Artificial intelligence-based tailing wastewater analysis method and system

Abstract

The invention provides a tailing wastewater analysis method and system based on artificial intelligence, and the method comprises the steps: carrying out the important content analysis and tailing component construction of to-be-analyzed tailing wastewater data containing a target item, and obtaining the analysis important content information and original tailing components corresponding to the target item; the original tailing components comprise analysis important contents and correlation contents among the analysis important contents; optimizing each piece of analysis important content information in the original tailing components of the target item to obtain a tailing component optimization result of the target item; and determining tailing pollutant data of the target matter based on the tailing component optimization result of the target matter. According to the method, on the premise that the target is partially interfered or the target tailing pollutant data is fuzzy, the original tailing components of the target matter are optimized, important content which is about to be interfered and is not analyzed is analyzed, the complete tailing components of the target matter are obtained, and the analysis accuracy of the tailing pollutant data of the target matter is improved.

Description

Technical Field

[0001] The present application relates to the technical field of tailings wastewater analysis, and in particular, to a tailings wastewater analysis method and system based on artificial intelligence. Background Art

[0002] Tailings wastewater is a type of wastewater containing a large amount of pollutants or toxic substances, which not only affects the soil, but may also endanger people's lives and safety.

[0003] Therefore, tailings wastewater needs to be treated, but the specific pollutants or types and contents of toxic substances in tailings wastewater are currently unknown, and there may be inadequate treatment during treatment. Therefore, tailings wastewater needs to be analyzed. The current analysis technology requires the collection of wastewater and sending it to the laboratory for analysis, and then relevant technical personnel are required to make judgments. This wastes a lot of time and there is also the problem of inaccurate analysis. Therefore, an artificial intelligence-based tailings wastewater analysis method is urgently needed to improve the above technical problems. Summary of the invention

[0004] In order to improve the technical problems existing in the relevant technologies, the present application provides a tailings wastewater analysis method and system based on artificial intelligence.

[0005] In a first aspect, a tailings wastewater analysis method based on artificial intelligence is provided, and the tailings wastewater analysis method based on artificial intelligence comprises: Performing important content analysis and tailings composition construction on the tailings wastewater data to be analyzed that covers the target items, and obtaining the important analysis content information and original tailings composition corresponding to the target items; the original tailings composition includes important analysis contents and related contents between the important analysis contents; Optimizing each of the important analytical contents in the original tailings composition of the target item to obtain an optimized result of the tailings composition of the target item; the optimized result of the tailings composition includes all designated important contents and associated contents between the designated important contents, and the number of the important analytical contents covered in the original tailings composition is not greater than the number of the designated important contents covered in the optimized result of the tailings composition; Based on the tailings composition optimization result of the target item, the tailings pollutant data of the target item is determined.

[0006] In the present application, the tailings pollutant data of the target item is determined based on the tailings composition optimization result of the target item, including: Extracting descriptive knowledge of the tailings composition optimization result of the target item to obtain descriptive knowledge of the tailings pollutant composition of the target item; Extracting descriptive knowledge of tailings pollutant data covering the target items to obtain descriptive knowledge of tailings pollutant data of the target items; Performing description knowledge splicing on the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain spliced ​​description knowledge of the target item; Based on the spliced ​​description knowledge of the target item, the tailings pollutant data of the target item is determined.

[0007] It can be understood that when the tailings composition optimization result is based on the target item, the problem of inaccurate extraction of description knowledge is improved, so that the tailings pollutant data of the target item can be accurately determined.

[0008] In the present application, the optimization of each of the important analysis information in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item includes: The original tailings components are sequentially expanded and compressed using compression and expansion threads to generate tailings component optimization results for the target item.

[0009] It can be understood that when optimizing the important content information of each analysis in the original tailings composition of the target item, the problem of inaccurate compression and expansion threads is avoided, so that the tailings composition optimization result of the target item can be accurately obtained.

[0010] In this application, the training method of compressing and expanding threads includes: Obtaining a first training contaminant data set, the first training contaminant data set comprising first example tailings contaminant data having a first training target, the first training target being associated with annotated tailings composition and a tailings contaminant data catalog; The compression and expansion threads are used to sequentially expand and compress the annotated tailings components of the first example tailings pollutant data to generate a regression analysis tailings component of the first training target; the regression analysis tailings component includes all designated important contents of the first training target and associated contents between the designated important contents; Based on the regression analysis tailings composition of the first training target, obtaining the regression analysis tailings pollutant data of the first training target; The compression and expansion threads are trained based on the deviation value between the regression analysis tailings pollutant data and the tailings pollutant data directory corresponding to the first training target.

[0011] It is understandable that the compression and expansion threads can be trained to improve the performance of the compression and expansion threads.

[0012] In the present application, the description knowledge of the tailings pollutant composition description knowledge and the tailings pollutant data description knowledge corresponding to the target item are spliced ​​to obtain the spliced ​​description knowledge of the target item, including: Using a description knowledge splicing thread to perform description knowledge splicing on the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item, to obtain the spliced ​​description knowledge of the target item; It can be understood that when the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item are spliced, the problem of inaccurate splicing is improved, so that the spliced ​​description knowledge of the target item can be obtained more accurately.

