Image data processing method and system for digital management of environmental protection equipment

Abstract

The invention relates to the technical field of image processing, in particular to an image data processing method and system for digital management of environmental protection equipment, and the method comprises the steps: collecting a real-time image of an aeration area through a waterproof camera disposed above an aeration tank, detecting the definition of the camera regularly, and carrying out the cleaning when the definition is lower than a preset value; a self-adaptive Retinex algorithm is adopted for the real-time image of the aeration area to eliminate water surface reflection interference, noise caused by water turbidity is reduced through a multi-frame weighted average method, and a processed image is obtained; segmenting a bubble region in the image based on a U-Net network and calculating a bubble distribution parameter; comparing the bubble distribution parameter with a bubble distribution threshold value, and generating an aeration rate optimization instruction according to average dissolved oxygen data obtained by a dissolved oxygen sensor; and controlling the air blower based on the aeration rate optimization instruction. The power of the air blower can be better adjusted according to the bubble condition of the aeration tank, so that more energy is saved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an image data processing method and system for digital management of environmental protection equipment. Background Art

[0002] In the sewage treatment process, the aeration tank is a key component, mainly used to supply the oxygen required for microorganisms to degrade organic matter and keep the mixed liquid in a suspended state to promote biological reactions. However, the energy consumption problem of the aeration tank is particularly prominent, among which the blower as an oxygen supply equipment is the main source of energy consumption. For urban sewage treatment plants that use the activated sludge treatment process, the electricity consumed by the blower operation can account for approximately 60% of the electricity consumption of the entire sewage treatment plant. This is because the aeration system needs to continuously supply a large amount of air to the aeration tank to ensure that there is enough oxygen supplied to the microorganisms for metabolism. Therefore, reducing the energy consumption of the blower and improving its operating efficiency have become important measures for sewage treatment plants to save energy, reduce consumption and improve economic benefits.

[0003] The existing method is to use a variable frequency drive (VFD) or speed control device to adjust the blower speed and power output according to the set control strategy. This cannot be flexibly adjusted according to the actual situation in the aeration tank, and the energy saving effect is limited. Summary of the Invention

[0004] The purpose of the present invention is to provide an image data processing method and system for digital management of environmental protection equipment, aiming to better adjust the power of the blower according to the bubble situation in the aeration tank, thereby saving more energy.

[0005] To achieve the above objectives, in a first aspect, the present invention provides an image data processing method for digital management of environmental protection equipment, comprising: collecting real-time images of the aeration area by a waterproof camera deployed above the aeration tank, and regularly detecting the clarity of the camera, and cleaning the image when the clarity is lower than a preset value;

[0006] The real-time image of the aeration area is subjected to an adaptive Retinex algorithm to eliminate the interference of water surface reflections, and a multi-frame weighted averaging method is used to reduce the noise caused by water turbidity to obtain a processed image;

[0007] Based on the U-Net network, the bubble area in the image is segmented and the bubble distribution parameters are calculated, including the bubble distribution density, average diameter and rising speed;

[0008] Comparing the bubble distribution parameter with the bubble distribution threshold and generating an aeration optimization instruction based on average dissolved oxygen data obtained from a dissolved oxygen sensor;

[0009] The blower is controlled based on the aeration rate optimization instruction.

[0010] After the blower is controlled based on the aeration volume optimization instruction, the method further includes establishing a bubble distribution energy consumption correlation model, and dynamically optimizing the bubble distribution threshold using the blower current value as feedback.

[0011] The specific steps of collecting real-time images of the aeration area by deploying a waterproof camera above the aeration tank and regularly testing the clarity of the camera and cleaning when the clarity is lower than a preset value include:

[0012] A waterproof camera is deployed above the aeration tank to collect image data of the aeration area at a fixed frequency;

[0013] Calculating clarity indicators of the aeration area image data at every preset time period, the clarity indicators including edge sharpness, contrast, and blur; the preset time period is dynamically adjusted according to the corrosiveness level of the aeration tank environment;

[0014] When the detected clarity index falls below the preset threshold, the camera self-cleaning mechanism is triggered;

[0015] If the clarity index still does not meet the requirements after self-cleaning, a maintenance alarm will be issued and the abnormal status of the camera will be recorded.

[0016] The specific steps of dynamically adjusting the preset time period according to the corrosiveness level of the aeration tank environment include:

[0017] Establishing an aeration tank corrosivity rating assessment model based on historical sewage data, wherein the historical sewage data includes sewage pH value, hydrogen sulfide concentration, chloride ion content, and ambient humidity data;

[0018] Calculate the corrosivity index based on the currently collected sewage data and the aeration tank corrosivity rating assessment model;

[0019] A corresponding preset time period is matched in a mapping table based on the corrosivity index.

