Method and system for identifying alumen ustum on surface of sewage

Abstract

The invention relates to the technical field of image processing, and discloses a sewage surface alumen ustum identification method and system. The sewage surface alumen ustum recognition method comprises the steps that regulation and control parameters and original water quality parameters in the alumen ustum forming process are obtained; extracting alumen ustum characteristics of the coagulative precipitation process video, and identifying the optimized key frame as a to-be-identified image based on a genetic algorithm; training a CNN neural network model by taking the regulation and control parameters, the original water quality parameters and the key frame image data as inputs and taking the alumen ustum quality as an output to obtain an alumen ustum quality identification model; the suspended solid content is obtained through a suspended solid sensor, and the alum peanut forming stage is determined; judging whether the alumen ustum quality recognition result is qualified or not based on the alumen ustum quality recognition model in the transition stage; and when the judgment result is no, adjusting the parameters according to the alumen ustum quality identification result. The accuracy of various parameters in the coagulating sedimentation process is judged and optimized according to the quality of alumen ustum, and the sewage treatment efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a method and system for identifying alum bloom on the surface of sewage. Background Art

[0002] The water treatment process of a water plant generally includes coagulation, sedimentation, filtration, disinfection and other links. The coagulation and sedimentation link is used to remove suspended particles, colloidal substances and some soluble organic matter in the water. Coagulants are added to the water to make the fine suspended particles and colloidal substances in the water aggregate to form larger flocs (i.e., alum flocs), and gravity is used to precipitate them and separate them from the water.

[0003] The formation of alum flocs is the result of charge neutralization and adsorption bridging during the coagulation process. Ideal alum flocs should have moderate size, regular shape and good sedimentation performance. However, affected by factors such as the type of coagulant (such as aluminum salt, iron salt, polyacrylamide), the amount of coagulant added, the stirring speed, and the original water quality, the alum flocs have diverse morphologies, ranging from small and loose micro-flocs to larger and dense flocs, thus affecting the overall water treatment effect. Therefore, by effectively identifying alum flocs, it is convenient to timely control and adjust the coagulation and sedimentation links to improve the efficiency of alum floc formation.

[0004] Existing image recognition technology and sensors are difficult to accurately distinguish different forms of alum flowers, especially when they are mixed. The shape of alum flowers is usually irregular, snowflake-shaped or lumpy, which makes it difficult to apply traditional mathematical models based on regular shapes (such as spheres). It is also difficult to extract effective features from complex forms for identification and classification. In addition, the water quality parameters of raw water (such as turbidity, pH value, temperature, dissolved organic matter content, etc.) will also significantly affect the formation and morphology of alum flowers. Under different water quality conditions, the same dose of coagulant will produce completely different effects. Moreover, the formation of alum flowers has a hysteresis, and it is difficult to accurately identify the adjustment time node, and the adjustment efficiency is low. Therefore, how to efficiently identify alum flowers has become an urgent problem to be solved. Summary of the invention

[0005] The present invention aims to provide a method and system for identifying alum flocs on the surface of sewage, obtain key frames of the alum floc formation process through a genetic algorithm, identify the alum floc quality based on the key frames, and judge and optimize the accuracy of various parameters in the coagulation and sedimentation process through the alum floc quality to improve the sewage treatment efficiency.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: A method for identifying alum bloom on the surface of sewage, comprising: Data acquisition step, obtaining the control parameters and original water quality parameters in the process of alum floc formation; The image acquisition step includes collecting a video of the coagulation and sedimentation process, extracting the features of the alum flower, constructing a key frame optimization model based on a genetic algorithm, determining a fitness function based on the features of the alum flower, identifying and optimizing key frames through the key frame optimization model, and extracting the key frames as images to be identified; The model training step takes the control parameters, original water quality parameters and key frame image data as input, takes the alum flower quality as output, trains the CNN neural network model, and obtains the alum flower quality recognition model; A stage identification step, obtaining the suspended solid content through a suspended solid sensor, and determining the alum floc formation stage through the suspended solid content, wherein the alum floc formation stage includes a reaction stage and a transition stage, wherein the transition stage is a stage where the alum floc stage changes; A judgment step is to identify the quality of the alum flower in the transition stage, obtain the alum flower quality identification result of the current stage based on the alum flower quality identification model, and judge whether the alum flower quality identification result is qualified; In the control step, when the judgment result is no, the parameters are adjusted according to the alum flower quality identification result.