[0013] The training method for describing the knowledge splicing thread includes: Obtaining a second training pollutant data set, the second training pollutant data set comprising second example tailings pollutant data having a second training target, the second training target being associated with annotated tailings pollutant data description knowledge, annotated tailings pollutant component description knowledge, and a tailings pollutant data catalog; Using the description knowledge splicing thread, the annotated tailings pollutant data description knowledge and the annotated tailings pollutant component description knowledge corresponding to the second training target are spliced ​​to obtain the regression analysis splicing description knowledge of the second training target; Using a pollutant attribute recognition unit to perform pollutant attribute recognition on the regression analysis splicing description knowledge to obtain regression analysis tailings pollutant data of the second training target; The description knowledge splicing thread is trained based on the deviation value between the regression analysis tailings pollutant data and the tailings pollutant data directory corresponding to the second training target.

[0014] It can be understood that the performance of the description knowledge splicing thread is improved by the second training pollutant dataset.

[0015] In the present application, the tailings pollutant data covering the target item is subjected to description knowledge extraction to obtain the tailings pollutant data description knowledge of the target item, including: Using a description knowledge extraction thread to extract description knowledge of the tailings pollutant data of the target item, to obtain description knowledge of the tailings pollutant data of the target item; It can be understood that when extracting descriptive knowledge of tailings pollutant data covering the target item, the problem of inaccurate extraction is improved, so that the descriptive knowledge of tailings pollutant data covering the target item can be accurately obtained.

[0016] The training method for describing the knowledge extraction thread includes: obtaining a third training pollutant data set, the third training pollutant data set comprising third example tailings pollutant data having a third training target, the third training target being associated with a tailings pollutant data directory; The original description knowledge extraction unit in the original neural network is used to perform special extraction on the tailings pollutant data of the third example to obtain a regression analysis description knowledge network of the tailings pollutant data of the third example; Using the recognition unit in the original neural network to perform recognition according to the regression analysis description knowledge network of the third example tailings pollutant data, to obtain the regression analysis tailings pollutant data of the third training target in the third example tailings pollutant data; The original neural network is trained based on the deviation value between the regression analysis tailings pollutant data corresponding to the third training target in the third example tailings pollutant data and the tailings pollutant data directory, and the original description knowledge extraction unit that has completed the training is determined as the description knowledge extraction thread.

[0017] It can be understood that the performance of the description knowledge extraction thread is improved by training the description knowledge extraction thread with the third training pollutant data set.

[0018] In this application, the tailings wastewater data to be analyzed covering the target items are analyzed for important content and tailings composition is constructed to obtain the important content information of the analysis corresponding to the target items and the original tailings composition, including: obtaining a real-time tailings wastewater data set, wherein the real-time tailings wastewater data set includes a plurality of real-time tailings wastewater data continuously covering the target matter; Performing important content analysis on the current real-time tailings sewage data to obtain information on the important content of the analysis of the target item; the information on the important content of the analysis includes location information of the important content of the analysis and type information of the important content of the analysis; The original tailings composition of the target item is generated based on the information of the important content of the analysis by using a specified rule; the original tailings composition includes a plurality of component elements and the associated content between the main component elements; the important content of the analysis is determined as the main component element; Based on the sorting information of the real-time tailings wastewater data in the real-time tailings wastewater data set, the target item is associated with the original tailings components in each of the real-time tailings wastewater data.

[0019] It can be understood that when conducting important content analysis and tailings composition construction on the tailings wastewater data to be analyzed covering the target items, the problem of inaccurate construction is avoided, so that the important content information of the analysis corresponding to the target items and the original tailings composition can be obtained more accurately.

[0020] In the present application, the optimization of each of the important analysis information in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item also includes: Based on the current real-time tailings wastewater data and the important content information of the analysis of the target item in the real-time tailings wastewater data adjacent to the current real-time tailings wastewater data, the original tailings composition of the target item in the current real-time tailings wastewater data is optimized to obtain the tailings composition optimization result of the target item.

[0021] It can be understood that when optimizing the important analysis content information in the original tailings composition of the target item, the problem of inaccurate optimization is improved, so that the tailings composition optimization result of the target item can be obtained more accurately.

[0022] In the present application, after the step of splicing the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain the spliced ​​description knowledge of the target item, it also includes: According to the sorting information of the real-time tailings sewage data, the splicing description knowledge corresponding to the target item in each of the real-time tailings sewage data is sorted to obtain a splicing description knowledge sequence of the target item; The step of determining tailings pollutant data of the target item based on the spliced ​​description knowledge of the target item includes: Based on the spliced ​​description knowledge sequence of the target item, the tailings pollutant data of the target item is determined.

[0023] It can be understood that the tailings pollutant data of the target item can be accurately obtained through the sorting information of the real-time tailings wastewater data.

