Underwater alum bloom state identification method and monitoring device, computer equipment and storage medium

Abstract

The present invention discloses an underwater alum floc state recognition method, a monitoring device, a computer device and a storage medium that can automatically identify underwater alum flocs and guide the dosage of coagulant according to the alum floc state. The method performs image preprocessing on underwater alum floc images, performs multiple multi-scale sampling, extracts features from the sampling images according to different feature description factors, obtains feature vectors, selects feature vectors with high correlation and inputs them into a fuzzy layer, obtains alum floc value, and compensates the alum floc value according to the difference between turbidity meter data and the alum floc value and the process data of the water plant, and performs iterative training until accurate alum floc value data is obtained. A alum floc monitoring system is also disclosed, including an acquisition module, a sampling module, a feature extraction module, a feature processing module, a result processing module, a return module and a detection module. The present invention analyzes the state of underwater alum flocs in real time and objectively through the combination of machine vision and artificial intelligence, and provides a guiding basis for the dosage of coagulant.

Description

Technical Field

[0001] The invention belongs to the field of tap water production, and in particular relates to a method for identifying the state of underwater alum flowers based on machine vision, a monitoring device, a computer device and a storage medium. Background Art

[0002] The water treatment process of a water plant generally includes drug administration, coagulation sedimentation, filtration and other links. The coagulation sedimentation link refers to the addition of coagulants to destroy the stability of the colloidal particles in the water, forming flocs (i.e. alum flowers) under certain conditions and using gravity to precipitate them and separate them from the water.

[0003] At present, domestic water plants mainly rely on manual inspections to monitor the status of alum flocs. This is mainly done by manually observing the alum flocs at the end of the reaction tank and judging the flocculation effect based on factors such as the size, shape, and structure of the alum flocs. However, this method requires workers to have many years of experience and is easily affected by human factors. Summary of the invention

[0004] The purpose of the present invention is to provide an underwater alum floc state identification method, a monitoring device, a computer device and a storage medium which can automatically identify underwater alum flocs and guide the placement of coagulants according to the underwater alum floc state.

[0005] The present invention provides a method for identifying underwater alum bloom status, comprising the following steps:

[0006] Perform image preprocessing on the collected underwater alum flower images to obtain a region of interest map;

[0007] The region of interest map is input into a convolutional neural network, and the region of interest map is sampled multiple times to obtain sampling maps of multiple scales;

[0008] Selecting feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images to obtain a feature vector composed of extraction results corresponding to each feature description factor;

[0009] The fuzzy neural network screens the feature vector to obtain a feature vector with high correlation, and inputs the feature vector with high correlation into the fuzzy layer of the fuzzy neural network to obtain the alum flower value;

[0010] The floc value is corrected with reference to the turbidity meter data, the corrected floc value and the amount of coagulant added in the water body are recorded in a reasonable matching data sequence, and the parameters of the fuzzy neural network are updated;

[0011] The updated fuzzy neural network is used to return to the starting step and start again, and iterative training is performed until the set termination condition is met, and the training is terminated to obtain the alum flower state discrimination network;

[0012] The collected water body images are input into the alum floc state discrimination network to obtain the alum floc value and the amount of coagulant added.

[0013] In one embodiment, the method for correcting the floc value is to detect the water body with a turbidity meter to obtain turbidity meter data, calculate the difference between the turbidity meter data and the floc value, and make corresponding compensation for the floc value based on the process data of the water plant.

[0014] In one embodiment, the calculation formula used for the correction method of the alum flower value is:

[0015] y new =y×(1+α×(NTU-y)),

[0016] Among them, y is the alum flower value before correction and compensation, a is the process compensation parameter, NTU is the turbidity meter data, y new It is the alum bloom value after correction and compensation.

[0017] In one of the embodiments, the method for screening the feature vectors is to first screen out the feature vectors with high correlation, and then adjust the weights corresponding to each feature in the feature vector.

[0018] In one embodiment, the parameters of the fuzzy neural network include membership function parameters of the fuzzy layer, fuzzy rules, and parameters of the anti-fuzzy function.

