the water area data image using the ISODATA classification algorithm after band synthesis, and to establish a water area extraction model;

[0041] An area acquisition module is used to use a water area extraction model to divide the water area data image into water body, building and ground raster data, convert the water body raster data into vector surface data, calculate the area of ​​the vector surface data, and obtain the water area of ​​the small reservoir;

[0042] The optimization module is used to optimize the water body extraction model by constructing a validation dataset and evaluating the water body extraction accuracy using a confusion matrix.

[0043] In a third aspect, the present invention provides a computing device, comprising:

[0044] Memory and processor;

[0045] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the small reservoir water area monitoring method are implemented.

[0046] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the small reservoir water area monitoring method.

[0047] Compared with the prior art, the present invention has the following beneficial effects: the present invention realizes accurate monitoring of the water area of ​​small reservoirs through ISODATA classification algorithm optimization. The present invention not only improves the accuracy and efficiency of water extraction, but also effectively overcomes the limitations of traditional monitoring methods in monitoring large areas and remote small reservoirs by dynamically adjusting cluster centers, optimizing classification strategies and constructing verification data sets; at the same time, the use of confusion matrix for accuracy evaluation provides a scientific basis for the continuous optimization of the water area extraction model, ensures the reliability and practicability of the monitoring results, and provides strong technical support for the rational scheduling of water resources and the effective management of small reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0049] Figure 1 A schematic diagram of the overall process of a method for monitoring the water area of ​​a small reservoir according to an embodiment of the present invention;

[0050] Figure 2 A schematic diagram of a water body extraction result of a PlanetScope image of a small reservoir water area monitoring method according to an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of water body extraction results from Sentinel-2 images of a small reservoir water area monitoring method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0055] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0056] At the same time, in the description of the present invention, it should be noted that the orientations or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the system or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0057] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] Example 1

[0059] Reference Figure 1-3 , is an embodiment of the present invention, and provides a method for monitoring the water area of ​​a small reservoir, comprising:

[0060] S100: Acquire water area data images from different sources;

[0061] Specifically, when acquiring water data images from different sources, a variety of satellite remote sensing image data can be included, including but not limited to PlanetScope images, Sentinel-2 images, Landsat series images, and MODIS, etc. These images have their own advantages in monitoring water body dynamics and extracting water area due to their different spatial resolutions, spectral band coverage, and revisit cycles;

[0062] In the embodiment of the present application, PlanetScope images and Sentinel-2 images are selected as data sources. These two images, with their high resolution and rich spectral information, can more accurately identify and extract water areas, and further confusion matrix accuracy evaluation can ensure the accuracy and reliability of water area extraction results;

[0063] Data preprocessing is performed on PlanetScope and Sentinel-2 images. Preprocessing mainly includes geometric correction, image stitching and cropping. The essence of geometric connection is geometric correction. Through this method, different images can be corrected to a unified coordinate system. The images involved in mosaicking can be acquired at different times or by different sensors, but there must be overlap between different images. Through mosaicking, a larger range of ground images can be obtained. The key to mosaicking images is geometric connection and correction of chromatic aberration.

[0064] S102: After band synthesis of the water area data image, the water area data image is classified using the ISODATA classification algorithm to establish a water area extraction model;

[0065] Preferably, multiple water area data image bands with specific physical characteristics are selected, and water area data images of different bands are fused into a new multi-band image according to a rule;

[0066] Specifically, blue light bands were selected from PlanetScope images and Sentinel-2 images. These bands are particularly sensitive to the reflection of water bodies and can clearly outline the contours of water bodies. These bands were directly mapped to the blue channel of the RGB color space, making the water bodies appear bright blue in the images, which is convenient for subsequent visual interpretation;

[0067] Preferably, an initial cluster center K is preset, and the pixel points of the water area data image are classified into the initial cluster center according to the spectral characteristics to obtain the initial clustering result; according to the initial clustering result, the cluster center value is updated, which is expressed as:

[0068]

[0069] Where M is the spectral feature, c im represents the m-th dimension feature of the i-th category, x i Indicates that it belongs to category S i Pixels of

[0070] In the embodiment of the present application, the initial cluster center K is set to 500;

[0071] Preferably, calculate the average distance between the cluster centers and the total average distance from all pixel points to the cluster centers, and split and merge the initial cluster centers according to the average distance and the total average distance;

[0072] Preferably, splitting and merging the initial cluster centers includes: if the current number of cluster centers k <= K / 2, perform a splitting operation on the current cluster centers; if the current number of cluster centers k >= 2K, perform a merging operation on the current cluster centers; if K / 2 < k < 2K, make a further judgment according to the number of iterations of the ISODATA classification algorithm. When the number of iterations is odd, perform a splitting operation, and when the number of iterations is even, perform a merging operation;

[0073] The splitting operation includes calculating the standard deviation of the pixel points in all clusters of the water area data image in the spectral band, and obtaining a standard deviation vector, expressed as:

[0074]

[0075] where, X lm is the m-th spectral feature of the i-th pixel point in the cluster S i , C im is the mean of the m-th spectral feature in the cluster S i , σ im is the standard deviation of the m-th spectral feature of the cluster S i , m is the dimension of the spectral feature, N i is the pixel point belonging to the cluster category.

[0076] Preferably, according to the standard deviation vector, obtain the maximum component in each standard deviation vector to obtain a component set, and judge whether the splitting calculation is completed for the component set;

[0077] The merging operation includes calculating the mutual distance between all cluster centers in the water area data image, expressed as,

[0078] D ij =|z i -z j |, i = 1, 2,..., N c -1, j = i + 1,..., N c

[0079] where, D ij is the mutual distance, z i is the i-th cluster center, z j is the j-th cluster center, N c is the total number of cluster centers;

[0080] Preferably, compare the mutual distance D ij with the set minimum distance θ cFor comparison, D ij <θ c The cluster centers corresponding to the mutual distances are arranged, and the new cluster center is the weighted average of the two original cluster centers. If the mutual distances are the same, the two cluster centers are merged, and the new cluster center is the weighted average of the two original cluster centers.

[0081] It should be noted that by selecting the blue light bands in PlanetScope images and Sentinel-2 images, the recognition of water bodies in images is significantly enhanced, making the outlines of water bodies clearly presented. By using the ISODATA classification algorithm, presetting and dynamically adjusting the initial cluster centers, and combining splitting and merging strategies, the complex classification problems in water data images are effectively solved. This process not only improves the accuracy and stability of classification, but also simplifies the water extraction model by optimizing the number of cluster centers, making the model more efficient and accurate in extracting water information.

[0082] S104: using a water area extraction model to divide the water area data image into water body, building and ground raster data, converting the water body raster data into vector surface data, calculating the area of ​​the vector surface data, and obtaining the water area of ​​the small reservoir;

[0083] In the embodiment of the present application, different data source images will have different extraction effects when using different water body extraction methods to extract the water area of ​​small reservoirs of different sizes and shapes. i and different area sizes, S i The calculation method is as follows:

[0084]

[0085] Among them, S i is the shape index, A is the water area of ​​the small reservoir, and P is the perimeter of the small reservoir;

[0086] In the present embodiment, S i It is the ratio of the area of ​​a small reservoir to the square of its perimeter. Its value is between 0 and 1. When the value is 1, it means that the small reservoir is a circular reservoir, and the closer the value is to 1, the more regular the small reservoir is.