[0024] In a second aspect, a tailings wastewater analysis system based on artificial intelligence is provided, comprising a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute it to implement the above-mentioned tailings wastewater analysis method based on artificial intelligence.

[0025] The tailings wastewater analysis method and system based on artificial intelligence provided by the embodiment of the present application analyzes important contents and constructs tailings components for the tailings wastewater data to be analyzed covering the target items, and obtains the important content information of the analysis corresponding to the target items and the original tailings components; the original tailings components include the important content of the analysis and the related content between the important content of the analysis; the important content information of each analysis in the original tailings components of the target items is optimized to obtain the tailings component optimization result of the target items; the tailings component optimization result includes all the specified important contents and the related content between the specified important contents, and the number of the important content of the analysis covered in the original tailings components is not greater than the number of the specified important contents covered in the tailings component optimization result; based on the tailings component optimization result of the target items, the tailings pollutant data of the target items is determined. Under the premise that the target is partially disturbed or the target tailings pollutant data is fuzzy, the original tailings components of the target items are optimized, and the important contents that have been disturbed and not analyzed are analyzed to obtain the complete tailings components of the target items, thereby improving the analysis accuracy of the tailings pollutant data of the target items. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0027] Figure 1 A flow chart of an artificial intelligence-based tailings wastewater analysis method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0029] See also Figure 1 , shows a tailings wastewater analysis method based on artificial intelligence, which may include the technical solutions described in the following steps S1-S3.

[0030] S1: Conduct important content analysis and tailings composition construction on the tailings wastewater data to be analyzed that covers the target items, and obtain the important analysis content information and original tailings composition corresponding to the target items; the original tailings composition includes the important analysis content and the related content between the important analysis contents.

[0031] Among them, the important content information of the analysis in this application refers to the important content of large wastewater volume, high suspended solids content, and more types of harmful substances with lower concentrations. The original tailings composition includes the important content of the analysis, which refers to heavy metal ions (copper, zinc, lead, nickel, iron, barium, arsenic, cadmium, etc.) and mineral processing agents (xanthate, black medicine, cyanide, pine root oil, cresol, copper sulfate, heavy metal salts, sodium sulfate, sulfuric acid, lime, etc.).

[0032] Exemplarily, the tailings wastewater data to be analyzed is formed by digitally describing the tailings data after detection. The main harmful substances in the tailings wastewater are heavy metal ions, various organic and inorganic flotation agents used in ore flotation, including highly toxic cyanide, cyanochromium compounds, etc. The wastewater also contains various insoluble coarse and fine dispersed impurities. Tailings wastewater often contains sulfates, chlorides or hydroxides of sodium, magnesium, calcium, etc. The acid in the tailings wastewater is mainly formed by sulfur-containing minerals being oxidized by air and mixed with water.

[0033] S2: Optimize the important analytical content information in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item; the tailings composition optimization result includes all designated important contents and the related contents between the designated important contents, and the number of important analytical contents covered in the original tailings composition is not greater than the number of designated important contents covered in the tailings composition optimization result.

[0034] Exemplarily, all designated important contents are understood as all important contents in the database. Designated important contents are understood as important contents in the actual tailings composition optimization results. The purpose of optimization in this application is to remove the interference data in the important content information of each analysis in the original tailings composition, for example: to remove the information that is not pollutant data, so as to reduce the interference in the data and improve the accuracy for subsequent data type determination.

[0035] S3: Based on the tailings composition optimization results of the target item, determine the tailings pollutant data of the target item.

[0036] Exemplarily, the tailings composition optimization result is data without interference, so the pollutant data can be accurately determined by comparing it with the pollutant composition in the database.

[0037] The tailings wastewater analysis method based on artificial intelligence provided in this embodiment optimizes the original tailings composition of the target item, analyzes the important contents that have been interfered with and not analyzed, and obtains the complete tailings composition of the target item, thereby improving the analysis accuracy of the tailings pollutant data of the target item.

[0038] In a possible implementation example, the training method for compressing and expanding threads specifically includes the following steps.

[0039] S101: Obtain a first training pollutant data set, the first training pollutant data set comprising first example tailings pollutant data having a first training target, the first training target being associated with annotated tailings components and a tailings pollutant data directory.

[0040] Specifically, a first example tailings pollutant data is obtained, and the first example tailings pollutant data specifically covers a first training target. The first training target in the first example tailings pollutant data may be partially disturbed. The first training target may be a movable target such as an animal. At least one first training target may be covered in the first example tailings pollutant data. The important content analysis thread performs an important content analysis on the first training target to obtain the analyzed important content information of the first training target. The analyzed important content information includes the location information of the analyzed important content and the type of the analyzed important content.

[0041] Based on the designated rule and the types of the important analysis contents corresponding to the first training target, associated contents are established between the important analysis contents to form the annotated tailings component of the first training target.

[0042] S102: Use compression and expansion threads to sequentially expand and compress the annotated tailings components of the first example tailings pollutant data to generate a regression analysis tailings component of the first training target; the regression analysis tailings component includes all designated important contents of the first training target and related contents between the designated important contents.