[0019] In one of the embodiments, the sampling method is to obtain a sampling map by sequentially passing the region of interest map through a convolution layer and a pooling layer of a convolutional neural network.

[0020] In one of the embodiments, the method for extracting feature results for each feature description factor is to use a local binary pattern to calculate texture information corresponding to the feature description factor in the image to obtain the feature result.

[0021] The present invention also provides an alum bloom monitoring device using the underwater alum bloom state identification method, comprising:

[0022] The acquisition module is used to perform image preprocessing on the acquired underwater alum flower images to obtain an area of ​​interest map;

[0023] A sampling module is used to input the region of interest map into the convolutional neural network, sample the region of interest map multiple times, and obtain sampling maps of multiple scales;

[0024] A feature extraction module, used to select feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images, and obtain a feature vector composed of extraction results corresponding to each feature description factor;

[0025] A feature processing module is used for the fuzzy neural network to screen the feature vector to obtain a feature vector with high correlation, and input the feature vector with high correlation into the fuzzy layer of the fuzzy neural network to obtain the alum value;

[0026] A result processing module is used to correct the floc value, record the corrected floc value and the amount of coagulant added in the water body into a reasonable matching data sequence, and update the parameters of the fuzzy neural network;

[0027] A return module is used to use the updated fuzzy neural network to return to the starting step and start again, and perform iterative training until the set termination condition is met, and then the training is terminated to obtain the alum flower state discrimination network;

[0028] The detection module is used to input the collected water body image into the alum bloom state discrimination network to obtain the alum bloom value and the amount of coagulant added.

[0029] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the underwater alum bloom state identification method when executing the computer program.

[0030] The present invention also provides a computer-readable storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the underwater alum bloom state identification method are implemented.

[0031] The underwater alum flower state identification method provided by the present invention has the following advantages and effects:

[0032] Through the combination of machine vision and artificial intelligence, the captured alum flowers images are processed at multiple scales and features, the status of underwater alum flowers is analyzed in real time and objectively to obtain the alum flower value, and the guidance basis for the dosage of coagulant is given, making the addition of coagulants simple and feasible, and no longer needing to rely on professional knowledge and experience to judge the amount of dosage.

[0033] The alum flower monitoring device provided by the present invention has the following advantages and effects:

[0034] This monitoring system can monitor in real time and automatically analyze and process the alum value in the water body, and judge whether the amount of coagulant added is appropriate based on the alum value and provide guidance on the adjustment direction. Machine vision is used to replace manual inspection. Compared with the post-settlement turbidity meter, it can predict the post-settlement turbidity earlier, overcoming the problem of untimely manual inspection and delayed judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of a first embodiment of a method for identifying underwater alum bloom status of the present invention;

[0036] Figure 2 It is a flow chart of a second embodiment of the underwater alum bloom state identification method of the present invention;

[0037] Figure 3 It is a structural diagram of an embodiment of an alum flower monitoring system of the present invention;

[0038] Figure 4 It is a reasonable matching curve diagram of an embodiment of the underwater alum flower state identification method of the present invention;

[0039] Figure 5 is a schematic diagram of a computer device of an embodiment;

[0040] Figure 6 It is a schematic diagram of the fuzzy neural network processing process in the embodiment of the underwater alum bloom state recognition method of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Mentioning "embodiment" in this article means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The appearance of this phrase in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0042] The underwater alum bloom state recognition method of the present invention can be applied to the underwater alum bloom state detection system of the open-air water plant. The first embodiment is as follows: Figure 1 As shown, the specific implementation steps are as follows:

[0043] S201, continuously collect underwater alum flower images through the image acquisition device of the water plant. The image acquisition device of this embodiment is mainly a waterproof industrial-grade underwater camera, which is fixed on a camera bracket, and the camera bracket is installed on a camera rod, which is fixed to the channel between the reaction tank and the sedimentation tank of the water plant, such as fixed to the guardrail of the channel through a U-shaped lock.

