A water body analysis system and method based on edge computing and integrated sensing and computing

Abstract

The present invention discloses a sensing and computing integrated water body analysis system and method based on edge computing, and relates to the technical field of water body detection. The present invention collects water body environment data by remote sensing technology, places edge water body detection equipment at optimal placement points, and constructs a distributed edge detection network; extracts the most significant substance type in the water body image under each light, matches the light type with the most significant substance type, and generates a light-substance matching database; calculates the most significant substance abnormality threshold of the abnormal water body image under each light according to the most significant substance matched by each light in the light-substance matching database; uses abnormal clarity to perform shallow analysis and judgment on the clarity of real-time water body images, and performs abnormal warning; selects the most significant abnormal threshold corresponding to the light type with the largest number in the central cloud to perform in-depth analysis and judgment on the real-time water body, and outputs the final judgment result.

Description

Technical Field

[0001] The present invention relates to the field of water body detection technology, and specifically to a water body analysis system and method based on edge computing and integrated sensing and computing. Background Art

[0002] Water is the source of life and an indispensable resource for human life, industrial production, agricultural irrigation and other activities. However, with the acceleration of industrialization and urbanization and the growth of population, a large amount of pollutants are discharged into water bodies, such as industrial wastewater, domestic sewage, agricultural non-point source pollution, etc., which leads to the deterioration of water quality and poses a serious threat to human health and ecological environment. Therefore, accurate and timely monitoring of water quality is crucial for the protection and management of water resources. With the continuous advancement of science and technology, various advanced analytical instruments, sensor technologies, information technologies, etc. have emerged, providing strong support for the development of water body detection technology. At present, the monitoring method of water body information is mainly based on physical and chemical monitoring technology. Traditional monitoring methods mainly rely on monitoring stations arranged on the monitoring section, and use manual or semi-automatic monitoring methods to obtain water resource information at the station. However, this method requires high investment costs, long time, and limited coverage of water areas. In addition, the setting of hydrological stations will be affected by the natural environment. It is difficult to set up hydrological stations in some places with harsh natural environments, and it cannot meet the requirements of spatial continuity monitoring and supervision of water bodies. Nowadays, most water quality detection mainly relies on water quality analyzers, but water quality analyzers are expensive and not convenient to use. They cannot detect small water bodies. As a result, the use of visual sensors to collect images for water body detection has been born, but there are still some problems. Visual images are overly dependent on environmental factors. When the weather is gloomy or at night, the light is dim, resulting in unclear images collected, which will greatly cause water body detection equipment to misjudge and fail to provide accurate water body analysis results. Summary of the invention

[0003] The purpose of the present invention is to provide a water body analysis system and method based on edge computing to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A water body analysis method based on edge computing and sensing integration, the method comprising the following steps:

[0006] S100, using remote sensing technology to collect water environment data, calculating the optimal placement point of edge water body detection equipment according to the water environment data, placing the edge water body detection equipment at the optimal placement point, and constructing a distributed edge detection network; the distributed edge detection network includes edge water body detection equipment and a central cloud;

[0007] Furthermore, the specific steps for constructing a distributed edge detection network are as follows:

[0008] S101, using remote sensing technology to collect water environment data, extracting the edge shape of the water body in the water environment, dividing the edge of the water body into n broken lines connected, assuming that the broken line constituting the edge of the water body is , represents the 1st, 2nd, 3rd, ..., nth broken lines that constitute the edge of the water body, where n is a positive integer; the coordinates of the connection points of each broken line in the edge of the water body are measured as follows: , It represents the 1st, 2nd, 3rd, ..., nth connection points in all the polylines at the edge of the water body. The optimal placement point of the edge water body detection equipment is calculated based on all the connection points. The formula is:

[0009] ;

[0010] In the formula, represents the calculated optimal placement point coordinates of the edge water body detection device, i represents any one of the n connection points, ;

[0011] S102, after calculating the best placement point of the edge water body detection device, collect the farthest visible distance of the visual sensor in the edge water body detection device as L, place the first edge water body detection device at the best placement point, and assume that the detection range of each edge water body detection device is a circle with a radius of L; when the area of ​​the water body is larger than the detection area of ​​the edge water body detection device at the best placement point, place six identical edge water body detection devices tangent to each other around the first edge water body detection device, so that the first edge water body detection device and the six tangent edge water body detection devices form the vertices and center point of a regular hexagon, and calculate the number of edge water body detection devices that need to be placed in the water body, the formula is:

[0012] ;

[0013] In the formula, sm represents the number of edge water body detection devices that need to be placed in the water body, the calculation result is rounded down, Q represents the area of ​​the water body, and L represents the maximum detection distance of the edge water body detection device. After placing the edge water body detection device in the water body, each edge water body detection device is connected to the central cloud wirelessly to form a distributed edge detection network.

[0014] Using wireless communication and analyzing the water environment to build a distributed edge detection network can quickly respond to abnormal situations and detect them close to the water data source, reducing the delay of data transmission and processing. The distributed edge detection network can continuously monitor the water and capture these changes in real time. The use of multi-node collaboration greatly improves the detection accuracy.