[0020] The specific steps of using the adaptive Retinex algorithm to eliminate the interference of water surface reflections on the real-time image of the aeration area and reducing the noise caused by water turbidity by the multi-frame weighted averaging method to obtain the processed image include:

[0021] Perform multi-scale Gaussian filtering decomposition on the real-time image of the aeration area to obtain illumination component and reflection component;

[0022] The bubble edge information in the reflection component is enhanced by a nonlinear mapping function, while the illumination intensity of the overexposed area in the illumination component is compressed.

[0023] The processed illumination component and reflection component are fused to output an enhanced image with reflection interference eliminated.

[0024] Continuously collect N frames of enhanced images and calculate the temporal stability weight of each pixel, using the first weight for high-fluctuation areas and the second weight for stable background areas;

[0025] Pixel-level fusion based on weight distribution, retaining the real bubble features while suppressing the suspended particle noise;

[0026] The optical flow method is used to detect the flow direction of the water body, and multi-frame alignment compensation is performed along the flow trajectory to obtain the processed image.

[0027] The specific steps of segmenting the bubble area in the image based on the U-Net network and calculating the bubble distribution parameters include:

[0028] Perform semantic segmentation on the preprocessed image through the U-Net network and output the binary mask of the bubble area;

[0029] Mark the connected domains in the binary mask and extract the contour coordinates and area attributes of individual bubbles;

[0030] The minimum circumscribed ellipse is fitted based on the contour coordinates of a single bubble, and its major axis and minor axis are calculated as the bubble diameter characteristics;

[0031] Count the number of bubbles per unit area and generate bubble distribution density data;

[0032] Track the displacement of matching bubbles in continuous frame images, calculate the pixel displacement using the Euler optical flow algorithm, and eliminate abnormal trajectories caused by water flow disturbances;

[0033] The pixel displacement is converted into rising speed by combining the camera calibration parameters.

[0034] The specific steps of tracking the displacement of matching bubbles in continuous frame images, calculating pixel displacement using the Euler optical flow algorithm, and eliminating abnormal trajectories caused by water flow disturbances include:

[0035] Create a spatiotemporal cube data block for consecutive multi-frame images, with a block size of 32×32 pixels;

[0036] The bubble displacement field is calculated using the Euler optical flow algorithm based on the space-time cube data block;

[0037] Calculate the median and standard deviation of the displacement field of the current frame to obtain the judgment value;

[0038] Mark the image frame displacement field that is not within the judgment value range as abnormal;

[0039] The bubble trajectories marked as abnormal for three consecutive frames are removed.

[0040] The specific steps of comparing the bubble distribution parameter with the bubble distribution threshold and generating an aeration volume optimization instruction according to the average dissolved oxygen data obtained from the dissolved oxygen sensor include:

[0041] Establish a standard database of bubble distribution parameters and store reference DO concentration benchmark values ​​under different sewage treatment conditions;

[0042] If the difference between the current DO concentration and the reference DO concentration is lower than the first threshold, an instruction to increase the aeration power is generated; if the current DO concentration and the reference DO concentration are higher than the second threshold, an instruction to reduce the aeration power is generated;

[0043] The command is quantified into the blower speed adjustment value through the PID controller.

[0044] The specific steps of establishing a bubble distribution energy consumption correlation model and dynamically optimizing the bubble distribution threshold using the blower current value as feedback include:

[0045] Collect the three-phase current value of the blower in real time and calculate the effective value of the three-phase current value;

[0046] Establish a multivariate regression model between bubble distribution density, average diameter, rising speed and effective value of three-phase current;

[0047] When the actual current is in the first range, the first threshold is increased and the quality mode is triggered to improve the aeration quality;

[0048] When the actual current is within the second range, the second threshold is reduced and the energy-saving mode is triggered, with priority given to maintaining the lower limit of the DO concentration.

[0049] In a second aspect, the present invention also provides an image data processing system for digital management of environmental protection equipment, comprising an image acquisition module, an image processing module, a bubble parameter calculation module, an aeration instruction optimization module, and a control module;

[0050] The image acquisition module is used to collect real-time images of the aeration area through a waterproof camera deployed above the aeration tank, and regularly detect the clarity of the camera, and clean it when the clarity is lower than a preset value;

[0051] The image processing module is used to use an adaptive Retinex algorithm to eliminate water surface reflection interference from the real-time image of the aeration area, and to reduce noise caused by water turbidity through a multi-frame weighted averaging method to obtain a processed image;

[0052] The bubble parameter calculation module is used to segment the bubble area in the image based on the U-Net network and calculate the bubble distribution parameters, which include bubble distribution density, average diameter and rising speed;

[0053] The aeration instruction optimization module is used to compare the bubble distribution parameter with the bubble distribution threshold and generate an aeration amount optimization instruction based on the average dissolved oxygen data obtained from the dissolved oxygen sensor;

[0054] The control module is used to control the blower based on the aeration amount optimization instruction.

