Mangrove wetland pollutant monitoring method and system based on edge computing

Abstract

The present invention relates to the field of wetland pollution monitoring technology, and specifically discloses a mangrove wetland pollutant monitoring method and system based on edge computing, the method comprising acquiring a wetland area, creating a monitoring point based on the wetland area; acquiring collected data containing a time tag based on the monitoring point, and creating a data graph containing a time tag; identifying the data graph, determining the polluted area and its pollution type, and updating the benchmark acquisition frequency of the monitoring point according to the polluted area and its pollution type; recording the time tag of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the benchmark acquisition frequency, and adjusting the recognition process of the data graph. After the present invention acquires the monitoring data of each sensor, it identifies the monitoring data, thereby reducing the data acquisition frequency of each sensor in a non-polluted area or an area with a low degree of pollution, improving the endurance of some sensors, and effectively reducing the inspection pressure.

Description

Technical Field

[0001] The present invention relates to the technical field of wetland pollution monitoring, and specifically to a mangrove wetland pollutant monitoring method and system based on edge computing. Background Art

[0002] Pollutant monitoring in mangrove wetlands is to protect mangrove ecosystems from pollution and ensure the maintenance of their ecological functions and biodiversity. As an important ecosystem, mangrove wetlands can provide a variety of ecological services, such as preventing coastal erosion, providing habitats, and supporting fishery resources.

[0003] Most of the existing mangrove pollutant monitoring processes use sensors with data transmission functions to obtain pollutant information and upload it to the main control terminal. In actual applications, the data acquisition frequency and data upload frequency of the sensor are preset, and regular upload is sufficient, which makes the battery life also fixed and needs to be replaced regularly. When the number of sensors is large, each replacement requires a large workload. How to improve the battery life of sensors during pollution monitoring is the technical problem that the technical solution of the present invention aims to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a mangrove wetland pollutant monitoring method and system based on edge computing to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for monitoring pollutants in mangrove wetlands based on edge computing, the method comprising:

[0007] Acquire a wetland area, and create a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed;

[0008] Acquire collected data with time tags based on monitoring points and create a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; and the initial data acquisition frequency of all monitoring points is the same;

[0009] Identify the data graph, determine the polluted area and its pollution type, and update the baseline acquisition frequency of the monitoring point according to the polluted area and its pollution type;

[0010] Record the time tags of the collected data fed back by each monitoring point, calculate the actual acquisition frequency, compare the actual acquisition frequency with the benchmark acquisition frequency, and adjust the recognition process of the data graph;

[0011] When acquiring collected data at the monitoring point, a frequency adjustment ratio is generated according to the change amount of the collected data; the frequency adjustment ratio is proportional to the change amount.

[0012] As a further solution of the present invention: the step of obtaining the wetland area and creating monitoring points based on the wetland area includes:

[0013] Obtain the boundary information of the wetland, create a wetland outline, and use the inner area of ​​the wetland outline as the wetland area;

[0014] Determine the innermost contour based on the total length of the wetland contour;

[0015] Create a preset number of equally spaced contour lines based on the innermost contour and the wetland contour, and calculate the distances between adjacent contour lines;

[0016] For each contour line, a monitoring point is created based on the distance.

[0017] As a further solution of the present invention: the step of acquiring the collected data with time tags based on the monitoring points and creating a data graph with time tags includes:

[0018] Creating a reference map based on the positional relationship of the monitoring points; each pixel in the reference map corresponds one-to-one to each monitoring point;

[0019] Receive the initial collected data uploaded by the monitoring point, input the initial collected data into a preset conversion model, and obtain the conversion value;

[0020] Statistically convert the values, determine the color values ​​of the corresponding pixels in the reference image, and obtain the initial data image; wherein the data types of the collected data at the same monitoring point can be up to three types;

[0021] When any monitoring point obtains new collected data with a time tag, the collected data is input into the conversion model to obtain a conversion value, and the color value of the corresponding pixel point in the initial data map is updated based on the conversion value as the data map corresponding to the time tag;

[0022] The data graph at the latest moment is used as the new initial data graph.