[0015] S200, using a multispectral camera to extract all light types and corresponding bands of natural light, collecting water samples, collecting historical material types in the water, collecting historical water images taken by a multispectral camera under different light, extracting the most significant material type in the water image under each light, matching the light type with the most significant material type, and generating a light-material matching database;

[0016] Furthermore, the specific steps for generating a light-matter matching database are:

[0017] S201, using a multi-spectral camera to extract all light types existing in natural light , represents the 1st, 2nd, 3rd, ..., mth light in natural light, where m is a positive integer; the corresponding wavelength band of each light in natural light measured by the spectrometer is , Indicates the wavelength bands of the 1st, 2nd, 3rd, ..., mth light rays in the measured natural light, In the above equation, q represents the minimum value of each light band, and p represents the maximum value of each light band;

[0018] S202. Collect all the substances present in the water. , Indicates the 1st, 2nd, 3rd, .., jth substance in the collected water body, where j is a positive integer; collects image data of water bodies under different light in history, grays the image data using the Sobel operator, and calculates the gradient amplitude of the edge gray value of each object in the water body under different light. , Represents the calculated gradient amplitude of the edge grayscale value of the 1st, 2nd, 3rd, ..., jth material in the water body; the material corresponding to the maximum gradient amplitude in the image is judged to be the most significant material type, and the gradient amplitude of the material edge grayscale value in the water body image under each light is calculated in turn to obtain the most significant material in the image under each light, and the light type and material type are matched as , Represents the matching results of the 1st, 2nd, 3rd, ..., mth light and matter, and uses the matching results to generate a light-matter matching database.

[0019] Different substances in water bodies have different displays under different lighting conditions. Some substances will have greatly improved clarity or have significantly prominent colors under certain lighting conditions. Therefore, matching the most prominent substances under different lighting conditions can increase the accuracy and breadth of water body detection, allowing detection equipment to detect water bodies under different lighting conditions.

[0020] S300, collecting water body images under each light when the water body is abnormal in history, and calculating the most significant substance abnormality threshold of the abnormal water body image under each light according to the most significant substance matched by each light in the light-substance matching database;

[0021] Furthermore, the specific steps for calculating the most significant material anomaly threshold of the abnormal water image under each light are:

[0022] S301, collect water body images under each light when the water body is abnormal and normal in history, extract the most significant substance type in the image under each light, compare the most significant substances in the two images when the water body is in different states, and set the most significant substance matching state data of each light to be , It represents the 1st, 2nd, 3rd, ..., kth state data contained in the most significant substance matched by each light, where k is a positive integer; after the most significant substance in two states, the state with the largest difference between the most significant substance in the two states is extracted as the abnormal state feature; after comparison, the abnormal state feature of the most significant substance matched by each light is obtained as follows: , Indicates the most significant abnormal state characteristics of the material matched by the first, second, third, ..., mth light rays obtained by comparison;

[0023] S302, after determining the most significant abnormal state characteristics of the substance matched by each light, extract the most significant abnormal state characteristics of the substance under different light in the image data of the historical water body anomaly, and calculate the abnormal threshold of the most significant substance under each light respectively, the formula is:

[0024] ;

[0025] In the formula, and represents the upper and lower limits of the most significant material abnormality threshold under each light, E represents the most significant material abnormal state feature value in the image under each light in the collected history, h represents the number of historical abnormal state feature values ​​collected, and Ep represents the average value of h abnormal state features in the collected history; after calculation, the abnormal threshold of the most significant material under each light is , Represents the calculated anomaly threshold for the most significant material match for each ray.

[0026] S400, collecting visual images of water body anomalies in history, and calculating abnormal clarity using the visual images of water body anomalies; using the visual sensor in the edge water body detection equipment to collect water body images in real time, obtaining the clarity of the real-time water body images, and using the abnormal clarity to perform shallow analysis and judgment on the clarity of the real-time water body images, and issue abnormal warnings;

[0027] Furthermore, the specific steps of using abnormal clarity to conduct shallow analysis and judgment on the clarity of real-time water body images are as follows:

[0028] S401, collect visual images of water body anomalies in history, grayscale the collected visual images of water body anomalies, and calculate the grayscale value variance of each collected historical visual image, the formula is: , in the formula, represents the grayscale value variance of the historical visual image, d represents the number of grayscale values ​​in the collected historical visual images, H represents the grayscale value in the collected historical visual images, Represents the average grayscale value in the historical visual image; the grayscale value variance of each visual image in the collected history is obtained by calculation; the grayscale value variance is used to represent the clarity of each visual image;

[0029] S402, using the clarity of each visual image in the history to calculate the abnormal clarity, the formula is: , in the formula, Indicates the abnormal clarity of the calculated water body image, represents the average clarity of all visual images in history, represents the standard deviation of clarity of all visual images in history;

[0030] S403, using the visual sensor in the edge water body detection device to collect water body images in real time, and obtaining the clarity of the real-time water body image as Qs, and using the calculated abnormal clarity to make a shallow judgment on the clarity of the real-time water body image. When the water body detection equipment at the edge determines that there is an abnormal risk in the water body, it will issue an early warning; At this time, the edge water body detection equipment determines that the water body is normal.