[0043] Exemplarily, the expansion processing is to perform local processing on the annotated tailings components of the first example tailings pollutant data to display the annotated tailings component information of the first example tailings pollutant data more comprehensively, and the compression processing is to extract the key information of the annotated tailings components of the first example tailings pollutant data.

[0044] S103: using a pollutant attribute identification unit to identify pollutant attributes of the regression analysis tailings components of the first training target, and obtaining the regression analysis tailings pollutant data of the first training target.

[0045] For example, the pollutant attribute can understand why a substance has what kind of pollution.

[0046] Specifically, the state recognition thread performs regression analysis based on the material content of each component element in the specified tailings component of the first training target to obtain the regression analysis tailings pollutant data of the first training target, wherein the regression analysis tailings pollutant data may be the probability of each specified tailings pollutant data.

[0047] S104: Training the compression and expansion threads based on the deviation value between the regression analysis tailings pollutant data and the tailings pollutant data directory corresponding to the first training target.

[0048] Among them, it is to improve the performance of compression and expansion threads.

[0049] Specifically, the deviation value is calculated by regression analysis of tailings pollutant data and the operation status directory corresponding to the first training target, and the parameters of the compression and expansion threads are optimized according to the deviation value to achieve the training of the compression and expansion threads. When the regression analysis deviation value of the compression and expansion threads is less than the specified error, the optimization of the compression and expansion threads can be stopped. Or when the number of iterations reaches the specified number, the optimization of the compression and expansion threads can be stopped.

[0050] In a possible implementation, the compression and expansion thread can also be trained based on the training real-time tailings sewage data set. Specifically, the training real-time tailings sewage data set covering the first training target is input into the compression and expansion thread, and the first training target has an annotated tailings pollutant component map in each training real-time tailings sewage data, and the first training target has a tailings pollutant data directory. Among them, the first training target has a corresponding annotated tailings pollutant component map sequence in the training real-time tailings sewage data. The compression and expansion thread respectively expands and compresses the tailings pollutant component map in the annotated tailings pollutant component map sequence of the first training target to obtain the regression analysis tailings components of each tailings pollutant component map. According to the sorting information of the annotated tailings pollutant component map sequence, each regression analysis tailings component is sorted to generate the regression analysis tailings component sequence of the first training target. The pollutant attribute recognition unit is used to identify the pollutant attributes of the regression analysis tailings component sequence of the first training target to obtain the regression analysis tailings pollutant data of the first training target. The compression and expansion thread is trained based on the deviation value between the regression analysis tailings pollutant data corresponding to the first training target and the tailings pollutant data directory.

[0051] Specifically, when the training example of the training compression and expansion thread is to train a real-time tailings wastewater data set, the compression and expansion thread determines the regression analysis tailings composition of the first training target in the current real-time tailings wastewater data based on the original tailings composition of the first training target in the current real-time tailings wastewater data and the original tailings composition of the first training target in the real-time tailings wastewater data adjacent to the current real-time tailings wastewater data set, thereby generating the regression analysis tailings composition of the first training target in the current real-time tailings wastewater data.

[0052] In a possible implementation example, a training method for describing a knowledge splicing thread specifically includes the following steps.

[0053] S201: Obtain a second training pollutant data set, the second training pollutant data set includes second example tailings pollutant data with a second training target, the second training target is associated with annotated tailings pollutant data description knowledge, annotated tailings pollutant component description knowledge and a tailings pollutant data catalog.

[0054] Among them, descriptive knowledge is understood as features.

[0055] Specifically, a second example tailings pollutant data is obtained, and the second example tailings pollutant data specifically covers the second training target. The second training target in the second example tailings pollutant data can be partially disturbed. At least one second training target can be covered in the second example tailings pollutant data. The second training target has annotated tailings pollutant data description knowledge and annotated tailings pollutant component description knowledge. Among them, the annotated tailings pollutant data description knowledge of the second training target is obtained by extracting the description knowledge of the tailings pollutant data covering the second training target using the target description knowledge extraction thread. The annotated tailings pollutant component description knowledge of the second training target is obtained by optimizing and extracting the description knowledge of the original tailings component of the second training target using the compression and expansion threads trained by the above steps. The original tailings component of the second training target is generated based on the important content of the analysis of the second training target according to the specified rules. Among them, the target description knowledge extraction thread can be a convolutional neural network (CNN). The second example tailings pollutant data in this embodiment can be the same as the first example tailings pollutant data, or it can be different; the second training target can be the same as the first training target, or it can be different.

[0056] S202: Using the description knowledge splicing thread, the description knowledge of the annotated tailings pollutant data and the description knowledge of the annotated tailings pollutant components corresponding to the second training target are spliced ​​to obtain the regression analysis splicing description knowledge of the second training target.

