An intelligent analysis system based on liquid detection

Abstract

The present invention relates to an intelligent analysis system based on liquid detection, and relates to the technical field of liquid detection, comprising: a database establishment module, used for establishing a sample liquid database, and storing liquid images of multiple sample liquids; a liquid detection module, comprising multiple liquid detection units respectively arranged at multiple positions of a target area, the liquid detection unit comprising a liquid detection device and an edge computing device, the liquid detection device being used for sampling a liquid to be detected, and obtaining liquid characteristics of the liquid to be detected, the edge computing device being used for establishing a liquid image of the liquid to be detected based on the liquid characteristics of the liquid to be detected, and identifying the liquid to be detected based on the liquid images of multiple sample liquids, and uploading the identification result of the liquid to be detected; a cloud processing module, used for synchronizing the identification result of the liquid to be detected uploaded by the edge computing device to other liquid detection units, and having the advantages of automatically completing liquid detection and improving the efficiency of water environment pollution detection.

Description

Technical Field

[0001] The present invention relates to the technical field of liquid detection, and in particular to an intelligent analysis system based on liquid detection. Background Art

[0002] In recent years, with the continuous development of industrialization, regional economy and the acceleration of urbanization, water pollution has become increasingly prominent, seriously affecting the drinking water safety of urban and rural residents, and further exacerbating the urban water shortage. Ecological rivers have good overall landscape effects, reasonable ecosystem organizational structure and good operation functions; they can maintain elasticity, stability and a certain degree of self-recovery ability for long-term or sudden disturbances; the overall functions of rivers are diverse and complex, and can meet the reasonable target requirements of all beneficiaries. However, with economic development, the amount of pollutants discharged in life and production has also increased exponentially. Due to the large-scale discharge of pollutants and unreasonable land use methods, rivers, wetlands, lakes and other water bodies have been seriously polluted and ecologically damaged, and the ecological functions of rivers, wetlands and lakes have been lost or cannot be performed normally. Ecological rivers are constructed by adding ecological principles to traditional river construction and regulation, and ecologically designing projects according to the current status and functions of rivers, to construct river water ecosystems and riverbank ecosystems that meet the ecological characteristics of the basin and region. Although the design of ecological rivers is reasonable, there is still a possibility of pollution. After being polluted, the polluted river sewage needs to be treated to prevent the pollution from becoming more and more serious. In addition to treating sewage, it is also necessary to trace the source of pollution and control pollution emissions in a timely manner. However, the time of pollution emissions is uncontrollable, and there may be other pollution discharge points downstream of the polluted river.

[0003] The existing conventional monitoring method for water environment pollution is still through manual monitoring, combined with some auxiliary detection equipment, such as carrying a water quality detector to the corresponding part to carry out water environment characteristic data detection work. However, it is limited by manual labor and is not conducive to the rapid completion of water environment characteristic data detection information collection.

[0004] Therefore, it is necessary to provide an intelligent analysis system based on liquid detection to automatically complete liquid detection and improve the efficiency of water environment pollution detection. Summary of the invention

[0005] The present invention provides an intelligent analysis system based on liquid detection, comprising: a database establishment module, used to establish a sample liquid database, wherein the sample liquid database is used to store liquid images of multiple sample liquids; a liquid detection module, comprising multiple liquid detection units respectively arranged at multiple positions of a target area, wherein the liquid detection unit comprises a liquid detection device and an edge computing device, the liquid detection device is used to sample the liquid to be detected and obtain the liquid characteristics of the liquid to be detected, the edge computing device is used to establish the liquid image of the liquid to be detected based on the liquid characteristics of the liquid to be detected, and based on the liquid images of the multiple sample liquids, identify the liquid to be detected and upload the identification result of the liquid to be detected; a cloud processing module, used to synchronize the identification result of the liquid to be detected uploaded by the edge computing device to other liquid detection units.