[0044] S202, preprocessing the collected underwater alum flower images to obtain a region of interest map, i.e., a ROI region image. The preprocessing includes using a mean filtering method to denoise the image, using a histogram equalization operation to adjust the image brightness distribution to achieve image enhancement, removing the distorted areas at the edge of the image due to deformation and blur, and if there is text information or a border at the top and bottom of the captured image, these areas should also be removed, and finally a ROI region map containing only valid water body information areas is obtained.

[0045] The specific calculation steps of the mean filtering method are: calculating the pixel mean of the window area, and then assigning the mean value to the pixel at the center point of the window to obtain the image after the mean filtering operation;

[0046] The calculation formula of the mean filtering method is:

[0047]

[0048] Among them, Sxy represents the filter window with the center point at (x, y) and a size of m*n, g(s, t) represents the original image, and f(x, y) represents the image obtained after mean filtering.

[0049] S203. Input the region of interest map (ROI region map) into a convolutional neural network. A convolution layer and a pooling layer of the convolutional neural network constitute a sampling unit. One sampling unit obtains a sampling map. The convolutional neural network of this embodiment has three groups of sampling units with different scale parameters, thereby obtaining three sampling maps of different sizes. The sizes of these three sampling maps are 1 / 2, 1 / 4, and 1 / 8 of the size of the ROI region map, respectively.

[0050] Specifically, the convolutional layer operation formula is:

[0051]

[0052] Among them, x[m, n] is the input image, h[m, n] is the convolution kernel, and y[m, n] is the output result of the convolution layer.

[0053] After the image is processed by the convolution layer, the data is passed to the pooling layer, where the maximum pooling strategy is used to complete the data sampling. The operation formula used by the pooling layer is:

[0054] f(x[i,j])=max{x[i,j],x[i+1,j],x[i,j+1],x[i+1,j+1]}.

[0055] S204, select feature description factors to perform global feature extraction and local feature extraction on the three sample images obtained, the feature description factors mainly include the shape, size, thickness, distribution, texture, structure, etc. of the alum flower, and adopt the local binary pattern calculation method to obtain the feature vector P = [p1, p2...pn] composed of the extraction results of each feature description factor. This embodiment extracts both global features and local features. Of course, global feature extraction or local feature extraction can also be performed on the sample images.

[0056] The specific extraction process is:

[0057]

[0058] S205, such as Figure 6 As shown, the fuzzy neural network autonomously screens the features Q=[q1, q2…qm], m<n, with a large correlation with the water quality description based on the obtained feature vector, and adjusts the weights corresponding to each feature. The features and the weights corresponding to the features form a highly correlated feature vector, and the highly correlated feature vector is input into the fuzzy layer of the fuzzy neural network for fuzzification, and the obtained fuzzy variables are input into the rule layer of the fuzzy neural network to obtain the output description, and then the output description is quantified to obtain the alum value.

[0059] S206. After calculating the difference between the alum floc value and the recorded data of the water quality turbidity meter of the water body, make correction compensation according to the process data of the water plant's treated water, and according to the amount of coagulant added in the water body, obtain the reasonable matching data of the corrected alum floc value and the amount of coagulant added as a guiding basis for the amount of coagulant added, and update the fuzzy layer membership function parameters, fuzzy rules, and anti-fuzzy function parameters of the fuzzy neural network accordingly, thereby updating the parameters of the fuzzy neural network.

[0060] Specifically, the revised calculation formula is:

[0061] y new =y×(1+α×(NTU-y))

[0062] Among them, y is the floc value before correction and compensation, a is the process compensation parameter, NTU is the turbidity meter data, and ynew is the floc value after correction and compensation.

[0063] S207, using the fuzzy neural network to return to the starting step S201 and restart, and perform iterative training until the number of iterations is greater than 5000 times or the aforementioned difference is less than 0.01, and then the training is terminated to obtain the alum flower state discrimination network.

[0064] S208, inputting the collected water image into the alum floc state discrimination network, obtaining the alum floc value and the amount of coagulant added based on reasonable matching data.

[0065] By using this method, a combination of machine vision and artificial intelligence can be used to process the captured alum flower images to obtain more objective indicators in real time. By matching it with the amount of coagulant added, it can also obtain Figure 4 The reasonable matching curve shown is used as a guide for the dosage of coagulant, making the addition of coagulant simple and feasible. It is no longer necessary to rely on professional knowledge and experience to judge the dosage. The external automatic control system can also be used to guide the automatic control system to complete the automatic addition of coagulant.