[0057] Specifically, the annotated tailings pollutant data description knowledge and the annotated tailings pollutant component description knowledge corresponding to the second training target are simultaneously input into the description knowledge splicing thread, and the description knowledge splicing thread splices the annotated tailings pollutant data description knowledge and the annotated tailings pollutant component description knowledge to obtain the regression analysis splicing description knowledge of the second training target. Among them, the description knowledge splicing thread can be a fully connected layer.

[0058] In this embodiment, the description knowledge splicing thread has at least two input terminals to facilitate splicing of the tailings pollutant component description knowledge and the tailings pollutant data description knowledge of the second training target. And the tailings pollutant component description knowledge of the second training target is spliced ​​with the tailings pollutant data description knowledge through the description knowledge splicing thread to obtain richer semantic information, thereby facilitating improving the accuracy of the pollutant attribute analysis of the second training target.

[0059] S203: Using a pollutant attribute recognition unit to perform pollutant attribute recognition on the regression analysis splicing description knowledge, and obtaining the regression analysis tailings pollutant data of the second training target.

[0060] Specifically, the state recognition thread is used to perform state recognition on the regression analysis splicing description knowledge of the second training target to generate the regression analysis tailings pollutant data of the second training target. The regression analysis tailings pollutant data may be the probability of each designated tailings pollutant data.

[0061] S204: Training the description knowledge splicing thread based on the deviation value between the tailings pollutant data and the tailings pollutant data catalog corresponding to the second training target.

[0062] The deviation value is understood as the error value.

[0063] The loss value between the tailings pollutant data and the tailings pollutant data catalog based on the regression analysis of the second training objective is used to optimize the parameters in the description knowledge splicing thread to achieve the training of the description knowledge splicing thread.

[0064] When the loss value of the description knowledge splicing thread is less than the specified value, the description knowledge splicing thread can be stopped from being optimized. Or when the number of iterations reaches the specified number, the description knowledge splicing thread can be stopped from being optimized. Alternatively, a specified validation data set can be input into the description knowledge splicing thread, and the training effect of the description knowledge splicing thread can be evaluated based on the specified validation data set, and then the tailings pollutant composition and hyperparameters of the description knowledge splicing thread can be adjusted to ensure that the description knowledge splicing thread can effectively splice different types of description knowledge and improve the accuracy of target state classification.

[0065] In the above embodiment, the description knowledge splicing thread can also be trained based on the training real-time tailings wastewater data set. Specifically, the annotated tailings pollutant data description knowledge sequence and the annotated tailings pollutant component description knowledge sequence corresponding to the training real-time tailings wastewater data set covering the second training target are input into the description knowledge splicing thread. Among them, the annotated tailings pollutant data description knowledge sequence is the tailings pollutant data description knowledge sequence composed of the tailings pollutant data description knowledge corresponding to the second training target in each real-time tailings wastewater data according to the time series; the annotated tailings pollutant component description knowledge sequence is the tailings pollutant component description knowledge sequence composed of the tailings pollutant component description knowledge corresponding to the second training target in each real-time tailings wastewater data according to the time series, and the second training target has a tailings pollutant data directory. The description knowledge splicing thread performs description knowledge splicing on the tailings pollutant component description knowledge and tailings pollutant data description knowledge of the second training target in the same real-time tailings wastewater data to generate the regression analysis splicing description knowledge in the real-time tailings wastewater data. The regression analysis splicing description knowledge corresponding to each real-time tailings wastewater data can be obtained through the above steps. The regression analysis splicing description knowledge is used to generate the regression analysis splicing description knowledge sequence of the second training target according to the time sequence of the real-time tailings sewage data. The pollutant attribute recognition unit is used to identify the pollutant attributes of the regression analysis splicing description knowledge sequence of the second training target to obtain the regression analysis tailings pollutant data of the second training target. The description knowledge splicing thread is trained based on the deviation value between the regression analysis tailings pollutant data corresponding to the second training target and the tailings pollutant data directory.

[0066] The above steps can realize the training of compression and expansion threads and description knowledge splicing threads.

[0067] In a possible implementation example, a training method for describing a knowledge extraction thread specifically includes the following steps.

[0068] A third training pollutant data set is obtained, the third training pollutant data set includes a third example tailings pollutant data with a third training target, and the third training target is associated with a tailings pollutant data directory; the original description knowledge extraction unit in the original neural network is used to extract the third example tailings pollutant data, and a regression analysis description knowledge network of the third example tailings pollutant data is obtained; the recognition unit in the original neural network is used to identify according to the regression analysis description knowledge network of the third example tailings pollutant data, and the regression analysis tailings pollutant data of the third training target in the third example tailings pollutant data is obtained; the original neural network is trained based on the deviation value between the regression analysis tailings pollutant data corresponding to the third training target in the third example tailings pollutant data and the tailings pollutant data directory, and the original description knowledge extraction unit that has completed the training is determined as a description knowledge extraction thread. The description knowledge extraction thread in this embodiment can be the CNN thread in the above embodiment.

[0069] Specifically, in step S1, important content analysis and tailings composition construction are performed on the tailings wastewater data to be analyzed that covers the target items, and the specific implementation method of obtaining the important content information of the analysis corresponding to the target items and the original tailings composition is as follows.