[0007] The principles and advantages of this scheme are as follows: in actual application, the data acquisition step obtains the control parameters and original water quality parameters in the process of alum floc formation, covering multiple key factors affecting alum floc formation, and providing a rich information basis for subsequent analysis; the image acquisition step uses computer vision technology to automatically extract features and identify key frames, reducing human intervention, and customizing the selection of key frames that best reflect the changes in the state of alum flocs through genetic algorithms, thereby improving the relevance and effectiveness of the data; the model training step uses multimodal data input to facilitate the model to learn the complex relationship between information from different sources, and can consider multiple types of data at the same time, making full use of all available information, improving prediction accuracy, and adjusting the model structure and output form according to specific needs, which is suitable for different application scenarios; in the stage identification step, the alum floc generation stage is identified. The alum floc generation stage includes a reaction stage and a transition stage. The transition stage is the stage where the alum floc stage changes. The alum floc quality is identified in the transition stage, thereby ensuring the integrity of the feature collection in the previous stage and ensuring that the identification is timely. Use the trained model to identify the quality of the alum flocs at the current stage, and judge whether it is qualified based on the set standards. The evaluation results of the alum flocs quality can be quickly obtained, which is convenient for taking measures in a timely manner. The results given by the model rather than subjective judgments ensure the consistency and fairness of the evaluation; the above process is continuously cycled through the control steps to gradually optimize the entire coagulation and sedimentation process and improve the stability and efficiency of the system. In this application, the accuracy of various parameters in the coagulation and sedimentation process is judged and optimized by the quality of the alum flocs, which can ensure the coagulation effect and the quality of the effluent, while overcoming the problem of hysteresis in the formation of alum flocs.

[0008] Preferably, as an improvement, the image acquisition step includes: The stage acquisition sub-step is to acquire the current alum flower formation stage, wherein the alum flower formation stage includes a rapid mixing stage after the addition of coagulant, a slow mixing stage, a sedimentation stage, and a final inspection stage before filtration; In the key frame optimization model construction sub-step, the fitness function and constraints are defined in combination with the alum flower generation stage, and the initial frame combination is randomly selected as the initial individual of the population. The population is optimized through selection, crossover and mutation operations to find the best key frame combination.

[0009] Technical effect: Through the key frame optimization model, it is easy to find the best key frame combination, so as to obtain sufficient information while reducing the amount of image recognition and saving resources.

[0010] Preferably, as an improvement, the alum flower characteristics include morphological characteristics, texture characteristics, and aggregation density characteristics of alum flowers.

[0011] Technical effect: It is convenient to fully obtain the characteristics of alum flowers.

[0012] Preferably, as an improvement, the fitness function includes:

[0013]

[0014]

[0015] in, , , is the weight value, and are the pixel values ​​of the i-th frame and the i-1-th frame respectively, and N is the total number of pixels; is the time decay factor, T is the time window size, is the change of image features within the time window T; represents the difference between the image morphology j of the i-th frame and the i-1-th frame, is the weight value of morphology j in the alum flower generation stage d; Indicates the amount of color change, Indicates the amount of texture change, is the balance coefficient.

[0016] Technical effect: Through the fitness function, comprehensive consideration is given to multiple angles such as morphological changes, color and texture changes, time and space information, and pixel quality to capture the optimal key frame combination.

[0017] Preferably, as an improvement, the constraint conditions include:

[0018]

[0019]

[0020] in, is the index of the kth keyframe, is the minimum interval set, is the number of keyframes actually selected, is the maximum number of settings, is the quality score of the i-th frame, Minimum quality standard.

[0021] Technical effect: The above constraints can prevent key frames from being too dense, while limiting the number of key frames to ensure that the key frames have high image quality.