[0087] The results of water area extraction from PlanetScope image and Sentinel-2 image using the method of the present invention are as follows: Figure 2 Figure 3As shown in the figure, by comparing the results obtained by the original true color image and other water body extraction methods, it can be found that the ISODATA classification method can extract the main water body of the reservoir, but the details are slightly different; in the five types of small reservoirs, the narrow and small water bodies can be effectively classified, and the water and land can be effectively separated at the junction of water and land;

[0088] When using the NDWI water body index method to extract water bodies, it is easy to mistakenly classify small water bodies in reservoirs and water bodies at the junction of water and land as non-water bodies, resulting in a smaller water area extraction result. When the water area is less than 15ha, the water body extraction inside the water body will be discontinuous, showing a "spotted" phenomenon.

[0089] S106: Optimizing the water extraction model by constructing a validation data set and evaluating the water extraction accuracy using a confusion matrix;

[0090] Verification points are randomly generated within the management scope of the reservoir, and the types of features at the randomly generated verification points are interpreted using visual interpretation methods to calculate the confusion matrix;

[0091] The overall classification accuracy and Kappa coefficient generated by the confusion matrix were used to evaluate the extraction results of the water area extraction model in different images and different types of small reservoirs.

[0092] Specifically, the historical images in Google Earth software were used to draw the validation data set through manual visual interpretation. Due to the small amount of validation data, it cannot be guaranteed that the observation date of the validation data set is consistent with the observation date of each remote sensing image. Therefore, the remote sensing image data three days before and after the observation date of the validation data set were used for pairwise comparison. According to the validation data set, small reservoirs of different sizes and shapes were classified and the root mean square error (RMSE), variance, mean absolute percentage error (MAPE) and percentage error (Perecent Error) were calculated. The calculation formula is:

[0093]

[0094] Among them, x i is the water area of ​​the small reservoir extracted from the validation dataset, y i It is the water area of ​​small reservoirs extracted from different remote sensing images.

[0095] The above is a schematic scheme of a method for monitoring the water area of ​​a small reservoir in this embodiment. It should be noted that the technical scheme of the system for monitoring the water area of ​​a small reservoir and the technical scheme of the method for monitoring the water area of ​​a small reservoir are of the same concept. For details not described in detail in the technical scheme of the system for monitoring the water area of ​​a small reservoir in this embodiment, please refer to the description of the technical scheme of the method for monitoring the water area of ​​a small reservoir.

[0096] The small reservoir water area monitoring system in this embodiment includes:

[0097] A data acquisition module is used to acquire water area data images from different sources;

[0098] The image classification and extraction module is used to classify the water area data image by using the ISODATA classification algorithm after band synthesis, and to establish a water area extraction model;

[0099] The area acquisition module is used to use the water area extraction model to divide the water area data image into water body, building and ground raster data, convert the water body raster data into vector surface data, calculate the area of ​​the vector surface data, and obtain the water area of ​​the small reservoir;

[0100] The optimization module is used to optimize the water body extraction model by constructing a validation dataset and evaluating the water body extraction accuracy using a confusion matrix.

[0101] This embodiment also provides a computing device suitable for monitoring the water area of ​​a small reservoir, including:

[0102] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method for monitoring the water area of ​​a small reservoir as proposed in the above embodiment.

[0103] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for monitoring the water area of ​​a small reservoir as proposed in the above embodiment is implemented.

[0104] The storage medium proposed in this embodiment and the method for realizing the monitoring of the water area of ​​a small reservoir proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0105] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0106] Example 2

[0107] Referring to Table 1-2, an embodiment of the present invention provides a method for monitoring the water area of ​​a small reservoir. In order to verify its beneficial effects, comparison results of several schemes are provided.

[0108] The small reservoirs in Yizheng City were selected as the research area. Yizheng City is located on the north bank of the lower reaches of the Yangtze River and the top of the Yangtze River Delta. Yizheng City is adjacent to the Yangtze River in the south, Yangzhou City's Hanjiang District in the east, Nanjing City's Liuhe District and Jiangning District in the west, and Gaoyou City and Tianchang City in Anhui Province in the north. The terrain is high in the northwest and low in the southeast. It is divided into hilly and mountainous areas and plain areas along the mountains and rivers. This paper mainly studies the monitoring of the water area of ​​small reservoirs.