[0031] When the water body is first analyzed at a shallow level, the convenience of the edge water body detection equipment can quickly find the abnormal edge position without information transmission, increasing the real-time nature of abnormal judgment, and simple abnormal judgment can be handled through shallow analysis, greatly reducing the workload of the system and reducing energy consumption.

[0032] S500: After the edge water body detection device issues an early warning, the number of light rays in different bands of natural light is collected in real time, and the edge water body detection device sends the number of light rays and image data of different light rays to the central cloud; the central cloud selects the most significant abnormal threshold corresponding to the light type with the largest number to conduct in-depth analysis and judgment of the real-time water body, and outputs the final judgment result.

[0033] Furthermore, the specific steps for in-depth analysis and judgment of real-time water bodies are as follows:

[0034] S501. After the edge water body detection device issues an early warning, the spectral radiometer is used to collect the number of light rays in different bands of natural light in real time, and the light band with the largest number of light rays is extracted. The light type Rs corresponding to the light band with the largest number of light rays in real time and the most significant substance type Ws that matches are searched in the light-matter matching database; the edge water body detection device uses the single wavelength mode of the spectral imager to select the band with the most real-time light rays, and collects the water body image under the light Rs in real time and transmits it to the central cloud; after receiving the real-time water body image, the central cloud extracts the abnormal state characteristics of the most significant substance Ws; and calculates the abnormal state characteristic value of the most significant substance Ws in the real-time water body image to obtain the real-time abnormal state characteristic value Es;

[0035] S502, using the abnormal threshold of the most significant substance Ws in the real-time water body image to perform in-depth analysis and judgment on the real-time abnormal state feature value, when When the water body is abnormal, the warning signal of the edge water body detection device is transmitted to the user; When the edge water body detection device warns that there is no abnormality in the water body, a warning cancellation signal is sent to the edge water body detection device that warns, and the edge water body detection device stops warning.

[0036] Performing deep judgment after shallow judgment can further deepen the results of shallow judgment and prevent the one-sidedness of shallow judgment. The combination of shallow judgment and deep judgment can give full play to the advantages of both. Shallow judgment provides a wider potential edge range, avoiding the situation where deep judgment may miss some edges due to the complexity of the model; deep judgment screens and refines the results of shallow judgment, reducing the misjudgment caused by shallow judgment being easily disturbed by brightness and other factors.

[0037] A sensing and computing integrated water body analysis system based on edge computing, the sensing and computing integrated water body analysis system includes a data collection module, an edge detection network construction module, a light matching module, a threshold calculation module, a shallow layer judgment module and a depth judgment module;

[0038] The data collection module is used to collect image data of water body anomalies in history, and to collect types and bands of natural light in history;

[0039] The edge detection network construction module is used to analyze the environmental data of the water body, calculate the optimal placement point of the edge water body detection device, place the edge water body detection device in the water body, and build a distributed edge detection network;

[0040] The light matching module is used to analyze the light types and wavelengths in natural light to determine the most significant water substance under each light.

[0041] The threshold calculation module is used to obtain the abnormal state characteristics of the most significant substance in each light by using the image data of the water body under different light in history, and calculate the abnormal threshold;

[0042] The shallow layer judgment module is used to use the edge water body detection equipment to judge the shallow layer abnormality of the water body and issue an early warning;

[0043] The depth judgment module is used to use the central cloud to make depth anomaly judgment on the water body of the edge water body detection device that issues an early warning.

[0044] The edge detection network building module includes an optimal placement point unit and a network building unit;

[0045] The optimal placement point unit is used to analyze the water body environment data, divide the water body edge into broken lines, and calculate the optimal placement point using the coordinates of the connecting points of the broken lines;

[0046] The network construction unit is used to place edge water body detection equipment in the water body using the best placement point and calculate the required number of equipment.

[0047] The shallow layer judgment module includes an abnormal clarity calculation unit and a shallow layer judgment unit;

[0048] The abnormal clarity calculation unit is used to calculate the abnormal clarity by using the abnormal water body images in history;

[0049] The shallow layer judgment unit is used to collect real-time water body images and use abnormal clarity to make shallow layer judgments on the real-time water body.

[0050] The depth judgment module includes a real-time light judgment unit and a depth judgment unit;

[0051] The real-time light judging unit is used to judge the quantity of light in different bands in real time to obtain the real-time light type and the most significant substance type;

[0052] The depth judgment unit is used to judge the real-time depth of the water body by using the abnormal state characteristics of the most significant substance in real time, and to process the early warning of the edge water body detection equipment according to the judgment result.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. The present invention can quickly respond to abnormal situations by constructing a distributed edge detection network, and can perform detection at a location close to the water data source, reducing the delay in data transmission and processing. In addition, the distributed edge detection network can continuously monitor the water body and capture these changes in real time.

[0055] 2. The present invention analyzes the most significant substances that match under different light, and uses the most significant substances that match under different light to judge the water body, thereby increasing the accuracy and universality of water body detection, enabling the detection equipment to detect water bodies under different light conditions, and avoiding water body judgment errors caused by light changes when judging with a single light.