[0070] In a possible implementation example, tailings wastewater data to be analyzed in a target area is obtained in real time through a tailings pollutant data acquisition device, and the tailings wastewater data to be analyzed covers the target items. Target analysis is performed on the tailings wastewater data to be analyzed to obtain analysis boxes of each target item. The analysis box only covers one target item. Important content analysis is performed on the analysis box covering the target item to obtain important content information of the analysis of the target item.

[0071] In one possible implementation, an offline real-time tailings sewage data set is obtained, and the real-time tailings sewage data set includes a plurality of real-time tailings sewage data that continuously cover target items. All real-time tailings sewage data in the real-time tailings sewage data set are traversed, and each real-time tailings sewage data is sequentially determined as the current real-time tailings sewage data. Target analysis is performed on the current real-time tailings sewage data to obtain an analysis frame covering the target item, and only one target item is covered in the analysis frame. Important content analysis is performed on the analysis frame covering the target item to obtain information on the important content of the analysis of the target item. Information on the important content of the analysis of the target item in each real-time tailings sewage data is obtained through the above steps.

[0072] When the information obtained is a real-time tailings wastewater data set, based on the sorting information of the real-time tailings wastewater data in the real-time tailings wastewater data set, the original tailings components of the target item in each real-time tailings wastewater data are associated to generate an original tailings component sequence of the target item.

[0073] Specifically, in step S2, the original tailings composition is optimized based on the important content information of the analysis of the target item, and the specific implementation method of obtaining the tailings composition optimization result of the target item is as follows.

[0074] In one possible implementation, the original tailings composition corresponding to the target item is input into the compression and expansion thread trained in the above steps S101 to S104. The compression and expansion thread performs expansion and compression processing on the original tailings composition in turn to generate an optimized tailings composition result for the target item.

[0075] In a possible implementation, if the category information of the important content to be analyzed does not belong to the external component element, the original content of the important content to be analyzed is set to zero.

[0076] The original material content of each core component in the specific original tailings composition.

[0077] In a specific embodiment, the original tailings composition of the target item in the current real-time tailings wastewater data is optimized based on the important content information of the analysis of the target item in the current real-time tailings wastewater data and the real-time tailings wastewater data adjacent to the current real-time tailings wastewater data to obtain the tailings composition optimization result of the target item. Specifically, the material content of each designated important content in the tailings composition optimization result of the target item in the current real-time tailings wastewater data can be determined based on the following formula.

[0078] Specifically, the specific implementation method of determining the tailings pollutant data of the target item based on the tailings composition optimization result of the target item in step S3 is as follows.

[0079] A process of a specific embodiment of step S3 in the tailings wastewater analysis method based on artificial intelligence is provided.

[0080] S31: splicing the tailings composition optimization result of the target item with the tailings wastewater data to be analyzed covering the target item to obtain the splicing description knowledge of the target item.

[0081] A process of a specific embodiment of step S31 in the tailings wastewater analysis method based on artificial intelligence is provided.

[0082] S311: extracting descriptive knowledge of the tailings composition optimization results of the target item to obtain descriptive knowledge of the tailings pollutant components of the target item.

[0083] In a possible implementation example, each designated important content in the tailings composition optimization result of the target item has a corresponding material content. The material content of the designated important content reflects the probability that the target item is in each designated tailings pollutant data. By extracting the material content of all designated important contents in the tailings composition optimization result of the target item, the tailings pollutant composition description knowledge of the target item is generated.

[0084] S312: Extract descriptive knowledge of tailings pollutant data covering the target items to obtain descriptive knowledge of tailings pollutant data of the target items.

[0085] In one possible implementation example, descriptive knowledge is extracted from the analysis box covering the target item to obtain tailings pollutant data descriptive knowledge of the target item.

[0086] In a specific embodiment, a CNN thread is used to extract descriptive knowledge from an analysis box covering a target item, thereby obtaining tailings pollutant data descriptive knowledge of the target item.

[0087] S313: splicing the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain the spliced ​​description knowledge of the target item.

[0088] In a possible implementation example, the tailings pollutant composition description knowledge and the tailings pollutant data description knowledge corresponding to the target item are input into the description knowledge splicing thread trained from the above steps S201 to S204 to obtain the spliced ​​description knowledge of the target item.

[0089] S32: Determine the tailings pollutant data of the target item based on the splicing description knowledge of the target item.

[0090] Specifically, the description knowledge splicing thread is used to splice the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain the spliced ​​description knowledge of the target item. The tailings pollutant data identification thread obtained by pre-training is used to determine the tailings pollutant data of the target item based on the spliced ​​description knowledge of the target item.

[0091] In a specific embodiment, according to the sorting information of the real-time tailings sewage data, the splicing description knowledge corresponding to the target item in each real-time tailings sewage data is sorted to obtain the splicing description knowledge sequence of the target item. The tailings pollutant data identification thread obtained by pre-training is used to determine the tailings pollutant data of the target item based on the splicing description knowledge sequence of the target item.