[0022] Technical effect: It is easy to carry out feature extraction from static and dynamic perspectives, thereby improving feature integrity and accuracy.

[0023] Preferably, as an improvement, the model training step includes: The model selection sub-step obtains the training data set and model parameters based on the current bloom generation stage; The image preprocessing sub-step is to enhance the image and label it; In the model training sub-step, a branch structure is created to process numerical data and image data, the outputs of the numerical data branch and the image data branch are concatenated to form the final feature representation, an output layer is added to the concatenated features according to task requirements, and model training is performed.

[0024] Technical effect: It is convenient to adopt different models in different stages, highlight the stage characteristics, and improve the accuracy of alum flower recognition.

[0025] Preferably, as an improvement, the model training step also includes a data enhancement sub-step, which randomly transforms the input data to obtain an enhanced data set.

[0026] Technical effect: Generate more training samples, thereby enhancing the generalization ability of the model.

[0027] Preferably, as an improvement, the model training step also includes a sub-step of dynamically adjusting feature weights, which dynamically adjusts the importance weights of each input feature by learning a meta-model, and gradually optimizes the combination of input features according to environmental feedback using a reinforcement learning algorithm, and dynamically adjusts the input feature weights of the model based on the optimization results.

[0028] Technical effect: It is easy to improve the performance of the model when dealing with complex and changing tasks.

[0029] Preferably, as an improvement, the alum flower quality identification model also includes a secondary output layer. When the judgment result obtained in the judgment step is no, the secondary output is triggered, and the secondary output is a control method.

[0030] Technical effect: Easy to make adjustments quickly.

[0031] It also includes a system for identifying alum bloom on the surface of sewage, which uses the method for identifying alum bloom on the surface of sewage. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The figure is a flow chart of a method for identifying alum bloom on the surface of sewage. DETAILED DESCRIPTION

[0033] The following is further described in detail through specific implementation methods: The embodiment is basically as shown in the attached Figure 1 As shown, a method for identifying alum bloom on the surface of sewage comprises: Data acquisition steps, obtain the control parameters and original water quality parameters in the process of alum floc formation; the control parameters include the type and dosage of coagulant, stirring speed, and water flow rate; the original water quality parameters include turbidity, pH value, temperature, and dissolved organic matter content. The control parameters and original water quality parameters are closely related to the formation of alum flocs. The type and dosage of coagulant are determined according to the quality of raw water. Many particles in the water carry negative charges, repel each other and are difficult to settle. After adding positively charged coagulants, the negative charges on the surface of the particles can be neutralized, reducing electrostatic repulsion and promoting collision and aggregation between particles; some polymer coagulants (such as polyacrylamide) have long-chain structures and can "bridge" multiple particles together through physical adsorption to form larger flocs; different coagulants have different optimal pH ranges. For example, aluminum salt coagulants work best between pH 6.0-7.5, while iron salt coagulants work best at pH 6.0-7.5. The effect is better between 5.0-8.5; fast stirring helps the coagulant to disperse evenly, while slow stirring helps the formation and growth of alum flocs; the higher the temperature, the stronger the Brownian motion between particles, which is conducive to the formation of alum flocs, but too high a temperature will also lead to incomplete hydrolysis of the coagulant; too much or too little coagulant will affect the quality of alum flocs, too much will lead to counter-ion effect, causing the particles to disperse again; too little will not effectively neutralize the negative charge on the particle surface. By obtaining the control parameters and original water quality parameters, it is convenient to analyze and adjust the influencing factors when the alum floc quality is poor.

[0034] The stage identification step is to obtain the suspended solid content through the suspended solid sensor, and determine the alum floc formation stage through the suspended solid content. The alum floc formation stage includes a reaction stage and a transition stage, and the transition stage is the stage where the alum floc stage changes. The transition stage is used for subsequent alum floc quality identification. The reaction stage is the rapid mixing stage after the addition of the coagulant, the slow mixing stage, the sedimentation stage, and the final inspection stage before filtration. In this embodiment, the alum floc formation stage is identified through the suspended solid content, and the alum floc reaction stage is identified by setting the suspended solid content threshold range. The stage in which the suspended solid content fluctuates up and down the threshold or the suspended solid content reaches the transition threshold is defined as the transition stage, and the transition threshold is determined based on historical data.