[0066] Figure 2Another embodiment of the underwater alum bloom state recognition method of the present invention is shown. The difference from the first embodiment is that the second embodiment directly uses the alum bloom state discrimination network that has been trained, so there is no need for corresponding training steps, and the alum bloom value and reasonable matching data are directly obtained. The steps include:

[0067] S211, collecting underwater alum flower images;

[0068] S212, preprocessing the alum flower image to obtain a region of interest map;

[0069] S213, the alum flower state recognition network samples the region of interest map multiple times to obtain sampling maps of multiple scales;

[0070] S214, the alum flower state recognition network selects feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images to obtain a feature vector composed of feature extraction results;

[0071] S215, the alum flower state recognition network screens the feature vector to obtain a feature vector with high correlation;

[0072] S216. The alum floc state recognition network performs fuzzy processing on the highly correlated feature vectors to obtain the alum floc value and the amount of coagulant added.

[0073] The present invention also includes an alum bloom monitoring device using the underwater alum bloom state recognition method. Figure 3 The structure diagram of the device embodiment is shown, which includes:

[0074] The acquisition module S101 is used to perform image preprocessing on the acquired underwater alum flower image to obtain an area of ​​interest map;

[0075] The sampling module S102 is used to input the region of interest map into the convolutional neural network, sample the region of interest map multiple times, and obtain sampling maps of multiple scales;

[0076] A feature extraction module S103 is used to select feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images to obtain a feature vector composed of extraction results corresponding to each feature description factor;

[0077] The feature processing module S104 is used for the fuzzy neural network to screen the feature vector to obtain a feature vector with high correlation, and input the feature vector with high correlation into the fuzzy layer of the fuzzy neural network to obtain the alum flower value;

[0078] The result processing module S105 is used to correct the floc value, record the corrected floc value and the amount of coagulant added in the water body into a reasonable matching data sequence, and update the parameters of the fuzzy neural network;

[0079] Return module S106, used to use the updated fuzzy neural network to return to the starting step and restart, and perform iterative training until the set termination condition is met, and then the training is terminated to obtain the alum flower state discrimination network;

[0080] The detection module S107 is used to input the collected water body image into the alum bloom state discrimination network to obtain the alum bloom value and the amount of coagulant added.

[0081] Optionally, an early warning module is also included, which can issue an early warning reminder for the alum flower value exceeding the warning value, such as: lighting an alarm and / or sounding an alarm.

[0082] Each module in the alum bloom monitoring device using the underwater alum bloom state identification method can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the modules. In one embodiment, a computer device is provided, which can be a terminal, and its internal structure diagram can be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for identifying the state of underwater alum flowers is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc. Those skilled in the art can understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. Based on the example described above, in one embodiment, a computer device is also provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any one of the underwater alum bloom state recognition methods in the above-mentioned embodiments is implemented.

[0083] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. For example, in the embodiment of the present invention, the program can be stored in the storage medium of the computer system and executed by at least one processor in the computer system to implement the process of the embodiment including the above-mentioned underwater alum bloom state identification method. Wherein, the storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM). Accordingly, in one embodiment, a computer storage medium or a computer-readable storage medium is also provided, on which a computer program is stored, wherein when the program is executed by the processor, any one of the underwater alum bloom state identification methods in the above-mentioned embodiments is implemented.