[0023] As a further solution of the present invention: the step of identifying the data graph, determining the pollution area and its pollution type, and updating the benchmark acquisition frequency of the monitoring point according to the pollution area and its pollution type includes:

[0024] Traverse the pixels in the data graph and calculate the difference of the pixels;

[0025] The data graph is divided into regions according to the difference of pixel points to obtain sub-regions;

[0026] Calculating the area, color value mean and color value standard deviation of the sub-region, comparing the color value mean and color value standard deviation with preset color value parameters, and matching to obtain the polluted area and its pollution type;

[0027] Update the baseline acquisition frequency of monitoring points according to the polluted area and its pollution type;

[0028] Among them, the mapping relationship between color value parameters and pollution types is obtained in advance by staff statistics.

[0029] As a further solution of the present invention: the rule for updating the benchmark acquisition frequency of the monitoring point according to the pollution area and its pollution type includes: In the formula, For monitoring point The reference acquisition frequency at It is The middle point of the pollution area The characteristic value of is determined by the pollution type. The higher the pollution level of the pollution type, The bigger; is the total number of polluted areas, and is the size of any contaminated area; As a further solution of the present invention: the steps of recording the time tags of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the reference acquisition frequency, and adjusting the recognition process of the data graph include:

[0030] Record the time tags of the collected data fed back by each monitoring point, calculate the difference of the time tags, and determine the actual acquisition frequency based on the difference;

[0031] Compare the actual acquisition frequency with the benchmark acquisition frequency, and reversely calculate the frequency adjustment ratio;

[0032] Adjust the color value parameters according to the frequency adjustment ratio;

[0033] The adjustment process of the color value parameter is proportional to the inverse of the frequency adjustment ratio.

[0034] The technical solution of the present invention also provides a mangrove wetland pollutant monitoring system based on edge computing, the system comprising:

[0035] A monitoring point creation module is used to obtain a wetland area and create a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed;

[0036] A data graph creation module is used to obtain collected data with time tags based on monitoring points and create a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; the initial data acquisition frequency of all monitoring points is the same;

[0037] A reference frequency determination module, used to identify the data graph, determine the pollution area and its pollution type, and update the reference acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0038] The recognition process adjustment module is used to record the time tags of the collected data fed back by each monitoring point, calculate the actual acquisition frequency, compare the actual acquisition frequency with the reference acquisition frequency, and adjust the recognition process of the data graph;

[0039] When acquiring collected data at the monitoring point, a frequency adjustment ratio is generated according to the change amount of the collected data; the frequency adjustment ratio is proportional to the change amount.

[0040] As a further solution of the present invention: the monitoring point creation module includes:

[0041] A contour creation unit is used to obtain boundary information of the wetland, create a wetland contour, and use the inner area of ​​the wetland contour as the wetland area;

[0042] An inner contour determining unit, used for determining the innermost contour according to the total length of the wetland contour;

[0043] An equidistant line construction unit, used for creating a preset number of equidistant contour lines based on the innermost contour and the wetland contour, and calculating the distances between adjacent contour lines;

[0044] The point selection unit is used to create a monitoring point for each contour line based on the distance.

[0045] As a further solution of the present invention: the data graph creation module includes:

[0046] A reference map creation unit, used to create a reference map according to the positional relationship of the monitoring points; each pixel point in the reference map corresponds to each monitoring point one by one;

[0047] A data conversion unit is used to receive the initial collected data uploaded by the monitoring point, input the initial collected data into a preset conversion model, and obtain a conversion value;

[0048] The numerical statistics unit is used to statistically convert numerical values, determine the color values ​​of corresponding pixels in the reference image, and obtain the initial data image; wherein, the data types of the collected data of the same monitoring point are at most three;

[0049] A data map updating unit, for when any monitoring point acquires new collected data with a time tag, inputs the collected data into the conversion model to obtain a conversion value, and based on the conversion value, updates the color value of the corresponding pixel point in the initial data map as the data map corresponding to the time tag;

[0050] The base updating unit is used to use the data graph at the latest moment as a new initial data graph.