[0056] 3. The present invention detects water bodies by combining shallow judgment and deep judgment. The combination of shallow judgment and deep judgment can give full play to the advantages of both. Shallow judgment provides a wider potential edge range, avoiding the situation where deep judgment may miss some edges due to the overly complex model; deep judgment screens and refines the results of shallow judgment, reducing the misjudgment caused by shallow judgment being easily disturbed by brightness and other factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a module distribution diagram of a water body analysis system based on edge computing and sensing integration of the present invention;

[0058] Figure 2 This is a schematic diagram of the steps of a water body analysis method based on edge computing and sensing integration of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution.

[0061] A water body analysis method based on edge computing and sensing integration, the method comprising the following steps:

[0062] S100, using remote sensing technology to collect water environment data, calculating the optimal placement point of edge water body detection equipment according to the water environment data, placing the edge water body detection equipment at the optimal placement point, and constructing a distributed edge detection network; the distributed edge detection network includes edge water body detection equipment and a central cloud;

[0063] The specific steps to construct a distributed edge detection network are:

[0064] S101, using remote sensing technology to collect water environment data, extracting the edge shape of the water body in the water environment, dividing the edge of the water body into n broken lines connected, assuming that the broken line constituting the edge of the water body is , represents the 1st, 2nd, 3rd, ..., nth broken lines that constitute the edge of the water body, where n is a positive integer; the coordinates of the connection points of each broken line in the edge of the water body are measured as follows: , It represents the 1st, 2nd, 3rd, ..., nth connection points in all the polylines at the edge of the water body. The optimal placement point of the edge water body detection equipment is calculated based on all the connection points. The formula is:

[0065] ;

[0066] In the formula, represents the calculated optimal placement point coordinates of the edge water body detection device, i represents any one of the n connection points, ;

[0067] S102, after calculating the best placement point of the edge water body detection device, collect the farthest visible distance of the visual sensor in the edge water body detection device as L, place the first edge water body detection device at the best placement point, and assume that the detection range of each edge water body detection device is a circle with a radius of L; when the area of ​​the water body is larger than the detection area of ​​the edge water body detection device at the best placement point, place six identical edge water body detection devices tangent to each other around the first edge water body detection device, so that the first edge water body detection device and the six tangent edge water body detection devices form the vertices and center point of a regular hexagon, and calculate the number of edge water body detection devices that need to be placed in the water body, the formula is:

[0068] ;

[0069] In the formula, sm represents the number of edge water body detection devices that need to be placed in the water body, the calculation result is rounded down, Q represents the area of ​​the water body, and L represents the maximum detection distance of the edge water body detection device. After placing the edge water body detection device in the water body, each edge water body detection device is connected to the central cloud wirelessly to form a distributed edge detection network.

[0070] Using wireless communication and analyzing the water environment to build a distributed edge detection network can quickly respond to abnormal situations and detect them close to the water data source, reducing the delay of data transmission and processing. The distributed edge detection network can continuously monitor the water and capture these changes in real time. The use of multi-node collaboration greatly improves the detection accuracy.

[0071] S200, using a multispectral camera to extract all light types and corresponding bands of natural light, collecting water samples, collecting historical material types in the water, collecting historical water images taken by a multispectral camera under different light, extracting the most significant material type in the water image under each light, matching the light type with the most significant material type, and generating a light-material matching database;

[0072] The specific steps to generate a light-matter matching database are:

[0073] S201, using a multi-spectral camera to extract all light types existing in natural light , represents the 1st, 2nd, 3rd, ..., mth light in natural light, where m is a positive integer; in a multi-spectral camera, a spectroscopic element is used to classify natural light into multiple light sources; then, a spectrometer measures the wavelength of each light source so that each light source has a wavelength range. Suppose the wavelength of each light source in natural light measured by a spectrometer is , Indicates the wavelength bands of the 1st, 2nd, 3rd, ..., mth light rays in the measured natural light, In the above equation, q represents the minimum value of each light band, and p represents the maximum value of each light band;

[0074] S202. Collect all the substances present in the water. , represents the 1st, 2nd, 3rd, .., jth substance in the collected water, where j is a positive integer; the single wavelength mode of the spectral imager is used to select the light band, and the spectral imager is used to obtain the image of the light of the selected band. The image data of the water body under different light in history are collected many times, and the image data is grayed using the Sobel operator, and the gradient amplitude of the gray value of the edge of each object in the water body under different light is calculated. , Represents the calculated gradient amplitude of the edge grayscale value of the 1st, 2nd, 3rd, ..., jth material in the water body; the material corresponding to the maximum gradient amplitude in the image is judged to be the most significant material type, and the gradient amplitude of the material edge grayscale value in the water body image under each light is calculated in turn to obtain the most significant material in the image under each light, and the light type and material type are matched as , Represents the matching results of the 1st, 2nd, 3rd, ..., mth light and matter, and uses the matching results to generate a light-matter matching database.

[0075] Different substances in water bodies have different displays under different lighting conditions. Some substances will have greatly improved clarity or have significantly prominent colors under certain lighting conditions. Therefore, matching the most prominent substances under different lighting conditions can increase the accuracy and breadth of water body detection, allowing detection equipment to detect water bodies under different lighting conditions.