[0092] The tailings wastewater analysis method based on artificial intelligence provided in this embodiment optimizes the original tailings composition of the target item, analyzes the important contents that have been interfered with and not analyzed, and obtains the complete tailings composition of the target item, thereby improving the analysis accuracy of the tailings pollutant data of the target item.

[0093] On the basis of the above, a tailings wastewater analysis device based on artificial intelligence is provided, and the device comprises: A knowledge fragment extraction module is used to perform knowledge fragment extraction processing on the example collapse hidden danger data to obtain secondary knowledge fragments and important knowledge fragments of the example collapse hidden danger data; A label building module, used to build secondary description labels according to the secondary knowledge fragments corresponding to the several example collapse hazard data, and to build important description labels according to the important knowledge fragments corresponding to the several example collapse hazard data; A component acquisition module is used to perform important content analysis and tailings component construction on the tailings wastewater data to be analyzed covering the target items, and obtain the important analysis content information and original tailings components corresponding to the target items; the original tailings components include important analysis contents and related contents between the important analysis contents; A result optimization module is used to optimize the information of each of the important analysis contents in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item; the tailings composition optimization result includes all the designated important contents and the associated contents between the designated important contents, and the number of the important analysis contents covered in the original tailings composition is not greater than the number of the designated important contents covered in the tailings composition optimization result; A data determination module is used to determine the tailings pollutant data of the target item based on the tailings composition optimization result of the target item.

[0094] On the basis of the above, an artificial intelligence-based tailings wastewater analysis system is shown, including a processor and a memory that communicate with each other, and the processor is used to read and execute a computer program from the memory to implement the above method.

[0095] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.

[0096] In summary, based on the above scheme, the tailings wastewater data to be analyzed that covers the target items are analyzed for important content and the tailings composition is constructed to obtain the important content information of the analysis corresponding to the target items and the original tailings composition; the original tailings composition includes the important content of the analysis and the related content between the important content of the analysis; the important content information of each analysis in the original tailings composition of the target item is optimized to obtain the tailings composition optimization result of the target item; the tailings composition optimization result includes all the specified important content and the related content between the specified important content, and the number of important content of the analysis covered in the original tailings composition is not greater than the number of specified important content covered in the tailings composition optimization result; based on the tailings composition optimization result of the target item, the tailings pollutant data of the target item is determined. Under the premise that the target is partially interfered with or the target tailings pollutant data is blurred, this application optimizes the original tailings composition of the target item, analyzes the important content that has been interfered with and not analyzed, obtains the complete tailings composition of the target item, and improves the analysis accuracy of the tailings pollutant data of the target item.

[0097] It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present application can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software such as executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (e.g., firmware).

[0098] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other beneficial effects that may be obtained.

Claims

1. A tailings wastewater analysis method based on artificial intelligence, characterized in that: The tailings wastewater analysis method based on artificial intelligence includes: Performing important content analysis and tailings composition construction on the tailings wastewater data to be analyzed that covers the target items, and obtaining the important analysis content information and original tailings composition corresponding to the target items; the original tailings composition includes important analysis contents and related contents between the important analysis contents; Optimizing each of the important analytical contents in the original tailings composition of the target item to obtain an optimized result of the tailings composition of the target item; the optimized result of the tailings composition includes all designated important contents and associated contents between the designated important contents, and the number of the important analytical contents covered in the original tailings composition is not greater than the number of the designated important contents covered in the optimized result of the tailings composition; Based on the tailings composition optimization result of the target item, the tailings pollutant data of the target item is determined.

2. The tailings wastewater analysis method based on artificial intelligence according to claim 1 is characterized in that: The step of determining tailings pollutant data of the target item based on the tailings composition optimization result of the target item includes: Extracting descriptive knowledge of the tailings composition optimization result of the target item to obtain descriptive knowledge of the tailings pollutant composition of the target item; Extracting descriptive knowledge of tailings pollutant data covering the target items to obtain descriptive knowledge of tailings pollutant data of the target items; Performing description knowledge splicing on the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain spliced ​​description knowledge of the target item; Based on the spliced ​​description knowledge of the target item, the tailings pollutant data of the target item is determined.

3. The tailings wastewater analysis method based on artificial intelligence according to claim 1 is characterized in that: The step of optimizing the important analysis information in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item includes: The original tailings components are sequentially expanded and compressed using compression and expansion threads to generate tailings component optimization results for the target item.

4. The tailings wastewater analysis method based on artificial intelligence according to claim 3 is characterized in that: The training method of the compression and expansion thread includes: Obtaining a first training contaminant data set, the first training contaminant data set comprising first example tailings contaminant data having a first training target, the first training target being associated with annotated tailings composition and a tailings contaminant data catalog; The compression and expansion threads are used to sequentially expand and compress the annotated tailings components of the first example tailings pollutant data to generate a regression analysis tailings component of the first training target; the regression analysis tailings component includes all designated important contents of the first training target and associated contents between the designated important contents; Based on the regression analysis tailings composition of the first training target, obtaining the regression analysis tailings pollutant data of the first training target; The compression and expansion threads are trained based on the deviation value between the regression analysis tailings pollutant data and the tailings pollutant data directory corresponding to the first training target.