[0051] As a further solution of the present invention: the reference frequency determination module includes:

[0052] A difference calculation unit, used for traversing the pixel points in the data graph and calculating the difference of the pixel points;

[0053] A region segmentation unit is used to segment the data image into regions according to the difference of pixel points to obtain sub-regions;

[0054] A pollution matching unit is used to calculate the area, color value mean and color value standard deviation of the sub-area, compare the color value mean and color value standard deviation with preset color value parameters, and match the pollution area and its pollution type;

[0055] A frequency updating unit, used to update the base acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0056] Among them, the mapping relationship between color value parameters and pollution types is obtained in advance by staff statistics.

[0057] Compared with the prior art, the beneficial effects of the present invention are: after obtaining the monitoring data of each sensor, the present invention identifies the monitoring data, thereby reducing the data collection frequency of each sensor in non-polluted areas or areas with low pollution levels, increasing the endurance of some sensors, and effectively reducing inspection pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0059] Figure 1 The flowchart of the mangrove wetland pollutant monitoring method based on edge computing.

[0060] Figure 2 This is the first sub-process flowchart of the mangrove wetland pollutant monitoring method based on edge computing.

[0061] Figure 3 This is the second sub-process flowchart of the mangrove wetland pollutant monitoring method based on edge computing.

[0062] Figure 4 This is the third sub-process block diagram of the mangrove wetland pollutant monitoring method based on edge computing.

[0063] Figure 5 This is the fourth sub-process flowchart of the mangrove wetland pollutant monitoring method based on edge computing.

[0064] Figure 6This is a structural block diagram of the mangrove wetland pollutant monitoring system based on edge computing. DETAILED DESCRIPTION

[0065] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0066] Figure 1 The flowchart of the mangrove wetland pollutant monitoring method based on edge computing is shown in FIG. In an embodiment of the present invention, a mangrove wetland pollutant monitoring method based on edge computing includes:

[0067] Step S100: obtaining a wetland area, and creating a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed;

[0068] The wetland area is a map corresponding to the mangrove wetland that needs to be monitored for pollutants. It is created by the boundary information input by the managers of the mangrove wetlands. After the wetland area is obtained, the monitoring points are created by the wetland area. The monitoring points are the points where the monitoring equipment is installed. Commonly used equipment includes heavy metal monitoring equipment, organic pollutant monitoring equipment and nutrient monitoring equipment. The heavy metal monitoring equipment mainly monitors lead, mercury and cadmium, etc. The organic pollutant monitoring equipment mainly monitors polycyclic aromatic hydrocarbons and pesticide residues, etc. The nutrient monitoring equipment mainly monitors nitrogen and phosphorus, etc. It should be noted that the monitoring equipment involved in this application generally does not exceed three types. This is because this application will use images to represent these data, and the pixel color values ​​in the image are generally represented by no more than three parameters.

[0069] Step S200: acquiring collected data with time tags based on the monitoring points, and creating a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; and the initial data acquisition frequency of all monitoring points is the same;

[0070] Based on the monitoring point, the collected data with time tags are obtained, and the collected data is converted to the range of 0 to 255. As a display parameter, the collected data of all monitoring devices at the monitoring point are counted in the same way to obtain the pixel color value. Then, a map with a one-to-one correspondence between pixels and monitoring points is created, and the pixel color value can be inserted.

[0071] In the process of acquiring collected data, the initial acquisition frequency of each monitoring point is the same. As time goes by, the acquired collected data needs to be identified and the data acquisition frequency needs to be adjusted. Generally, in order to ensure the comprehensiveness of the data, the initial data acquisition frequency is generally the highest value. In the subsequent process, the data acquisition frequency of each monitoring point continues to decrease, thereby reducing the load of the monitoring equipment and improving its endurance.

[0072] Step S300: Identify the data graph, determine the pollution area and its pollution type, and update the benchmark acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0073] The data map is the centralized data of all monitoring points. By identifying the data map, the polluted area and its pollution type can be determined, and the benchmark acquisition frequency of the monitoring point can be updated according to the polluted area and its pollution type. In actual applications, the benchmark acquisition frequency of polluted areas is generally higher, and the benchmark acquisition frequency of polluted areas with lighter pollution types will be lower. In addition, the benchmark acquisition frequency of non-polluted areas is even lower.