[0006] Furthermore, the liquid detection device includes a feature acquisition component for acquiring liquid features of the liquid to be detected, wherein the feature acquisition component includes an ultrasonic detection device, an image acquisition device, a PH detection device and a spectrometer, the ultrasonic detection device is used to acquire turbidity features of the liquid to be detected, the image acquisition device is used to acquire image features of the liquid to be detected, the PH detection device is used to acquire PH features of the liquid to be detected, and the spectrometer is used to acquire a spectral characteristic curve of the liquid to be detected.

[0007] Furthermore, the liquid detection module obtains liquid characteristics of the liquid to be detected, including: obtaining image information of the liquid to be detected at multiple time points through an image acquisition device; determining the precipitation change parameters of the liquid to be detected based on the image information of the liquid to be detected obtained by the image acquisition device at two adjacent time points; when the precipitation change parameters of the liquid to be detected meet preset conditions, controlling the image acquisition device to obtain the image characteristics of the liquid to be detected, controlling the ultrasonic detection device to obtain the turbidity characteristics of the liquid to be detected, controlling the PH detection device to obtain the PH characteristics of the liquid to be detected, and controlling the spectrometer to obtain the spectral characteristic curve of the liquid to be detected.

[0008] Furthermore, the edge computing device identifies the liquid to be detected based on the liquid portraits of the multiple sample liquids, including: obtaining building information of the target area, wherein the building information includes building type and building location; determining priority values ​​of the liquid portraits of the multiple sample liquids based on the building information of the target area; and identifying the liquid to be detected based on the liquid portrait of the liquid to be detected and the liquid portraits of the multiple sample liquids according to the priority values ​​of the liquid portraits of the multiple sample liquids.

[0009] Furthermore, the edge computing device identifies the liquid to be detected based on the liquid image of the liquid to be detected and the liquid image of the multiple sample liquids according to the priority values ​​of the liquid image of the multiple sample liquids, including: calculating the liquid image similarities between the liquid image of the liquid to be detected and the liquid image of the multiple sample liquids in turn according to the priority values ​​of the liquid image of the multiple sample liquids, and judging whether there is a target sample liquid in the multiple sample liquids; if the target sample liquid does not exist in the multiple sample liquids, performing component detection on the liquid to be detected, obtaining the component characteristics of the liquid to be detected, and identifying the liquid to be detected.

[0010] Furthermore, the edge computing device calculates the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids, including: calculating the image feature similarity based on the image features included in the liquid image of the liquid to be detected and the image features included in the liquid image of the sample liquid; calculating the turbidity feature difference based on the turbidity feature included in the liquid image of the liquid to be detected and the turbidity feature included in the liquid image of the sample liquid; calculating the PH feature difference based on the PH feature included in the liquid image of the liquid to be detected and the PH feature included in the liquid image of the sample liquid; calculating the spectral characteristic curve similarity based on the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral characteristic curve included in the liquid image of the sample liquid; calculating the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids based on the image feature similarity, the turbidity feature difference, the PH feature difference and the spectral characteristic curve similarity.

[0011] Furthermore, the edge computing device calculates the similarity of spectral characteristic curves based on the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral characteristic curve included in the liquid image of the sample liquid, including: determining the spectral reflectance fluctuation characteristic sequence of the liquid to be detected based on the spectral characteristic curve included in the liquid image of the liquid to be detected; calculating the similarity of the spectral characteristic curves based on the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the sample liquid.

[0012] Furthermore, the liquid detection device also includes a component detection component, wherein the component detection component includes a first detection chamber and a second detection chamber, the first detection chamber is provided with a metal component detection device, and the second detection chamber is provided with a non-metal component detection device.

[0013] Furthermore, the cloud processing module synchronizes the recognition result of the liquid to be detected uploaded by the edge computing device to other liquid detection units: based on the water flow direction characteristics of the target area and the position information of each of the liquid detection units, establishes a detection unit association map corresponding to the multiple liquid detection units; based on the detection unit association map corresponding to the multiple liquid detection units, determines at least one associated liquid detection unit corresponding to the liquid detection unit; synchronizes the recognition result of the liquid to be detected uploaded by the edge computing device to at least one associated liquid detection unit corresponding to the liquid detection unit.