[0109] Preprocess the acquired PlanetScope image and Sentinel-2 image data.

[0110] Based on PlanetScope and Sentinel-2 image data, water bodies of five types of small reservoirs were extracted using ISODATA classification method;

[0111] The five types are: area>15ha and 0.15<S i <1, area>15ha and 0<S i <0.15, area <15ha and 0.4<S i <1, area <15ha and 0.2<S i <0.4 and area <15ha and 0<S i <0.2;

[0112] The accuracy of the confusion matrix for water area extraction under PlanetScope image and Sentinel-2 image using the ISODATA classification method of the present invention is shown in Table 1;

[0113] Table 1 Confusion matrix accuracy evaluation results of water area extraction based on PlanetScope and Sentinel-2 images

[0114]

[0115]

[0116] As can be seen from Table 1, for PlanetScope images, when the water area is less than 15ha, the OA of the ISODATA classification method is 58.00%-68.48%, and the Kappa coefficient is only 0.3831-0.4697; by comparing the confusion matrix accuracy evaluation results of the same method and different reservoir types, it can be found that when the reservoir water area is greater than 15ha, its OA and Kappa coefficient are generally higher than those of the reservoir with a water area of ​​less than 15ha, indicating that under the condition of a certain sample point, the increase of the reservoir water area can increase the result of its confusion matrix accuracy evaluation to a certain extent.

[0117] For Sentinel-2 images, the comparison of the confusion matrix accuracy evaluation results for the same method and different reservoir types shows that when the water area of ​​the reservoir is greater than 15ha, its OA and Kappa coefficients are generally greater than when the water area of ​​the reservoir is less than 15ha, indicating that when the number of verification points is the same, the size of the water area of ​​the reservoir can affect the confusion matrix accuracy evaluation results to a certain extent. Compared with the accuracy evaluation results obtained by using different methods based on PlanetScope images, it can be found that the OA and Kappa coefficients obtained by different methods based on Sentinel-2 images are quite different, which shows that the spatial resolution of remote sensing images has a great influence on the extraction of water bodies and the classification of other landforms.

[0118] The root mean square error results of the water area of ​​small reservoirs obtained by comparing PlanetScope images and Sentinel-2 images with the validation data set using the ISODATA classification method of the present invention are shown in Table 2;

[0119] Table 2 Root mean square error results of small reservoir water area obtained by comparing different remote sensing images with the validation data set

[0120]

[0121] It can be seen from Table 2 that, when the water area of ​​small reservoirs is the same, the higher the shape index of small reservoirs, the smaller the root mean square error of the water area of ​​small reservoirs obtained by comparing different images with the verification data set, that is, the higher the accuracy; as the water area of ​​small reservoirs increases, the root mean square error of the water area of ​​small reservoirs obtained by comparing different remote sensing images with the verification data set also increases;

[0122] Therefore, compared with the existing methods, the present invention gives full play to the advantages of remote sensing technology and greatly improves the monitoring efficiency. The machine learning algorithm has the characteristics of fast computing speed, good stability, and high prediction accuracy. The introduction of machine learning methods to extract water bodies effectively improves the water body accuracy and provides a new method for reservoir area monitoring.

[0123] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for monitoring the water area of ​​a small reservoir, characterized in that: Including: Obtain water area data images from different sources; After performing band synthesis on the water area data images, use the ISODATA classification algorithm to classify the water area data images and establish a water area extraction model; Use the water area extraction model to classify the water area data images into water bodies, buildings, and ground grid data, convert the water body grid data into vector surface data, calculate the area of the vector surface data, and obtain the water area of the small reservoir; Optimize the water area extraction model by constructing a validation dataset and using a confusion matrix to evaluate the water body extraction accuracy.