[0074] It should be noted that the benchmark acquisition frequency of the updated monitoring points is configured with the same benchmark acquisition frequency for all monitoring points in the polluted area after analyzing the polluted area.

[0075] Step S400: recording the time tags of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the reference acquisition frequency, and adjusting the recognition process of the data graph;

[0076] When each monitoring point uploads the collected data, it will also upload the time tag. The actual acquisition frequency can be calculated based on the time tag, and the actual acquisition frequency can be compared with the benchmark acquisition frequency to adjust the data graph recognition process. When each monitoring point acquires the collected data, a frequency adjustment ratio is generated based on the change in the collected data; the frequency adjustment ratio is proportional to the change.

[0077] The above content provides a two-party recognition architecture. Normally, after the monitoring point obtains the collected data and generates the data graph, with the help of the preset convolution kernel table (each convolution kernel in the convolution kernel table represents the image feature corresponding to a pollution type), the polluted area can be located in the data graph, and the pollution type of the polluted area can be determined, and then the benchmark acquisition frequency can be determined; on this basis, when each monitoring point obtains the collected data, it will also self-identify the collected data and adjust the benchmark acquisition frequency according to the changes in the collected data. If the change in the collected data is large, the frequency adjustment ratio will increase, and the actual acquisition frequency will be greater than the benchmark acquisition frequency. At this time, the corresponding area will be more easily judged as a polluted area, that is, in the convolution kernel matching process, the matching threshold will be lower (the matching process is that the convolution kernel traverses the data graph and calculates the matching degree. When the matching degree reaches the preset matching threshold, the matching area is regarded as a polluted area, and the pollution type corresponding to the convolution kernel is read as the pollution type of the matching area).

[0078] Figure 2 This is a first sub-process flowchart of a mangrove wetland pollutant monitoring method based on edge computing. The steps of obtaining a wetland area and creating a monitoring point based on the wetland area include:

[0079] Step S101: obtaining boundary information of the wetland, creating a wetland outline, and taking the inner area of ​​the wetland outline as the wetland area;

[0080] Step S102: determining the innermost contour according to the total length of the wetland contour;

[0081] Step S103: creating a preset number of equidistant contour lines based on the innermost contour and the wetland contour, and calculating the distances between adjacent contour lines;

[0082] Step S104: For each contour line, create a monitoring point based on the distance.

[0083] The mangrove wetland management personnel input the boundary information of the wetland to obtain the wetland outline. The internal area of ​​the wetland outline is the wetland area. Then, the total length of the wetland outline is obtained, and the length of the innermost outline is determined by the total length. The wetland outline is used as the outermost outline, and a preset number of equidistant contour lines are set between the outermost contour and the innermost contour. Monitoring points are set on each contour line at the distance between adjacent equidistant contour lines.

[0084] The above scheme is a uniform setting scheme based on the wetland contour. In fact, a grid of a preset size can also be set, and the intersection of the grids belonging to the wetland area can be used as monitoring points. This scheme can also obtain uniform monitoring points.

[0085] Figure 3The second sub-flow chart of the mangrove wetland pollutant monitoring method based on edge computing is as follows: the steps of acquiring collected data with time tags based on monitoring points and creating a data graph with time tags include:

[0086] Step S201: creating a reference map according to the positional relationship of the monitoring points; each pixel point in the reference map corresponds to each monitoring point one by one;

[0087] Step S202: receiving the initial collected data uploaded by the monitoring point, inputting the initial collected data into a preset conversion model, and obtaining a conversion value;

[0088] Step S203: Count the converted values, determine the color values ​​of the corresponding pixels in the reference image, and obtain an initial data image; wherein the data types of the collected data at the same monitoring point are at most three;

[0089] Step S204: when any monitoring point acquires new collected data with a time tag, the collected data is input into the conversion model to obtain a conversion value, and the color value of the corresponding pixel point in the initial data map is updated based on the conversion value as the data map corresponding to the time tag;

[0090] Step S205: Use the data graph at the latest time as a new initial data graph.