[0014] Furthermore, the edge computing device of the associated liquid detection unit is also used to: adjust the priority values ​​of the liquid images of the multiple sample liquids based on the identification results of the liquid to be detected uploaded by the edge computing device; and identify the liquid to be detected based on the liquid image of the liquid to be detected and the liquid image of the multiple sample liquids according to the adjusted priority values ​​of the liquid images of the multiple sample liquids.

[0015] Compared with the prior art, the intelligent analysis system based on liquid detection provided in this specification has at least the following beneficial effects:

[0016] 1. By establishing a sample liquid database, data support is provided for the subsequent identification of the liquid to be detected. Furthermore, multiple liquid detection units at multiple locations in the target area automatically complete the liquid detection at multiple locations in the target area, thereby improving the efficiency of water environment pollution detection;

[0017] 2. By setting the first detection chamber and the second detection chamber, the non-metallic components and the metal components of the liquid to be detected can be detected simultaneously, thereby improving the liquid detection efficiency of the liquid to be detected;

[0018] 3. When the precipitation change parameters of the liquid to be detected meet the preset conditions, the state of the liquid to be detected is relatively stable, and the image features, turbidity features, pH features and spectral characteristic curves of the liquid to be detected are more accurate;

[0019] 4. First determine the priority values ​​of the liquid images of the multiple sample liquids, and then identify the liquid to be detected based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids according to the priority values ​​of the liquid images of the multiple sample liquids, so that the liquid image of the sample liquid with a greater possibility of multiple sample liquids in the target area is compared with the liquid image of the liquid to be detected, thereby improving the recognition efficiency of the liquid to be detected;

[0020] 5. From the four dimensions of image feature similarity, turbidity feature difference, pH feature difference and spectral characteristic curve similarity, the calculated liquid image similarity between the liquid image of the liquid to be tested and the liquid image of the sample liquid is made more accurate;

[0021] 6. When a liquid detection unit detects a certain liquid, the associated liquid detection unit of the liquid detection unit has a higher possibility of detecting the liquid. Therefore, the priority values ​​of the liquid images of multiple sample liquids are adjusted to make the liquid recognition efficiency of the associated liquid detection unit corresponding to the liquid detection unit higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0023] Figure 1 It is a module diagram of an intelligent analysis system based on liquid detection shown in an embodiment of the present application;

[0024] Figure 2 is a structural diagram of a liquid detection unit shown in an embodiment of the present application;

[0025] Figure 3 is a flow chart of obtaining liquid characteristics of a liquid to be detected shown in an embodiment of the present application;

[0026] Figure 4 It is a flow chart of calculating the similarity of liquid images shown in one embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments.

[0028] Figure 1 is a module diagram of an intelligent analysis system based on liquid detection shown in an embodiment of the present application, such as Figure 1 As shown, an intelligent analysis system based on liquid detection may include a database establishment module, a liquid detection module and a cloud processing module.

[0029] The database establishment module is used to establish a sample liquid database.

[0030] The sample liquid database is used to store liquid images of various sample liquids, which may include image features, turbidity features, pH features, and spectral characteristic curves.

[0031] The liquid detection module may include a plurality of liquid detection units respectively disposed at a plurality of positions of the target area.

[0032] Specifically, the liquid detection module can determine the possibility of sewage discharge in different sub-areas in the target area based on the building information of the target area, and determine the locations of multiple liquid detection units based on the possibility of sewage discharge in different sub-areas in the target area, wherein the building information includes the building type and the building location. For example, at least one liquid detection unit is set in a sub-area where the possibility of sewage discharge is greater than a preset possibility threshold.

[0033] The liquid detection unit includes a liquid detection device and an edge computing device. The liquid detection device is used to sample the liquid to be detected and obtain the liquid characteristics of the liquid to be detected. The edge computing device is used to establish a liquid image of the liquid to be detected based on the liquid characteristics of the liquid to be detected, and identify the liquid to be detected based on the liquid images of multiple sample liquids, and upload the identification result of the liquid to be detected.

[0034] Edge computing devices may include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc.