[0035] The suspended solids content will change significantly in each stage of alum floc formation. In the rapid mixing stage after the addition of coagulant, some tiny particles will temporarily disperse due to the effect of the coagulant, resulting in a slight increase in the suspended solids content or a relatively stable level. At this stage, the suspended solids content in the water is monitored by an online suspended solids sensor. At the end of the rapid mixing stage, the suspended solids content remains at a relatively stable level or increases slightly; compared with the suspended solids content before treatment, the suspended solids content at this stage has not decreased significantly or increased significantly. In the slow mixing stage, the primary coagulants gradually gather to form larger, loosely structured alum flocs. With the formation of alum flocs, the suspended solid particles in the water begin to combine together to form larger particle aggregates. As the alum flocs grow, the suspended solids content will gradually decrease, because some particles have been combined into larger alum flocs, reducing the amount of suspended solids. The changes in the suspended solids content are continuously monitored, and the suspended solids content gradually decreases at this stage. The alum flocs formed in the sedimentation stage further mature and separate from the water to achieve solid-liquid separation. At this time, the suspended solids content drops significantly, and most of the alum flocs have settled to the bottom of the water. As time goes by, samples are taken at different depths of the sedimentation tank to measure the suspended solids content in the supernatant. The suspended solids content in the supernatant is close to the lowest point in this stage; in the final inspection stage before filtration, most of the alum flocs in the water should have settled, but there are still a small amount of residual particles. The suspended solids content in the water is accurately measured based on a high-precision suspended solids sensor. The suspended solids content at this stage should be very low (such as less than 5 mg / L).

[0036] The image acquisition step collects the video of the coagulation and sedimentation process, builds a key frame optimization model based on a genetic algorithm, identifies and optimizes key frames through the key frame optimization model, and extracts key frames as images to be identified; the image acquisition step includes a stage acquisition sub-step and a key frame optimization model construction sub-step.

[0037] The stage acquisition sub-step acquires the current alum floc formation stage, which includes the rapid mixing stage after the addition of coagulant, the slow mixing stage, the sedimentation stage, and the final inspection stage before filtration. The identification of alum flocs is carried out at several key stages in the tap water treatment process, and each stage has its specific purpose and significance. By identifying at these stages, the effectiveness and stability of the coagulation and sedimentation process can be ensured, thereby improving the efficiency of the entire water treatment system and the effluent quality.

[0038] The rapid mixing stage after the addition of coagulant is the initial stage of the coagulation reaction, which usually lasts for 30 seconds to 2 minutes. After the coagulant is added to the raw water, rapid mixing is required to allow the coagulant to be quickly and evenly dispersed in the water and fully contact with the suspended particles in the water. By identifying the water sample after mixing at this stage, it is possible to determine whether the coagulant is evenly dispersed, whether there is a local concentration that is too high or too low, and preliminarily observe whether microflocs are beginning to form, as well as their shape and size. Ensuring that the coagulant is evenly dispersed and the initial reaction is good is the basis for the success of subsequent steps. After rapid mixing, it enters the slow mixing stage (also called the flocculation stage). In this stage, slow stirring causes the micro-flocculation to gradually grow and form larger alum flowers. In this stage, the growth of the alum flowers is evaluated, and the size, shape and density of the alum flowers are observed to determine whether they are suitable for sedimentation. The stirring conditions are optimized, and the stirring speed and time are adjusted according to the shape of the alum flowers to ensure the best flocculation effect. The slow mixing stage is the key stage in the formation of alum flowers, which directly determines the quality and sedimentation performance of the alum flowers. By identifying problems at this stage, problems (such as insufficient or excessive stirring) can be discovered and corrected in a timely manner to ensure that the formed alum flowers have good sedimentation characteristics. This stage usually lasts 15-30 minutes. After slow mixing, the water sample enters the sedimentation tank, and the alum flocs settle to the bottom of the tank under the action of gravity, forming a sludge layer, and the clean water flows out from the top. The sedimentation stage is the key link to remove suspended matter, which directly affects the turbidity and water quality of the effluent. By identifying at this stage, the operation status of the sedimentation tank is ensured to avoid the decline of effluent quality due to too thick sludge layer or poor sedimentation of alum flocs. Identification is usually carried out within 5-10 minutes after the start of sedimentation. After sedimentation, the water sample will enter the filtration system for further purification. The final inspection before this is to ensure that the water sample after sedimentation meets the standards for entering the filtration system to prevent filter clogging. Identification is carried out before entering the filtration system after the sedimentation is completed, including effluent turbidity and fine alum flocs or particles.