[0076] S300, collecting water body images under each light when the water body is abnormal in history, and calculating the most significant substance abnormality threshold of the abnormal water body image under each light according to the most significant substance matched by each light in the light-substance matching database;

[0077] The specific steps for calculating the most significant material anomaly threshold of abnormal water images under each light are:

[0078] S301, collect water body images under each light when the water body is abnormal and normal in history, extract the most significant substance type in the image under each light, compare the most significant substances in the two images when the water body is in different states, and set the most significant substance matching state data of each light to be , It represents the 1st, 2nd, 3rd, ..., kth state data contained in the most significant substance matched by each light, where k is a positive integer; after the most significant substance in two states, the state with the largest difference between the most significant substance in the two states is extracted as the abnormal state feature; after comparison, the abnormal state feature of the most significant substance matched by each light is obtained as follows: , Indicates the most significant abnormal state characteristics of the material matched by the first, second, third, ..., mth light rays obtained by comparison;

[0079] The images produced by each substance in the water body will be different under different light. For example, the chlorophyll in the algae in the water body has a certain absorption capacity for green light, and will also produce a certain amount of scattering. Under the irradiation of blue and green light, the color of the algae will be brighter, and due to the scattering effect, the distribution of the algae in the water body will be more obvious. At this time, the most prominent substance corresponding to the green light is the algae;

[0080] S302, after determining the most significant abnormal state characteristics of the substance matched by each light, extract the most significant abnormal state characteristics of the substance under different light in the image data of the historical water body anomaly, and calculate the abnormal threshold of the most significant substance under each light respectively, the formula is:

[0081] ;

[0082] In the formula, and represents the upper and lower limits of the most significant material abnormality threshold under each light, E represents the most significant material abnormal state feature value in the image under each light in the collected history, h represents the number of historical abnormal state feature values ​​collected, and Ep represents the average value of h abnormal state features in the collected history; after calculation, the abnormal threshold of the most significant material under each light is , Represents the calculated anomaly threshold for the most significant material match for each ray.

[0083] S400, collecting visual images of water body anomalies in history, and calculating abnormal clarity using the visual images of water body anomalies; using the visual sensor in the edge water body detection equipment to collect water body images in real time, obtaining the clarity of the real-time water body images, and using the abnormal clarity to perform shallow analysis and judgment on the clarity of the real-time water body images, and issue abnormal warnings;

[0084] The specific steps for using abnormal clarity to conduct shallow analysis and judgment of the clarity of real-time water body images are as follows:

[0085] S401, collect visual images of water body anomalies in history, grayscale the collected visual images of water body anomalies, and calculate the grayscale value variance of each collected historical visual image, the formula is: , in the formula, represents the grayscale value variance of the historical visual image, d represents the number of grayscale values ​​in the collected historical visual images, H represents the grayscale value in the collected historical visual images, Represents the average grayscale value in the historical visual image; the grayscale value variance of each visual image in the collected history is obtained by calculation; the grayscale value variance is used to represent the clarity of each visual image;

[0086] S402, using the clarity of each visual image in the history to calculate the abnormal clarity, the formula is: , in the formula, Indicates the abnormal clarity of the calculated water body image, represents the average clarity of all visual images in history, represents the standard deviation of clarity of all visual images in history;

[0087] S403, using the visual sensor in the edge water body detection device to collect water body images in real time, and obtaining the clarity of the real-time water body image as Qs, and using the calculated abnormal clarity to make a shallow judgment on the clarity of the real-time water body image. When the water body detection equipment at the edge determines that there is an abnormal risk in the water body, it will issue an early warning; At this time, the edge water body detection equipment determines that the water body is normal.

[0088] When the water body is first analyzed at a shallow level, the convenience of the edge water body detection equipment can quickly find the abnormal edge position without information transmission, increasing the real-time nature of abnormal judgment, and simple abnormal judgment can be handled through shallow analysis, greatly reducing the workload of the system and reducing energy consumption.

[0089] S500: After the edge water body detection device issues an early warning, the number of light rays in different bands of natural light is collected in real time, and the edge water body detection device sends the number of light rays and image data of different light rays to the central cloud; the central cloud selects the most significant abnormal threshold corresponding to the light type with the largest number to conduct in-depth analysis and judgment of the real-time water body, and outputs the final judgment result.

[0090] The specific steps for in-depth analysis and judgment of real-time water bodies are:

[0091] S501. After the edge water body detection device issues an early warning, the spectral radiometer is used to collect the number of light rays in different bands of natural light in real time, and the light band with the largest number of light rays is extracted. The light type Rs corresponding to the light band with the largest number of light rays in real time and the most significant substance type Ws that matches are searched in the light-matter matching database; the edge water body detection device uses the single wavelength mode of the spectral imager to select the band with the most real-time light rays, and collects the water body image under the light Rs in real time and transmits it to the central cloud; after receiving the real-time water body image, the central cloud extracts the abnormal state characteristics of the most significant substance Ws; and calculates the abnormal state characteristic value of the most significant substance Ws in the real-time water body image to obtain the real-time abnormal state characteristic value Es;

[0092] S502, using the abnormal threshold of the most significant substance Ws in the real-time water body image to perform in-depth analysis and judgment on the real-time abnormal state feature value, when When the water body is abnormal, the warning signal of the edge water body detection device is transmitted to the user; When the edge water body detection device warns that there is no abnormality in the water body, a warning cancellation signal is sent to the edge water body detection device that warns, and the edge water body detection device stops warning.