[0035] Figure 2 is a structural diagram of a liquid detection unit shown in an embodiment of the present application, such as Figure 2 As shown, in some embodiments, the liquid detection device includes a feature acquisition component for acquiring liquid features of the liquid to be detected, wherein the feature acquisition component includes an ultrasonic detection device, an image acquisition device, a pH detection device and a spectrometer, the ultrasonic detection device is used to acquire turbidity features of the liquid to be detected, the image acquisition device is used to acquire image features of the liquid to be detected, the pH detection device is used to acquire pH features of the liquid to be detected, and the spectrometer is used to acquire a spectral characteristic curve of the liquid to be detected.

[0036] like Figure 2 As shown, in some embodiments, the liquid detection device also includes a component detection component, wherein the component detection component includes a first detection chamber and a second detection chamber, a metal component detection device is disposed in the first detection chamber, and a non-metal component detection device is disposed in the second detection chamber.

[0037] It can be understood that by providing the first detection chamber and the second detection chamber, the non-metallic components and the metallic components of the liquid to be detected can be detected synchronously, thereby improving the liquid detection efficiency of the liquid to be detected.

[0038] Figure 3 FIG. 1 is a flow chart of obtaining liquid characteristics of a liquid to be detected in an embodiment of the present application. Figure 3 As shown, in some embodiments, the liquid detection module obtains liquid characteristics of the liquid to be detected, including:

[0039] Acquire image information of the liquid to be detected at multiple time points by using an image acquisition device;

[0040] Determine the precipitation change parameter of the liquid to be detected based on the image information of the liquid to be detected acquired by the image acquisition device at two adjacent time points;

[0041] When the sedimentation change parameters of the liquid to be detected meet the preset conditions, the image acquisition device is controlled to obtain the image characteristics of the liquid to be detected, the ultrasonic detection device is controlled to obtain the turbidity characteristics of the liquid to be detected, the pH detection device is controlled to obtain the pH characteristics of the liquid to be detected, and the spectrometer is controlled to obtain the spectral characteristic curve of the liquid to be detected.

[0042] Specifically, for each time point, the edge computing device determines the sediment thickness of the liquid to be detected at that time point based on the image information of the liquid to be detected at that time point, and divides multiple time points into multiple time periods, and determines the sediment change parameters corresponding to the time period based on the sediment thickness of the liquid to be detected at multiple time points included in each time period.

[0043] For example, the precipitation change parameter can be calculated based on the following formula:

[0044]

[0045] in, The liquid to be tested is i The precipitation change parameters of each time period, The liquid to be tested is i The sediment thickness at the time point t1 included in the time period.

[0046] The liquid to be tested is i The sedimentation change parameter of the time period is less than the preset sedimentation change parameter threshold and the liquid to be detected is i The precipitation change parameter of the time period is less than that of the T i-1 When the precipitation change parameters of the liquid to be detected meet the preset conditions, the precipitation change parameters of the liquid to be detected meet the preset conditions.

[0047] It can be understood that when the precipitation change parameters of the liquid to be detected meet the preset conditions, the state of the liquid to be detected is relatively stable, and the image features, turbidity features, pH features and spectral characteristic curves of the liquid to be detected are more accurate.

[0048] In some embodiments, the edge computing device identifies the liquid to be detected based on liquid images of multiple sample liquids, including:

[0049] Acquire building information of a target area, wherein the building information includes building type and building location;

[0050] determining priority values ​​of liquid images of a plurality of sample liquids based on building information of a target area;

[0051] According to the priority values ​​of the liquid images of the multiple sample liquids, the liquid to be detected is identified based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids.

[0052] Specifically, the edge computing device can determine the possibility that there may be multiple sample liquids in the target area based on the building information of the target area, and the sample liquid with a greater possibility has a greater priority value of the liquid image of the sample liquid. The multiple sample liquids are sorted according to the priority value, and the liquid to be detected is identified based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids according to the sorting result.

[0053] In some embodiments, the edge computing device identifies the liquid to be detected based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids according to the priority values ​​of the liquid images of the multiple sample liquids, including:

[0054] According to the priority values ​​of the liquid images of the multiple sample liquids, the liquid image similarities between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids are calculated in sequence to determine whether the target sample liquid exists in the multiple sample liquids;

[0055] If the target sample liquid does not exist in the multiple sample liquids, a component detection is performed on the liquid to be detected to obtain the component characteristics of the liquid to be detected and identify the liquid to be detected.