2. The method for monitoring the water area of ​​a small reservoir according to claim 1, characterized in that: Performing band synthesis on the water area data images includes Select multiple bands of water area data images with specific physical characteristics, and fuse the water area data images of different bands into a new multi-band image according to rules.

3. The method for monitoring the water area of ​​a small reservoir according to claim 1 or 2, characterized in that: Using the ISODATA classification algorithm to classify the water area data images includes Preset an initial clustering center K, classify the pixel points of the water area data images according to spectral characteristics into the corresponding initial clustering centers, and obtain an initial clustering result; According to the initial clustering result, update the clustering center values, expressed as: Where M is the spectral feature, c im represents the m-th dimension feature of the i-th category, x i Indicates that it belongs to category S i Pixels of Calculate the average distance between clustering centers and the total average distance from all pixel points to the clustering centers, and split and merge the initial clustering centers according to the average distance and the total average distance.

4. The method for monitoring the water area of ​​a small reservoir according to claim 3, characterized in that: Splitting and merging the initial clustering centers includes If the current clustering center value k <= K / 2, perform a splitting operation on the current clustering center; If the current clustering center value k >= 2K, perform a merging operation on the current clustering center; If K / 2 < k < 2K, make a further judgment according to the number of iterations of the ISODATA classification algorithm. When the number of iterations is odd, perform a splitting operation, and when the number of iterations is even, perform a merging operation.

5. The method for monitoring the water area of ​​a small reservoir according to claim 4, characterized in that: The splitting operation includes Calculate the standard deviation of the pixel points in the spectral bands of all clusters in the water area data images to obtain a standard deviation vector, expressed as: Among them, X lm is the cluster S i The mth spectral feature of the Ith pixel in C im is the cluster S i The mean of the mth spectral feature in , σ im is the cluster S i The standard deviation of the mth spectral feature, m is the dimension of the spectral feature, N i are pixels belonging to the cluster category. According to the standard deviation vector, obtain the maximum component in each standard deviation vector to obtain a component set, and judge whether the splitting calculation is completed for the component set.

6. The method for monitoring the water area of ​​a small reservoir according to claim 4, characterized in that: The merging operation includes Calculate the mutual distances between all clustering centers in the water area data images, expressed as D ij =|z i -z j |,i=1,2,...,N c -1,j=i+1,...,N c Among them, D ij is the mutual distance, z i is the i-th cluster center, z j is the jth cluster center, N c is the total number of cluster centers; The mutual distance D ij and set the minimum distance θ c For comparison, D ij <θ c The values ​​of are arranged, and the cluster centers corresponding to the mutual distances are merged. The new cluster center is the weighted average of the two original cluster centers. If the mutual distances are the same, merge two clustering centers, and the new clustering center is the weighted average of the two original clustering centers.

7. The method for monitoring the water area of ​​a small reservoir according to claim 1, characterized in that: Evaluating the water body extraction accuracy by constructing a validation dataset and using a confusion matrix includes Randomly generate validation points within the management scope line of the reservoir, interpret the land cover types of the randomly generated validation points by visual interpretation, and calculate the confusion matrix; Use the overall classification accuracy and Kappa coefficient generated by the confusion matrix to evaluate the extraction results of the water area extraction model in different images and different types of small reservoirs.

8. A system for monitoring the water area of ​​a small reservoir, characterized in that: Including A data acquisition module for obtaining water area data images from different sources; An image classification and extraction module for performing band synthesis on the water area data images, using the ISODATA classification algorithm to classify the water area data images, and establishing a water area extraction model; An area acquisition module is used to use a water area extraction model to divide the water area data image into water body, building and ground raster data, convert the water body raster data into vector surface data, calculate the area of ​​the vector surface data, and obtain the water area of ​​the small reservoir; The optimization module is used to optimize the water body extraction model by constructing a validation dataset and evaluating the water body extraction accuracy using a confusion matrix.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the small reservoir water area monitoring method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for monitoring the water area of ​​a small reservoir as described in any one of claims 1 to 7.

Images