[0039] In the key frame optimization model construction sub-step, the fitness function and constraints are defined in combination with the alum flower generation stage, and the initial frame combination is randomly selected as the initial individual of the population. The population is optimized through selection, crossover and mutation operations to find the best key frame combination.

[0040] The alum flower characteristics include the morphological characteristics, texture characteristics, and aggregation density characteristics of the alum flower. The morphological characteristics include the area, perimeter, and shape factor of the alum flower:

[0041]

[0042]

[0043] Among them, A is the area of ​​the alum flower, P is the perimeter, is the area of ​​the alum flower, and C is the boundary of the alum flower; It is the shape factor, which is used to describe the regularity of the alum flower. For round alum flowers, the shape factor is close to 1; for irregular shapes, the shape factor is smaller.

[0044] Texture features include:

[0045] in, is the gray value of the center pixel, is the gray value of the neighborhood pixel, and s() is the sign function.

[0046] The aggregation density feature is the degree of aggregation of alum flowers in a certain area, which is closely related to the effect of the coagulant. The number of alum flowers in a specific area is counted using the connected component labeling algorithm, and the number of alum flowers per unit area or the coverage area ratio is calculated to obtain:

[0047] in, is the number of alum flowers, is the total area considered.

[0048] The fitness function includes:

[0049]

[0050]

[0051] in, , , is the weight value, and are the pixel values ​​of the i-th frame and the i-1-th frame respectively, and N is the total number of pixels. Considering that the formation of alum flowers is a dynamic process, it is necessary not only to pay attention to the feature changes within a single frame, but also to consider the changing trend in the time series. Time series analysis methods (such as autoregressive model AR, moving average model MA, etc.) are used to analyze the time series characteristics of alum flower formation. is the time decay factor, which is used to emphasize the importance of recent frames relative to earlier frames, T is the time window size, It is the change of image features within the time window T. By considering the time factor, the trend of bloom generation in the time dimension is captured to provide a more comprehensive perspective to select key frames. represents the difference between the image morphology j of the i-th frame and the i-1-th frame, is the weight value of morphology j in the alum flower generation stage d. During the alum flower generation process, its morphology will undergo significant changes. These changes are used to define the fitness function to capture the key morphological transition points. The importance of the frame is evaluated by comparing the changes in the morphological characteristics (area, perimeter and shape factor) of the alum flower area between adjacent frames. It can effectively capture the moment when the morphology changes significantly during the alum flower generation process, thereby helping to select the most representative key frame; Indicates the amount of color change, Indicates the amount of texture change. The color and texture of the alum flower are also important features. As the alum flower grows, its color becomes darker and the texture becomes more complex. Gray-level co-occurrence matrix (GLCM) or local binary pattern (LBP) is used to extract texture features, and the color change is analyzed in combination with the color histogram. It is the balance coefficient, which is used to adjust the proportion of the influence of color and texture on fitness.

[0052] The constraints include:

[0053]

[0054]

[0055] in, is the index of the kth keyframe, It is the minimum interval set. In order to avoid selecting frames that are too close as key frames, a minimum interval constraint is set to ensure that there is at least a fixed number of frames between every two key frames. This constraint can prevent key frames from being too dense and improve the efficiency of key frame selection. is the number of keyframes actually selected, It is the maximum number set. In order to control the number of key frames and avoid excessive burden of subsequent processing caused by too many key frames, the maximum number constraint is set to ensure the efficiency of the processing process; is the quality score of the i-th frame, As the minimum quality standard, in order to ensure the quality of the key frames, according to quality indicators such as clarity and contrast, ensure that the selected key frames have high image quality, which is convenient for subsequent feature extraction and analysis.