[0056] Specifically, in order from large to small priority values, first calculate the liquid image similarity between the liquid image of the liquid to be detected and the liquid image of the sample liquid with a larger priority value, and determine whether the liquid image similarity is greater than a preset liquid image similarity threshold. If so, the sample liquid is the target sample liquid, and the composition information of the target sample liquid is used as the identification result of the liquid to be detected. If not, calculate the liquid image similarity between the liquid image of the next sample liquid and the liquid image of the liquid to be detected.

[0057] If the target sample liquid does not exist among the multiple sample liquids, the metal component in the liquid to be detected is detected by the metal component detection equipment set in the first detection chamber, and the non-metal component in the liquid to be detected is detected by the non-metal component detection equipment set in the second detection chamber as the identification result of the liquid to be detected.

[0058] It can be understood that the priority values ​​of the liquid images of multiple sample liquids are first determined, and then the liquid to be detected is identified based on the liquid image of the liquid to be detected and the liquid images of multiple sample liquids according to the priority values ​​of the liquid images of the multiple sample liquids, so that the liquid image of the sample liquid with a higher possibility that there may be multiple sample liquids in the target area is compared with the liquid image of the liquid to be detected, thereby improving the recognition efficiency of the liquid to be detected.

[0059] Figure 4 is a flow chart of calculating the similarity of liquid images shown in an embodiment of the present application, such as Figure 4 As shown, in some embodiments, the edge computing device calculates the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of multiple sample liquids, including:

[0060] Calculating image feature similarity based on image features included in the liquid image of the liquid to be detected and image features included in the liquid image of the sample liquid;

[0061] Calculating a turbidity feature difference based on the turbidity feature included in the liquid image of the liquid to be detected and the turbidity feature included in the liquid image of the sample liquid;

[0062] Calculating a pH characteristic difference based on the pH characteristic included in the liquid image of the liquid to be detected and the pH characteristic included in the liquid image of the sample liquid;

[0063] Calculating the similarity of the spectral characteristic curves based on the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral characteristic curve included in the liquid image of the sample liquid;

[0064] Based on the image feature similarity, turbidity feature difference, pH feature difference and spectral characteristic curve similarity, the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of multiple sample liquids is calculated.

[0065] For example, the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of multiple sample liquids can be calculated based on the following formula:

[0066]

[0067] Among them, S h is the liquid image similarity between the liquid image of the liquid to be detected and the liquid image of the h-th sample liquid, S (image,h) is the similarity between the image of the liquid to be tested and the image feature of the h-th sample liquid, T current is the turbidity of the liquid image of the liquid to be tested, T h is the liquid image of the h-th sample liquid, including turbidity, pH current The pH value of the liquid image to be tested, pHh is the pH value of the liquid image of the h-th sample liquid, S (curve,h) is the similarity of the spectral characteristic curve between the liquid image of the liquid to be tested and the liquid image of the hth sample liquid, a 11 、a 12 、a 13 and a 14 All are preset weights, and M1 and M2 are preset parameters.

[0068] It can be understood that the similarity between the calculated liquid image of the liquid to be detected and the liquid image of the sample liquid is more accurate from the four dimensions of image feature similarity, turbidity feature difference, pH feature difference and spectral characteristic curve similarity.

[0069] In some embodiments, the edge computing device calculates the similarity of the spectral characteristic curves based on the spectral characteristic curves included in the liquid image of the liquid to be detected and the spectral characteristic curves included in the liquid image of the sample liquid, including:

[0070] Determining a spectral reflectance fluctuation characteristic sequence of the liquid to be detected based on a spectral characteristic curve included in the liquid image of the liquid to be detected;

[0071] Based on the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the sample liquid, the similarity of the spectral characteristic curve is calculated.