[0084] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered as the scope of this specification. It should be noted that the term "first\second\third" involved in the embodiment of the present application is only to distinguish similar objects, and does not represent a specific order for the object. It can be understood that the "first\second\third" can be interchanged with a specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged in appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. The terms "including" and "having" in the embodiments of the present application and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment comprising a series of steps or modules is not limited to the listed steps or modules, but optionally also includes steps or modules that are not listed, or optionally also includes other steps or modules inherent to these processes, methods, products or equipment. The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for identifying underwater alum bloom status, characterized in that: The steps include: Perform image preprocessing on the collected underwater alum flower images to obtain a region of interest map; Inputting the region of interest map into a convolutional neural network, sampling the region of interest map multiple times to obtain sampling maps of multiple scales; Selecting feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images to obtain a feature vector composed of extraction results corresponding to each feature description factor; The fuzzy neural network screens the feature vectors to obtain feature vectors with high correlation, and inputs the feature vectors with high correlation into the fuzzy layer of the fuzzy neural network to obtain the alum value; The floc value is corrected with reference to the turbidity meter data, the corrected floc value and the amount of coagulant added in the water body are recorded in a reasonable matching data sequence, and the parameters of the fuzzy neural network are updated; The updated fuzzy neural network is used to return to the starting step and start again, and iterative training is performed until the set termination condition is met, and the training is terminated to obtain the alum flower state discrimination network; The collected water body images are input into the alum bloom state discrimination network to obtain the alum bloom value and the amount of coagulant added; The reference turbidity meter data is used to correct the floc value, including: using a turbidity meter to detect the water body to obtain turbidity meter data, calculating the difference between the turbidity meter data and the floc value, and making corresponding compensation for the floc value according to the process data of the water plant; the calculation formula used is: and new =y×(1+α×(NTU-y)), Among them, y is the alum flower value before correction compensation, α is the process compensation parameter, NTU is the turbidity meter data, y new is the alum flower value after correction and compensation; The parameters of the fuzzy neural network include membership function parameters of the fuzzy layer, fuzzy rules, and parameters of the anti-fuzzy function.

2. The underwater alum bloom state identification method according to claim 1, characterized in that: Screening the feature vectors to obtain feature vectors with high correlation includes: firstly screening the feature vectors with high correlation according to the feature vectors, and then adjusting the weights corresponding to the features in the feature vectors with high correlation.

3. The underwater alum bloom state identification method according to any one of claims 1 to 2, characterized in that: The method for sampling the region of interest map is to sequentially pass the region of interest map through a convolutional layer and a pooling layer of a convolutional neural network to obtain a sampling map.

4. The underwater alum bloom state identification method according to any one of claims 1 to 2, characterized in that: Selecting feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images includes: using local binary patterns to calculate texture information corresponding to the feature description factors in the sampling images to obtain extraction results.

5. Alum flower monitoring device, characterized in that: include: The acquisition module is used to perform image preprocessing on the acquired underwater alum flower images to obtain an area of ​​interest map; A sampling module is used to input the region of interest map into the convolutional neural network, sample the region of interest map multiple times, and obtain sampling maps of multiple scales; A feature extraction module, used to select feature description factors to perform global feature extraction and / or local feature extraction on the obtained multiple sampling images, and obtain a feature vector composed of extraction results corresponding to each feature description factor; A feature processing module is used for the fuzzy neural network to screen the feature vector to obtain a feature vector with high correlation, and input the feature vector with high correlation into the fuzzy layer of the fuzzy neural network to obtain a flocculant value; A result processing module is used to correct the floc value with reference to the turbidity meter data, record the corrected floc value and the amount of coagulant added in the water body into a reasonable matching data sequence, and update the parameters of the fuzzy neural network; A return module is used to use the updated fuzzy neural network to return to the starting step and start again, and perform iterative training until the set termination condition is met, and then the training is terminated to obtain the alum flower state discrimination network; The detection module is used to input the collected water body image into the alum bloom state discrimination network to obtain the alum bloom value and the amount of coagulant added; The reference turbidity meter data is used to correct the floc value, including: using a turbidity meter to detect the water body to obtain turbidity meter data, calculating the difference between the turbidity meter data and the floc value, and making corresponding compensation for the floc value according to the process data of the water plant; the calculation formula used is: ynew=y×(1+α×(NTU-y)), Among them, y is the alum flower value before correction compensation, α is the process compensation parameter, NTU is the turbidity meter data, y new is the alum flower value after correction and compensation; The parameters of the fuzzy neural network include membership function parameters of the fuzzy layer, fuzzy rules, and parameters of the anti-fuzzy function.

6. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images