[0093] Performing deep judgment after shallow judgment can further deepen the results of shallow judgment and prevent the one-sidedness of shallow judgment. The combination of shallow judgment and deep judgment can give full play to the advantages of both. Shallow judgment provides a wider potential edge range, avoiding the situation where deep judgment may miss some edges due to the complexity of the model; deep judgment screens and refines the results of shallow judgment, reducing the misjudgment caused by shallow judgment being easily disturbed by brightness and other factors.

[0094] A sensing and computing integrated water body analysis system based on edge computing, the sensing and computing integrated water body analysis system includes a data collection module, an edge detection network construction module, a light matching module, a threshold calculation module, a shallow layer judgment module and a depth judgment module;

[0095] The data collection module is used to collect image data of water body anomalies in history, and to collect types and bands of natural light in history;

[0096] The edge detection network construction module is used to analyze the environmental data of the water body, calculate the optimal placement point of the edge water body detection device, place the edge water body detection device in the water body, and build a distributed edge detection network;

[0097] The light matching module is used to analyze the light types and wavelengths in natural light to determine the most significant water substance under each light.

[0098] The threshold calculation module is used to obtain the abnormal state characteristics of the most significant substance in each light by using the image data of the water body under different light in history, and calculate the abnormal threshold;

[0099] The shallow layer judgment module is used to use the edge water body detection equipment to judge the shallow layer abnormality of the water body and issue an early warning;

[0100] The depth judgment module is used to use the central cloud to make depth anomaly judgment on the water body of the edge water body detection device that issues an early warning.

[0101] The edge detection network building module includes an optimal placement point unit and a network building unit;

[0102] The optimal placement point unit is used to analyze the water body environment data, divide the water body edge into broken lines, and calculate the optimal placement point using the coordinates of the connecting points of the broken lines;

[0103] The network construction unit is used to place edge water body detection equipment in the water body using the best placement point and calculate the required number of equipment.

[0104] The shallow layer judgment module includes an abnormal clarity calculation unit and a shallow layer judgment unit;

[0105] The abnormal clarity calculation unit is used to calculate the abnormal clarity by using the abnormal water body images in history;

[0106] The shallow layer judgment unit is used to collect real-time water body images and use abnormal clarity to make shallow layer judgments on the real-time water body.

[0107] The depth judgment module includes a real-time light judgment unit and a depth judgment unit;

[0108] The real-time light judging unit is used to judge the quantity of light in different bands in real time to obtain the real-time light type and the most significant substance type;

[0109] The depth judgment unit is used to judge the real-time depth of the water body by using the abnormal state characteristics of the most significant substance in real time, and to process the early warning of the edge water body detection equipment according to the judgment result.

[0110] Example: The water quality in a reservoir is judged, the reservoir environment is analyzed, the edge of the reservoir is divided into 4 broken lines, the best placement point is calculated as (20, 40), and a distributed edge detection network is constructed;

[0111] The light in natural light is judged, including the blue light band (450 - 495nm), green light band (495 - 570nm), and infrared band (780nm - 1mm). The most significant substances of each light are blue light-organic matter, green light-algae, and infrared light-mineral particles.

[0112] By analyzing the water body image, the state data of the most significant substances including substance density, substance quantity, substance brightness, etc. are obtained; by analyzing multiple state data, the abnormal state characteristics of each most significant substance are respectively determined as organic matter-water body projection, algae-algae brightness, and mineral particles-particle density; the abnormal thresholds of each most significant substance are calculated as water body projection-25, algae brightness-10, and mineral particle concentration-15;

[0113] When using edge water body detection equipment to conduct shallow layer analysis and determine that there may be abnormalities in the water body, an early warning is issued; the amount of light in each band of natural light in the water body environment is collected in real time, and it is determined that the infrared band (780nm - 1mm) has the largest number. A spectral camera is used to collect water body images in the infrared band, and the corresponding mineral particle concentration is calculated to be 10. The corresponding abnormal threshold is used to determine that there is no abnormality in the water body, and a command is issued to stop the early warning of the edge water body detection equipment.