[0072] Specifically, the spectral characteristic curve can represent the curve of the spectral reflectance of the liquid changing with the wavelength. The spectral characteristic curve can be split into multiple curve segments, and for each curve segment, the spectral reflectance fluctuation parameter of the curve segment is calculated, and the spectral reflectance fluctuation characteristic sequence can be composed of the spectral reflectance fluctuation parameters of multiple curve segments.

[0073] For example, the spectral reflectance fluctuation parameter of a curve segment can be calculated based on the following formula:

[0074]

[0075] in, is the wavelength w x To wavelength w y The spectral reflectance fluctuation parameter of the curve segment between The wavelength w is the wavelength of the liquid x spectral reflectance.

[0076] The cloud processing module can be used to synchronize the identification results of the liquid to be detected uploaded by the edge computing device to other liquid detection units.

[0077] Specifically include:

[0078] Based on the water flow characteristics of the target area and the position information of each liquid detection unit, a detection unit association map corresponding to the multiple liquid detection units is established, wherein the detection unit association map can represent the upstream and downstream relationship of the multiple liquid detection units;

[0079] Determine at least one associated liquid detection unit corresponding to the liquid detection unit based on the detection unit association maps corresponding to the multiple liquid detection units, wherein the at least one associated liquid detection unit corresponding to the liquid detection unit may be a liquid detection unit upstream of the liquid detection unit and / or a liquid detection unit downstream of the liquid detection unit;

[0080] The identification result of the liquid to be detected uploaded by the edge computing device is synchronized to at least one associated liquid detection unit corresponding to the liquid detection unit.

[0081] In some embodiments, the edge computing device associated with the liquid detection unit is further used to:

[0082] Based on the recognition result of the liquid to be detected uploaded by the edge computing device, the priority values ​​of the liquid images of the multiple sample liquids are adjusted. For example, when the target sample liquid exists among the multiple sample liquids, the priority value of the liquid image corresponding to the target sample liquid is adjusted to the maximum. When the target sample liquid does not exist among the multiple sample liquids, the liquid image of the liquid to be detected can be added according to the recognition result of the liquid to be detected, and the liquid image of the liquid to be detected can be adjusted to the maximum.

[0083] According to the adjusted priority values ​​of the liquid images of the multiple sample liquids, the liquid to be detected is identified based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids.

[0084] It can be understood that when a liquid detection unit detects a certain liquid, the associated liquid detection unit of the liquid detection unit has a higher possibility of detecting this liquid. Therefore, the priority values ​​of the liquid images of multiple sample liquids are adjusted to make the liquid recognition efficiency of the associated liquid detection unit corresponding to the liquid detection unit higher.

[0085] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. An intelligent analysis system based on liquid detection, characterized in that: include: A database establishment module, used to establish a sample liquid database, wherein the sample liquid database is used to store liquid images of multiple sample liquids; A liquid detection module, comprising a plurality of liquid detection units respectively arranged at a plurality of positions of a target area, wherein the liquid detection unit comprises a liquid detection device and an edge computing device, the liquid detection device is used to sample a liquid to be detected and obtain liquid characteristics of the liquid to be detected, the edge computing device is used to establish a liquid image of the liquid to be detected based on the liquid characteristics of the liquid to be detected, and to identify the liquid to be detected based on the liquid images of the plurality of sample liquids, and upload the identification result of the liquid to be detected; A cloud processing module, used for synchronizing the identification result of the liquid to be detected uploaded by the edge computing device to other liquid detection units; The liquid detection device includes a feature acquisition component, which is used to acquire the liquid features of the liquid to be detected, wherein the feature acquisition component includes an ultrasonic detection device, an image acquisition device, a pH detection device and a spectrometer, the ultrasonic detection device is used to acquire the turbidity features of the liquid to be detected, the image acquisition device is used to acquire the image features of the liquid to be detected, the pH detection device is used to acquire the pH features of the liquid to be detected, and the spectrometer is used to acquire the spectral characteristic curve of the liquid to be detected; The liquid detection module obtains the liquid characteristics of the liquid to be detected, including: Acquiring image information of the liquid to be detected at multiple time points by an image acquisition device; Determining a sedimentation change parameter of the liquid to be detected based on image information of the liquid to be detected acquired by the image acquisition device at two adjacent time points; When the precipitation change parameter of the liquid to be detected meets the preset conditions, the image acquisition device is controlled to obtain the image characteristics of the liquid to be detected, the ultrasonic detection device is controlled to obtain the turbidity characteristics of the liquid to be detected, the pH detection device is controlled to obtain the pH characteristics of the liquid to be detected, and the spectrometer is controlled to obtain the spectral characteristic curve of the liquid to be detected; The edge computing device identifies the liquid to be detected based on the liquid images of the multiple sample liquids, including: Acquire building information of the target area, wherein the building information includes building type and building location; Based on the building information of the target area, the possibility of the presence of multiple sample liquids in the target area is judged, the sample liquid with a greater possibility has a greater priority value of the liquid image of the sample liquid, the multiple sample liquids are sorted according to the priority value, and the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids are identified according to the sorting result; The edge computing device identifies the liquid to be detected based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids according to the sorting result, including: According to the priority values ​​of the liquid images of the multiple sample liquids, the liquid image similarities between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids are calculated in sequence to determine whether the target sample liquid exists in the multiple sample liquids; If the target sample liquid does not exist in the multiple sample liquids, component detection is performed on the liquid to be detected to obtain component characteristics of the liquid to be detected and identify the liquid to be detected.