[0091] The monitoring points are read in sequence, and a reference map (two-dimensional array) is created according to the positional relationship of the monitoring points. Each pixel in the reference map corresponds to each monitoring point one by one. The initial collected data uploaded by each monitoring point is received, and the initial collected data is input into a preset conversion model to obtain a conversion value. This process is dedimensionalization to obtain a unified value from 0 to 255. Then, the value is used as a parameter of the pixel point and as a color value to insert the pixel point in the reference map to obtain an initial data map. The reason why it is called the initial data map is that when the data is first acquired, the data acquisition frequency of all monitoring points is the same, and one data needs to be acquired as a whole. The first data map obtained is called the initial data map.

[0092] Then, after the data graph is identified, the data acquisition frequency of each monitoring point will change. At this time, when any monitoring point obtains new collected data with a time tag, the collected data is converted into a conversion value with the help of the same conversion model, the initial data graph is read, and the corresponding pixel points are replaced. Further, the data graph at the latest moment is used as the new initial data graph; this process (step S204 and step S205) is continuously repeated to obtain data graphs at different moments.

[0093] Figure 4The third sub-process flowchart of the mangrove wetland pollutant monitoring method based on edge computing is as follows: the steps of identifying the data graph, determining the polluted area and its pollution type, and updating the benchmark acquisition frequency of the monitoring point according to the polluted area and its pollution type include:

[0094] Step S301: traverse the pixel points in the data graph and calculate the difference of the pixel points;

[0095] Step S302: dividing the data image into regions according to the difference of the pixel points to obtain sub-regions;

[0096] Step S303: calculating the area, color value mean and color value standard deviation of the sub-region, comparing the color value mean and color value standard deviation with preset color value parameters, and matching to obtain the polluted area and its pollution type;

[0097] Step S304: updating the baseline acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0098] Among them, the mapping relationship between color value parameters and pollution types is obtained in advance by staff statistics.

[0099] For the data graph at the latest moment, traverse the pixels in the data graph and calculate the difference of the pixels. The difference is generally based on the first-order difference. Of course, the second-order difference can also be introduced. The purpose is to identify the contour and divide the data graph into regions to obtain sub-regions; for each sub-region, calculate the regional area, color value mean and color value standard deviation of the sub-region, compare the color value mean and color value standard deviation with the preset color value parameters, and match the polluted area and its pollution type. The purpose of this process is to simplify the process and convert the existing commonly used convolution recognition process into a numerical comparison process. First, the relationship between the color value parameters and the pollution type is pre-stated. The color value parameters are the preset color value mean threshold and color value standard deviation threshold. The staff pre-statistics the color value parameters corresponding to some pollution types. In actual use, the color value parameters reached by the color value mean and the color value standard deviation are judged to obtain the pollution type, which is then used as the pollution type of the pollution area.

[0100] The benchmark acquisition frequency of the monitoring point is updated according to the polluted area and its pollution type. The more serious the pollution type, the higher the benchmark acquisition frequency. In addition, the update process is a global update process, and each area must be updated, including non-polluted areas. In order to simplify the operation, non-pollution can be regarded as a pollution type, so that each sub-area can get a corresponding pollution type.

[0101] Furthermore, the rules for updating the benchmark acquisition frequency of monitoring points according to the pollution area and its pollution type include: ; In the formula, For monitoring point The reference acquisition frequency at It is The middle point of the pollution area The characteristic value of is determined by the pollution type. The higher the pollution level of the pollution type, The bigger; is the total number of polluted areas, and is the size of any contaminated area; The principle of the above rule is that for a certain monitoring point, the influence of all polluted areas at the monitoring point is calculated. First, a characteristic value is determined according to the pollution type of the polluted area. The more serious the pollution, the larger the characteristic value; then, the influence of each point in the polluted area on the monitoring point is calculated, which is proportional to the characteristic value and inversely proportional to the distance; finally, the influence of all polluted areas on the monitoring point is accumulated, a coefficient is determined, and then multiplied by the preset standard frequency.

[0102] It should be noted that the coefficient is proportional to the degree of influence and has a value between 0 and 1. The standard frequency may be an initial acquisition frequency, and the initial acquisition frequency adopts the maximum frequency.