[0056] The selection operation is based on the fitness function, and individuals with higher fitness are selected as the parents of the next generation. The preferred selection method in this application includes but is not limited to the roulette selection method:

[0057] in, is the probability that the ith individual is selected.

[0058] The crossover operation generates a new individual by exchanging some genes of two parents. In this application, the one-point crossover method is preferably used:

[0059] The mutation operation increases the diversity of the population by randomly changing some genes of individuals. In this application, the mutation rate is set to 0.1.

[0060] The model training step takes the control parameters, original water quality parameters and key frame image data as input, takes the alum flower quality as output, trains the CNN neural network model, and obtains the alum flower quality recognition model; specifically, the model training step includes an image preprocessing sub-step, a model selection sub-step and a model training sub-step.

[0061] The model selection sub-step obtains the training data set and model parameters based on the current alum flower generation stage and labels them; the model parameters include feature weight parameters, so that different stages can focus on different influencing factors and improve the accuracy of the model.

[0062] In the image preprocessing sub-step, the image is enhanced and labeled; the enhancement process includes background removal and image enhancement. The background noise is removed by the mixed Gaussian model and the foreground alum flowers are highlighted:

[0063] in, is the current frame, is the background model, It is the foreground alum flower.

[0064] Use histogram equalization, filtering and denoising to enhance image contrast and clarity. Specifically:

[0065] in, is the grayscale value of the input image, is the grayscale value of the output image, is the grayscale probability density function of the input image.

[0066] The model training sub-step creates a branch structure to process numerical data and image data, uses a fully connected layer to process the standardized numerical data, the numerical data includes control parameters and original water quality parameters, uses a CNN convolutional neural network to extract image features, splices the outputs of the numerical data branch and the image data branch to form a final feature representation, and adds an output layer to the spliced ​​features according to task requirements. In this embodiment, the task requirement is multi-classification, that is, different levels of quality assessment. The loss function, optimizer and evaluation index are selected, and the model training is performed.

[0067] The model training step also includes a data enhancement sub-step, which performs random transformation on the input data, mainly including random transformation on the input image data, and the random transformation includes rotation, flipping, and cropping to obtain an enhanced data set.

[0068] The model training step also includes a sub-step of dynamically adjusting feature weights, which dynamically adjusts the importance weights of each input feature by learning a "meta-model", gradually optimizes the combination of input features based on environmental feedback using a reinforcement learning algorithm, and dynamically adjusts the input feature weights of the model based on the optimization results. Specifically: If the model is g(x;u), where x=[x1,x2,...,x n ] represents the input feature, u=[u1,u2,...,u n ] represents the corresponding weight vector, and the goal is to find the optimal weight vector u * .

[0069] The input features are weighted and then fed into the model:

[0070] Define the loss function to measure the difference between the model recognition result value and the true label, and find the optimal weight vector by optimizing the target:

[0071] in, ) is the loss function, is the optimal weight vector.

[0072] By dynamically adjusting the weights of input features, the model can automatically adjust its focus on different features according to specific application scenarios or data characteristics, thereby improving the adaptability and robustness of the model. For example, in the water quality treatment process, water quality parameters are more critical than control parameters in certain periods, while the opposite is true in other periods. Dynamically adjusting weights can help the model better capture these changes.

[0073] A judging step is to identify the quality of alum flocs in the transition stage, obtain the alum floc quality identification result of the current stage based on the alum floc quality identification model, and judge whether the alum floc quality identification result is qualified; a control step is to adjust parameters according to the alum floc quality identification result when the judgment result is no, the alum floc quality identification model also includes a secondary output layer, when the judgment result obtained in the judging step is no, the secondary output is triggered, the secondary output is a control method, and the control method includes adjusting the coagulant dosage, optimizing the stirring conditions, and adjusting the pH value.