5. The tailings wastewater analysis method based on artificial intelligence according to claim 2 is characterized in that: The step of performing description knowledge splicing on the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain the spliced ​​description knowledge of the target item includes: Using a description knowledge splicing thread to perform description knowledge splicing on the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item, to obtain the spliced ​​description knowledge of the target item; The training method for describing the knowledge splicing thread includes: Obtaining a second training pollutant data set, the second training pollutant data set comprising second example tailings pollutant data having a second training target, the second training target being associated with annotated tailings pollutant data description knowledge, annotated tailings pollutant component description knowledge, and a tailings pollutant data catalog; Using the description knowledge splicing thread, the annotated tailings pollutant data description knowledge and the annotated tailings pollutant component description knowledge corresponding to the second training target are spliced ​​to obtain the regression analysis splicing description knowledge of the second training target; Using a pollutant attribute recognition unit to perform pollutant attribute recognition on the regression analysis splicing description knowledge to obtain regression analysis tailings pollutant data of the second training target; The description knowledge splicing thread is trained based on the deviation value between the regression analysis tailings pollutant data and the tailings pollutant data directory corresponding to the second training target.

6. The tailings wastewater analysis method based on artificial intelligence according to claim 2 is characterized in that: The extracting of descriptive knowledge of tailings pollutant data covering the target items to obtain descriptive knowledge of tailings pollutant data of the target items includes: Using a description knowledge extraction thread to extract description knowledge of the tailings pollutant data of the target item, to obtain description knowledge of the tailings pollutant data of the target item; The training method for describing the knowledge extraction thread includes: Obtaining a third training pollutant data set, the third training pollutant data set comprising third example tailings pollutant data having a third training target, the third training target being associated with a tailings pollutant data directory; The original description knowledge extraction unit in the original neural network is used to perform special extraction on the tailings pollutant data of the third example to obtain a regression analysis description knowledge network of the tailings pollutant data of the third example; Using the recognition unit in the original neural network to perform recognition according to the regression analysis description knowledge network of the third example tailings pollutant data, to obtain the regression analysis tailings pollutant data of the third training target in the third example tailings pollutant data; The original neural network is trained based on the deviation value between the regression analysis tailings pollutant data corresponding to the third training target in the third example tailings pollutant data and the tailings pollutant data directory, and the original description knowledge extraction unit that has completed the training is determined as the description knowledge extraction thread.

7. The tailings wastewater analysis method based on artificial intelligence according to any one of claims 1-6, characterized in that: The important content analysis and tailings composition construction of the tailings wastewater data to be analyzed covering the target items are performed to obtain the important content information of the analysis corresponding to the target items and the original tailings composition, including: obtaining a real-time tailings wastewater data set, wherein the real-time tailings wastewater data set includes a plurality of real-time tailings wastewater data continuously covering the target matter; Performing important content analysis on the current real-time tailings sewage data to obtain information on the important content of the analysis of the target item; the information on the important content of the analysis includes location information of the important content of the analysis and type information of the important content of the analysis; The original tailings composition of the target item is generated based on the information of the important content of the analysis by using a specified rule; the original tailings composition includes a plurality of component elements and the associated content between the main component elements; the important content of the analysis is determined as the main component element; Based on the sorting information of the real-time tailings wastewater data in the real-time tailings wastewater data set, the target item is associated with the original tailings components in each of the real-time tailings wastewater data.

8. The tailings wastewater analysis method based on artificial intelligence according to claim 7 is characterized in that: The step of optimizing the important analysis information in the original tailings composition of the target item to obtain the tailings composition optimization result of the target item further includes: Based on the current real-time tailings wastewater data and the important content information of the analysis of the target item in the real-time tailings wastewater data adjacent to the current real-time tailings wastewater data, the original tailings composition of the target item in the current real-time tailings wastewater data is optimized to obtain the tailings composition optimization result of the target item.

9. The tailings wastewater analysis method based on artificial intelligence according to claim 2 is characterized in that: After the step of performing description knowledge splicing of the tailings pollutant component description knowledge and the tailings pollutant data description knowledge corresponding to the target item to obtain the spliced ​​description knowledge of the target item, the step further includes: According to the sorting information of the real-time tailings sewage data, the splicing description knowledge corresponding to the target item in each of the real-time tailings sewage data is sorted to obtain a splicing description knowledge sequence of the target item; The step of determining tailings pollutant data of the target item based on the spliced ​​description knowledge of the target item includes: Based on the spliced ​​description knowledge sequence of the target item, the tailings pollutant data of the target item is determined.

10. A tailings wastewater analysis system based on artificial intelligence, characterized in that: It comprises a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute it to implement the tailings wastewater analysis method based on artificial intelligence as described in any one of claims 1-9.

Images