[0114] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A water body analysis method based on edge computing and integrated sensing and computing, characterized by: The method comprises the following steps: S100, using remote sensing technology to collect water environment data, calculating the optimal placement point of edge water body detection equipment according to the water environment data, placing the edge water body detection equipment at the optimal placement point, and constructing a distributed edge detection network; the distributed edge detection network includes edge water body detection equipment and a central cloud; S200, using a multispectral camera to extract all light types and corresponding bands of natural light, collecting water samples, collecting historical material types in the water, collecting historical water images taken by a multispectral camera under different light, extracting the most significant material type in the water image under each light, matching the light type with the most significant material type, and generating a light-material matching database; S300, collecting water body images under each light when the water body is abnormal in history, and calculating the most significant substance abnormality threshold of the abnormal water body image under each light according to the most significant substance matched by each light in the light-substance matching database; The specific steps for calculating the most significant material anomaly threshold of abnormal water images under each light are: S301, collect water body images under each light when the water body is abnormal and normal in history, extract the most significant substance type in the image under each light, compare the most significant substances in the two images when the water body is in different states, and set the most significant substance matching state data of each light to be , It represents the 1st, 2nd, 3rd, ..., kth state data contained in the most significant substance matched by each light, where k is a positive integer; after the most significant substance in two states, the state with the largest difference between the most significant substance in the two states is extracted as the abnormal state feature; after comparison, the abnormal state feature of the most significant substance matched by each light is obtained as follows: , Indicates the most significant abnormal state characteristics of the material matched by the first, second, third, ..., mth light rays obtained by comparison; S302, after determining the most significant abnormal state characteristics of the substance matched by each light, extract the most significant abnormal state characteristics of the substance under different light in the image data of the historical water body anomaly, and calculate the abnormal threshold of the most significant substance under each light respectively, the formula is: ； In the formula, and represents the upper and lower limits of the most significant material abnormality threshold under each light, E represents the most significant material abnormal state feature value in the image under each light in the collected history, h represents the number of historical abnormal state feature values ​​collected, and Ep represents the average value of h abnormal state features in the collected history; after calculation, the abnormal threshold of the most significant material under each light is , represents the calculated anomaly threshold for the most significant material match for each ray; S400, collecting visual images of water body anomalies in history, and calculating abnormal clarity using the visual images of water body anomalies; using the visual sensor in the edge water body detection equipment to collect water body images in real time, obtaining the clarity of the real-time water body images, and using the abnormal clarity to perform shallow analysis and judgment on the clarity of the real-time water body images, and issue abnormal warnings; S500: After the edge water body detection device issues an early warning, the number of light rays in different bands of natural light is collected in real time, and the edge water body detection device sends the number of light rays and image data of different light rays to the central cloud; the central cloud selects the most significant abnormal threshold corresponding to the light type with the largest number to conduct in-depth analysis and judgment of the real-time water body, and outputs the final judgment result.

2. According to claim 1, a water body analysis method based on edge computing and sensing integration is characterized in that: The specific steps of constructing the distributed edge detection network in S100 are: S101, using remote sensing technology to collect water environment data, extracting the edge shape of the water body in the water environment, dividing the edge of the water body into n broken lines connected, assuming that the broken line constituting the edge of the water body is , represents the 1st, 2nd, 3rd, ..., nth broken lines that constitute the edge of the water body, where n is a positive integer; the coordinates of the connection points of each broken line in the edge of the water body are measured as follows: , It represents the 1st, 2nd, 3rd, ..., nth connection points in all the polylines at the edge of the water body. The optimal placement point of the edge water body detection equipment is calculated based on all the connection points. The formula is: ； In the formula, represents the calculated optimal placement point coordinates of the edge water body detection device, i represents any one of the n connection points, ; S102, after calculating the best placement point of the edge water body detection device, collect the farthest visible distance of the visual sensor in the edge water body detection device as L, place the first edge water body detection device at the best placement point, and assume that the detection range of each edge water body detection device is a circle with a radius of L; when the area of ​​the water body is larger than the detection area of ​​the edge water body detection device at the best placement point, place six identical edge water body detection devices tangent to each other around the first edge water body detection device, so that the first edge water body detection device and the six tangent edge water body detection devices form the vertices and center point of a regular hexagon, and calculate the number of edge water body detection devices that need to be placed in the water body, the formula is: ； In the formula, sm represents the number of edge water body detection devices that need to be placed in the water body, the calculation result is rounded down, Q represents the area of ​​the water body, and L represents the maximum detection distance of the edge water body detection device. After placing the edge water body detection device in the water body, each edge water body detection device is connected to the central cloud wirelessly to form a distributed edge detection network.

3. According to claim 1, a water body analysis method based on edge computing and sensing integration is characterized in that: The specific steps of generating the light-matter matching database in S200 are: S201, using a multi-spectral camera to extract all light types existing in natural light , represents the 1st, 2nd, 3rd, ..., mth light in natural light, where m is a positive integer; the corresponding wavelength band of each light in natural light measured by the spectrometer is , Indicates the wavelength bands of the 1st, 2nd, 3rd, ..., mth light rays in the measured natural light, In the above equation, q represents the minimum value of each light band, and p represents the maximum value of each light band; S202. Collect all the substances present in the water. , Indicates the 1st, 2nd, 3rd, .., jth substance in the collected water body, where j is a positive integer; collects image data of water bodies under different light in history, grays the image data using the Sobel operator, and calculates the gradient amplitude of the edge gray value of each object in the water body under different light. , Represents the calculated gradient amplitude of the edge grayscale value of the 1st, 2nd, 3rd, ..., jth material in the water body; the material corresponding to the maximum gradient amplitude in the image is judged to be the most significant material type, and the gradient amplitude of the material edge grayscale value in the water body image under each light is calculated in turn to obtain the most significant material in the image under each light, and the light type and material type are matched as , Represents the matching results of the 1st, 2nd, 3rd, ..., mth light and matter, and uses the matching results to generate a light-matter matching database.