[0074] It also includes a system for identifying alum bloom on the surface of sewage, which uses the method for identifying alum bloom on the surface of sewage.

[0075] The above is only an embodiment of the present invention, and the common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the technical solution of the present invention, several modifications and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for identifying alum bloom on the surface of sewage, characterized in that: include: Data acquisition step, obtaining the control parameters and original water quality parameters in the process of alum floc formation; The image acquisition step includes collecting a video of the coagulation and sedimentation process, extracting the features of the alum flower, constructing a key frame optimization model based on a genetic algorithm, determining a fitness function based on the features of the alum flower, identifying and optimizing key frames through the key frame optimization model, and extracting the key frames as images to be identified; The model training step takes the control parameters, original water quality parameters and key frame image data as input, takes the alum flower quality as output, trains the CNN neural network model, and obtains the alum flower quality recognition model; A stage identification step, obtaining the suspended solid content through a suspended solid sensor, and determining the alum floc formation stage through the suspended solid content, wherein the alum floc formation stage includes a reaction stage and a transition stage, wherein the transition stage is a stage where the alum floc stage changes; A judgment step is to identify the quality of the alum flower in the transition stage, obtain the alum flower quality identification result of the current stage based on the alum flower quality identification model, and judge whether the alum flower quality identification result is qualified; In the control step, when the judgment result is no, the parameters are adjusted according to the alum flower quality identification result.

2. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that, The image acquisition step comprises: The stage acquisition sub-step is to acquire the current alum flower formation stage, wherein the alum flower formation stage includes a rapid mixing stage after the addition of coagulant, a slow mixing stage, a sedimentation stage, and a final inspection stage before filtration; In the key frame optimization model construction sub-step, the fitness function and constraints are defined in combination with the alum flower generation stage, and the initial frame combination is randomly selected as the initial individual of the population. The population is optimized through selection, crossover and mutation operations to find the best key frame combination.

3. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that: The alum flower characteristics include morphological characteristics, texture characteristics, and aggregation density characteristics of alum flowers.

4. A method for identifying alum bloom on sewage surface according to claim 2, characterized in that, The fitness function includes: in, , , is the weight value, and are the pixel values ​​of the i-th frame and the i-1-th frame respectively, and N is the total number of pixels; is the time decay factor, T is the time window size, is the change of image features within the time window T; represents the difference between the image morphology j of the i-th frame and the i-1-th frame, is the weight value of morphology j in the alum flower generation stage d; Indicates the amount of color change, Indicates the amount of texture change, is the balance coefficient.

5. A method for identifying alum bloom on sewage surface according to claim 2, characterized in that, The constraints include: in, is the index of the kth keyframe, is the minimum interval set, is the number of keyframes actually selected, is the maximum number set, is the quality score of the i-th frame, Minimum quality standard.

6. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that, The model training step includes: The model selection sub-step obtains the training data set and model parameters based on the current bloom generation stage; The image preprocessing sub-step is to enhance the image and label it; In the model training sub-step, a branch structure is created to process numerical data and image data, the outputs of the numerical data branch and the image data branch are concatenated to form the final feature representation, an output layer is added to the concatenated features according to task requirements, and model training is performed.

7. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that: The model training step also includes a data enhancement sub-step, which randomly transforms the input data to obtain an enhanced data set.

8. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that: The model training step also includes a feature weight dynamic adjustment sub-step, which dynamically adjusts the importance weight of each input feature by learning a meta-model, uses a reinforcement learning algorithm to gradually optimize the combination of input features according to environmental feedback, and dynamically adjusts the input feature weights of the model based on the optimization results.

9. A method for identifying alum bloom on sewage surface according to claim 1, characterized in that: The alum flower quality identification model also includes a secondary output layer. When the judgment result obtained in the judgment step is no, the secondary output is triggered, and the secondary output is a control method.

10. A system for identifying alum bloom on the surface of sewage, characterized in that: A method for identifying alum bloom on the surface of sewage as described in any one of claims 1 to 9 is used.

Images