2. The intelligent analysis system based on liquid detection according to claim 1, characterized in that: The edge computing device calculates the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids, including: Calculating image feature similarity based on image features included in the liquid image of the liquid to be detected and image features included in the liquid image of the sample liquid; Calculating a turbidity feature difference based on the turbidity feature included in the liquid image of the liquid to be detected and the turbidity feature included in the liquid image of the sample liquid; Calculating a PH feature difference based on the PH feature included in the liquid image of the liquid to be detected and the PH feature included in the liquid image of the sample liquid; Calculating the similarity of the spectral characteristic curves based on the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral characteristic curve included in the liquid image of the sample liquid; Based on the image feature similarity, turbidity feature difference, pH feature difference and spectral characteristic curve similarity, the liquid image similarity between the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids is calculated.

3. The intelligent analysis system based on liquid detection according to claim 2, characterized in that: The edge computing device calculates the similarity of the spectral characteristic curves based on the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral characteristic curve included in the liquid image of the sample liquid, including: Determining a spectral reflectance fluctuation characteristic sequence of the liquid to be detected based on a spectral characteristic curve included in the liquid image of the liquid to be detected; Based on the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the liquid to be detected and the spectral reflectance fluctuation characteristic sequence of the spectral characteristic curve included in the liquid image of the sample liquid, the similarity of the spectral characteristic curve is calculated.

4. The intelligent analysis system based on liquid detection according to claim 3, characterized in that: The liquid detection device also includes a component detection component, wherein the component detection component includes a first detection chamber and a second detection chamber, a metal component detection device is arranged in the first detection chamber, and a non-metal component detection device is arranged in the second detection chamber.

5. The intelligent analysis system based on liquid detection according to claim 4, characterized in that: The cloud processing module synchronizes the identification result of the liquid to be detected uploaded by the edge computing device to other liquid detection units, including: Based on the water flow direction characteristics of the target area and the position information of each of the liquid detection units, establishing a detection unit association map corresponding to the multiple liquid detection units; Determine at least one associated liquid detection unit corresponding to the liquid detection unit based on the detection unit association maps corresponding to the multiple liquid detection units; The identification result of the liquid to be detected uploaded by the edge computing device is synchronized to at least one associated liquid detection unit corresponding to the liquid detection unit.

6. The intelligent analysis system based on liquid detection according to claim 5, characterized in that: The edge computing device associated with the liquid detection unit is also used for: Adjusting the priority values ​​of the liquid images of the plurality of sample liquids based on the identification results of the liquid to be detected uploaded by the edge computing device; According to the adjusted priority values ​​of the liquid images of the multiple sample liquids, the liquid to be detected is identified based on the liquid image of the liquid to be detected and the liquid images of the multiple sample liquids.

Images