[0103] Figure 5 The fourth sub-process flowchart of the mangrove wetland pollutant monitoring method based on edge computing is as follows: the steps of recording the time tags of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the benchmark acquisition frequency, and adjusting the recognition process of the data graph include:

[0104] Step S401: record the time tags of the collected data fed back by each monitoring point, calculate the difference of the time tags, and determine the actual acquisition frequency according to the difference;

[0105] Step S402: comparing the actual acquisition frequency with the reference acquisition frequency, and inversely calculating the frequency adjustment ratio;

[0106] Step S403: adjusting the color value parameters according to the frequency adjustment ratio;

[0107] The adjustment process of the color value parameter is proportional to the inverse of the frequency adjustment ratio.

[0108] In an example of the technical solution of the present invention, a specific solution for the identification process of the adjustment data graph is provided, which records the time tags of the collected data fed back by each monitoring point, calculates the difference of the time tags, and the difference is actually the instantaneous frequency. The latest instantaneous frequencies are obtained, and the average is calculated to obtain the actual acquisition frequency; the actual acquisition frequency is compared with the benchmark acquisition frequency. Since the actual acquisition frequency is obtained by independently adjusting the benchmark acquisition frequency according to the actual situation at the monitoring point, the frequency adjustment ratio can be reversed according to the comparison result, and then the color value parameter is adjusted. The color value parameter is the threshold in the pollution type identification process. Generally, the higher the frequency adjustment ratio, the greater the frequency change of the collected data. At this time, the area that was originally not polluted may have been polluted, and it is necessary to lower the threshold (color value parameter) so that each sub-area can be more easily marked as a polluted area. While issuing an alarm, a higher benchmark acquisition frequency can also be obtained.

[0109] Figure 6 The following is a structural block diagram of a mangrove wetland pollutant monitoring system based on edge computing. In an embodiment of the present invention, a mangrove wetland pollutant monitoring system based on edge computing is provided. The system 10 includes:

[0110] A monitoring point creation module 11 is used to obtain a wetland area and create a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed;

[0111] The data graph creation module 12 is used to obtain the collected data with time tags based on the monitoring points and create a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; and the initial data acquisition frequency of all monitoring points is the same;

[0112] A reference frequency determination module 13 is used to identify the data graph, determine the pollution area and its pollution type, and update the reference acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0113] The recognition process adjustment module 14 is used to record the time tags of the collected data fed back by each monitoring point, calculate the actual acquisition frequency, compare the actual acquisition frequency with the reference acquisition frequency, and adjust the recognition process of the data graph;

[0114] When acquiring collected data at the monitoring point, a frequency adjustment ratio is generated according to the change amount of the collected data; the frequency adjustment ratio is proportional to the change amount.

[0115] Furthermore, the monitoring point creation module 11 includes:

[0116] A contour creation unit is used to obtain boundary information of the wetland, create a wetland contour, and use the inner area of ​​the wetland contour as the wetland area;

[0117] An inner contour determining unit, used for determining the innermost contour according to the total length of the wetland contour;

[0118] An equidistant line construction unit, used for creating a preset number of equidistant contour lines based on the innermost contour and the wetland contour, and calculating the distances between adjacent contour lines;

[0119] The point selection unit is used to create a monitoring point for each contour line based on the distance.

[0120] Specifically, the data graph creation module 12 includes:

[0121] A reference map creation unit, used to create a reference map according to the positional relationship of the monitoring points; each pixel point in the reference map corresponds to each monitoring point one by one;

[0122] A data conversion unit is used to receive the initial collected data uploaded by the monitoring point, input the initial collected data into a preset conversion model, and obtain a conversion value;

[0123] The numerical statistics unit is used to statistically convert numerical values, determine the color values ​​of corresponding pixels in the reference image, and obtain the initial data image; wherein, the data types of the collected data of the same monitoring point are at most three;

[0124] A data map updating unit, for when any monitoring point acquires new collected data with a time tag, inputs the collected data into the conversion model to obtain a conversion value, and based on the conversion value, updates the color value of the corresponding pixel point in the initial data map as the data map corresponding to the time tag;

[0125] The base updating unit is used to use the data graph at the latest moment as a new initial data graph.