4. According to claim 2, a water body analysis method based on edge computing and sensing integration is characterized in that: The specific steps of using abnormal clarity to perform shallow analysis and judgment on the clarity of the real-time water body image in S400 are: S401, collect visual images of water body anomalies in history, grayscale the collected visual images of water body anomalies, and calculate the grayscale value variance of each collected historical visual image, the formula is: , in the formula, represents the grayscale value variance of the historical visual image, d represents the number of grayscale values ​​in the collected historical visual images, H represents the grayscale value in the collected historical visual images, Represents the average grayscale value in the historical visual image; the grayscale value variance of each visual image in the collected history is obtained by calculation; the grayscale value variance is used to represent the clarity of each visual image; S402. Abnormal clarity is calculated using the clarity of each visual image in history, and the formula is: , in the formula, Indicates the abnormal clarity of the calculated water body image, represents the average clarity of all visual images in history, represents the standard deviation of clarity of all visual images in history; S403, using the visual sensor in the edge water body detection device to collect water body images in real time, and obtaining the clarity of the real-time water body image as Qs, and using the calculated abnormal clarity to make a shallow judgment on the clarity of the real-time water body image. When the water body detection equipment at the edge determines that there is an abnormal risk in the water body, it will issue an early warning; At this time, the edge water body detection equipment determines that the water body is normal.

5. The method for water body analysis based on edge computing and sensing integration according to claim 4 is characterized in that: The specific steps of performing depth analysis and judgment on the real-time water body in S500 are: S501. After the edge water body detection device issues an early warning, the spectral radiometer is used to collect the number of light rays in different bands of natural light in real time, and the light band with the largest number of light rays is extracted. The light type Rs corresponding to the light band with the largest number of light rays in real time and the most significant substance type Ws that matches are searched in the light-matter matching database; the edge water body detection device uses the single wavelength mode of the spectral imager to select the band with the most real-time light rays, and collects the water body image under the light Rs in real time and transmits it to the central cloud; after receiving the real-time water body image, the central cloud extracts the abnormal state characteristics of the most significant substance Ws; The abnormal state characteristic value of the most significant substance Ws in the real-time water body image is calculated to obtain the real-time abnormal state characteristic value Es; S502, using the abnormal threshold of the most significant substance Ws in the real-time water body image to perform in-depth analysis and judgment on the real-time abnormal state feature value, when When the water body is abnormal, the warning signal of the edge water body detection device is transmitted to the user; When the edge water body detection device warns that there is no abnormality in the water body, a warning cancellation signal is sent to the edge water body detection device that warns, and the edge water body detection device stops warning.

6. A water body analysis system based on edge computing and sensing and computing integration using a water body analysis method based on edge computing and sensing and computing integration according to any one of claims 1 to 5, characterized in that: The sensing and computing integrated water body analysis system includes a data collection module, an edge detection network construction module, a light matching module, a threshold calculation module, a shallow layer judgment module, and a depth judgment module; The data collection module is used to collect image data of water body anomalies in history, and to collect types and bands of natural light in history; The edge detection network construction module is used to analyze the environmental data of the water body, calculate the optimal placement point of the edge water body detection device, place the edge water body detection device in the water body, and build a distributed edge detection network; The light matching module is used to analyze the light types and wavelengths in natural light to determine the most significant water substance under each light. The threshold calculation module is used to obtain the abnormal state characteristics of the most significant substance in each light by using the image data of the water body under different light in history, and calculate the abnormal threshold; The shallow layer judgment module is used to use the edge water body detection equipment to judge the shallow layer abnormality of the water body and issue an early warning; The depth judgment module is used to use the central cloud to make depth anomaly judgment on the water body of the edge water body detection device that issues an early warning.

7. The water body analysis system based on edge computing and sensing integration according to claim 6 is characterized by: The edge detection network construction module includes an optimal placement point unit and a network construction unit; The optimal placement point unit is used to analyze the water body environment data, divide the water body edge into broken lines, and calculate the optimal placement point using the coordinates of the connecting points of the broken lines; The network construction unit is used to place edge water body detection equipment in the water body using the best placement point and calculate the required number of equipment.

8. The water body analysis system based on edge computing and sensing integration according to claim 6 is characterized by: The shallow layer judgment module includes an abnormal clarity calculation unit and a shallow layer judgment unit; The abnormal clarity calculation unit is used to calculate the abnormal clarity by using the abnormal water body images in history; The shallow layer judgment unit is used to collect real-time water body images and use abnormal clarity to make shallow layer judgments on the real-time water body.

9. The water body analysis system based on edge computing and sensing integration according to claim 6 is characterized by: The depth judgment module includes a real-time light judgment unit and a depth judgment unit; The real-time light judging unit is used to judge the quantity of light in different bands in real time to obtain the real-time light type and the most significant substance type; The depth judgment unit is used to judge the real-time depth of the water body by using the abnormal state characteristics of the most significant substance in real time, and to process the early warning of the edge water body detection equipment according to the judgment result.

Images