[0126] Furthermore, the reference frequency determination module 13 includes:

[0127] A difference calculation unit, used for traversing the pixel points in the data graph and calculating the difference of the pixel points;

[0128] A region segmentation unit is used to segment the data image into regions according to the difference of pixel points to obtain sub-regions;

[0129] A pollution matching unit is used to calculate the area, color value mean and color value standard deviation of the sub-area, compare the color value mean and color value standard deviation with preset color value parameters, and match the pollution area and its pollution type;

[0130] A frequency updating unit, used to update the base acquisition frequency of the monitoring point according to the pollution area and its pollution type;

[0131] Among them, the mapping relationship between color value parameters and pollution types is obtained in advance by staff statistics.

[0132] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A mangrove wetland pollutant monitoring method based on edge computing, characterized in that: The method comprises: Acquire a wetland area, and create a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed; Acquire collected data with time tags based on monitoring points and create a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; and the initial data acquisition frequency of all monitoring points is the same; Identify the data graph, determine the polluted area and its pollution type, and update the baseline acquisition frequency of the monitoring point according to the polluted area and its pollution type; Record the time tags of the collected data fed back by each monitoring point, calculate the actual acquisition frequency, compare the actual acquisition frequency with the benchmark acquisition frequency, and adjust the recognition process of the data graph; When acquiring collected data at the monitoring point, a frequency adjustment ratio is generated according to the amount of change in the collected data; the frequency adjustment ratio is proportional to the amount of change; The step of identifying the data graph, determining the pollution area and its pollution type, and updating the benchmark acquisition frequency of the monitoring point according to the pollution area and its pollution type comprises: Traverse the pixels in the data graph and calculate the difference of the pixels; The data graph is divided into regions according to the difference of pixel points to obtain sub-regions; Calculating the area, color value mean and color value standard deviation of the sub-region, comparing the color value mean and color value standard deviation with preset color value parameters, and matching to obtain the polluted area and its pollution type; Update the baseline acquisition frequency of monitoring points according to the polluted area and its pollution type; Among them, the mapping relationship between color value parameters and pollution types is obtained by staff in advance; The rules for updating the baseline acquisition frequency of monitoring points according to the pollution area and its pollution type include: ; In the formula, For monitoring point The reference acquisition frequency at It is The middle point of the pollution area The characteristic value of is determined by the pollution type. The higher the pollution level of the pollution type, The bigger; is the total number of polluted areas, and is the size of any contaminated area; is the preset standard frequency; The steps of recording the time tags of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the reference acquisition frequency, and adjusting the recognition process of the data graph include: Record the time tags of the collected data fed back by each monitoring point, calculate the difference of the time tags, and determine the actual acquisition frequency based on the difference; Compare the actual acquisition frequency with the benchmark acquisition frequency, and reversely calculate the frequency adjustment ratio; Adjust the color value parameters according to the frequency adjustment ratio; The color value parameter is proportional to the inverse of the frequency adjustment ratio.

2. The method for monitoring pollutants in mangrove wetlands based on edge computing according to claim 1 is characterized in that: The step of obtaining the wetland area and creating monitoring points based on the wetland area includes: Obtain the boundary information of the wetland, create a wetland outline, and use the inner area of ​​the wetland outline as the wetland area; Determine the innermost contour based on the total length of the wetland contour; Create a preset number of equally spaced contour lines based on the innermost contour and the wetland contour, and calculate the distances between adjacent contour lines; For each contour line, a monitoring point is created based on the distance.

3. The method for monitoring pollutants in mangrove wetlands based on edge computing according to claim 1 is characterized in that: The step of acquiring the collected data with time tags based on the monitoring points and creating a data graph with time tags includes: Creating a reference map based on the positional relationship of the monitoring points; each pixel in the reference map corresponds one-to-one to each monitoring point; Receive the initial collected data uploaded by the monitoring point, input the initial collected data into a preset conversion model, and obtain the conversion value; Statistically convert the values, determine the color values ​​of the corresponding pixels in the reference image, and obtain the initial data image; wherein the data types of the collected data at the same monitoring point can be up to three types; When any monitoring point obtains new collected data with a time tag, the collected data is input into the conversion model to obtain a conversion value, and the color value of the corresponding pixel point in the initial data map is updated based on the conversion value as the data map corresponding to the time tag; The data graph at the latest moment is used as the new initial data graph.

4. A mangrove wetland pollutant monitoring system based on edge computing, characterized in that: The system comprises: A monitoring point creation module is used to obtain a wetland area and create a monitoring point based on the wetland area; the monitoring point is a point where a monitoring device is installed; A data graph creation module is used to obtain collected data with time tags based on monitoring points and create a data graph with time tags; wherein the data acquisition frequency of each monitoring point is a dynamic frequency, which is determined by the collected data of the monitoring point; the initial data acquisition frequency of all monitoring points is the same; A reference frequency determination module, used to identify the data graph, determine the pollution area and its pollution type, and update the reference acquisition frequency of the monitoring point according to the pollution area and its pollution type; The recognition process adjustment module is used to record the time tags of the collected data fed back by each monitoring point, calculate the actual acquisition frequency, compare the actual acquisition frequency with the reference acquisition frequency, and adjust the recognition process of the data graph; When acquiring collected data at the monitoring point, a frequency adjustment ratio is generated according to the amount of change in the collected data; the frequency adjustment ratio is proportional to the amount of change; The reference frequency determination module comprises: A difference calculation unit, used for traversing the pixel points in the data graph and calculating the difference of the pixel points; A region segmentation unit is used to segment the data image into regions according to the difference of pixel points to obtain sub-regions; A pollution matching unit is used to calculate the area, color value mean and color value standard deviation of the sub-area, compare the color value mean and color value standard deviation with preset color value parameters, and match the pollution area and its pollution type; A frequency updating unit, used to update the base acquisition frequency of the monitoring point according to the pollution area and its pollution type; Among them, the mapping relationship between color value parameters and pollution types is obtained by staff in advance; The rules for updating the baseline acquisition frequency of monitoring points according to the pollution area and its pollution type include: ; In the formula, For monitoring point The reference acquisition frequency at It is The middle point of the pollution area The characteristic value of is determined by the pollution type. The higher the pollution level of the pollution type, The bigger; is the total number of polluted areas, and is the size of any contaminated area; is the preset standard frequency; The contents of the recognition process of recording the time tags of the collected data fed back by each monitoring point, calculating the actual acquisition frequency, comparing the actual acquisition frequency with the reference acquisition frequency, and adjusting the data graph include: Record the time tags of the collected data fed back by each monitoring point, calculate the difference of the time tags, and determine the actual acquisition frequency based on the difference; Compare the actual acquisition frequency with the benchmark acquisition frequency, and reversely calculate the frequency adjustment ratio; Adjust the color value parameters according to the frequency adjustment ratio; The color value parameter is proportional to the inverse of the frequency adjustment ratio.

5. The mangrove wetland pollutant monitoring system based on edge computing according to claim 4 is characterized in that: The monitoring point creation module includes: A contour creation unit is used to obtain boundary information of the wetland, create a wetland contour, and use the inner area of ​​the wetland contour as the wetland area; An inner contour determining unit, used for determining the innermost contour according to the total length of the wetland contour; an equidistant line construction unit, for creating a preset number of equidistant contour lines based on the innermost contour and the wetland contour, and calculating the distances between adjacent contour lines; The point selection unit is used to create a monitoring point for each contour line based on the distance.

6. The mangrove wetland pollutant monitoring system based on edge computing according to claim 4 is characterized in that: The data graph creation module includes: A reference map creation unit, used to create a reference map according to the positional relationship of the monitoring points; each pixel point in the reference map corresponds to each monitoring point one by one; A data conversion unit is used to receive the initial collected data uploaded by the monitoring point, input the initial collected data into a preset conversion model, and obtain a conversion value; The numerical statistics unit is used to statistically convert numerical values, determine the color values ​​of corresponding pixels in the reference image, and obtain the initial data image; wherein, the data types of the collected data at the same monitoring point can be up to three types; A data map updating unit, for when any monitoring point acquires new collected data with a time tag, inputs the collected data into the conversion model to obtain a conversion value, and based on the conversion value, updates the color value of the corresponding pixel point in the initial data map as the data map corresponding to the time tag; The base updating unit is used to use the data graph at the latest moment as a new initial data graph.

Images