[0055] The present invention provides an image data processing method and system for digital management of environmental protection equipment. The method uses a waterproof camera deployed above an aeration tank to collect real-time images of the aeration area, regularly checks the camera's clarity, and cleans the image when the clarity falls below a preset value. An adaptive Retinex algorithm is used to eliminate interference from water surface reflections in the real-time images of the aeration area, and a multi-frame weighted averaging method is used to reduce noise caused by water turbidity to obtain a processed image. Bubble regions in the image are segmented and processed based on a U-Net network, and bubble distribution parameters are calculated. The bubble distribution parameters are compared with bubble distribution thresholds, and an aeration volume optimization instruction is generated based on average dissolved oxygen data obtained from a dissolved oxygen sensor. A blower is controlled based on the aeration volume optimization instruction.

[0056] This method achieves visual perception and intelligent control of the aeration tank's operating status through a collaborative, full-process approach encompassing image acquisition, image enhancement, bubble identification, parameter analysis, and control execution. Compared to traditional control methods that rely solely on DO feedback, this method provides a more comprehensive picture of aeration effectiveness, enabling more scientific adjustment of blower power, ultimately achieving energy savings and improved system efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 The present invention is a flowchart of an image data processing method for digital management of environmental protection equipment.

[0059] Figure 2 The present invention is a flowchart of collecting real-time images of the aeration area by deploying a waterproof camera above the aeration tank, regularly testing the clarity of the camera, and cleaning when the clarity is lower than a preset value.

[0060] Figure 3 This is a flow chart of the present invention in which the preset time period is dynamically adjusted according to the corrosivity level of the aeration tank environment.

[0061] Figure 4The present invention uses an adaptive Retinex algorithm to eliminate water surface reflection interference from the real-time image of the aeration area, and reduces noise caused by water turbidity through a multi-frame weighted averaging method to obtain a flow chart of the processed image.

[0062] Figure 5 This is a flow chart of the present invention for segmenting and processing bubble areas in an image based on a U-Net network and calculating bubble distribution parameters.

[0063] Figure 6 This is a flowchart of the present invention for tracking the displacement of matching bubbles in continuous frame images, calculating pixel displacement using the Euler optical flow algorithm, and eliminating abnormal trajectories caused by water flow disturbances.

[0064] Figure 7 The present invention is a flow chart of comparing the bubble distribution parameter with the bubble distribution threshold and generating an aeration amount optimization instruction according to the average dissolved oxygen data obtained from the dissolved oxygen sensor.

[0065] Figure 8 This is a flow chart of the present invention for establishing a bubble distribution energy consumption correlation model and dynamically optimizing the bubble distribution threshold using the blower current value as feedback. DETAILED DESCRIPTION

[0066] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0067] First embodiment

[0068] See also Figures 1 to 8 The present invention provides an image data processing method for digital management of environmental protection equipment, comprising:

[0069] S101 collects real-time images of the aeration area through a waterproof camera deployed above the aeration tank, and regularly checks the camera's clarity. When the clarity falls below a preset value, it cleans the image.

[0070] The specific steps include:

[0071] S201 collects image data of the aeration area at a fixed frequency through a waterproof camera deployed above the aeration tank;

[0072] A waterproof camera deployed above the aeration tank collects image data of the aeration area at a fixed frequency. This frequency can be set to suit actual monitoring needs, such as every 5 or 10 minutes, to obtain a continuous, stable image sequence for subsequent image quality assessment and process status analysis.

[0073] S202 calculates clarity indicators of the aeration area image data at every preset time period, wherein the clarity indicators include edge sharpness, contrast, and blur; the preset time period is dynamically adjusted according to the corrosiveness level of the aeration tank environment;

[0074] The clarity index of the aeration area image data is calculated at preset time intervals. This clarity index includes image characteristic parameters such as edge sharpness, contrast, and blurriness. Edge sharpness can be measured using image gradient operators (such as Sobel and Canny); contrast can be quantified using histogram distribution statistics; and blurriness can be estimated using frequency domain transforms (such as Fourier transforms) or local variance methods. This comprehensive evaluation of these multiple dimensions enables a more comprehensive and accurate assessment of image clarity and usability.

[0075] The specific steps of dynamically adjusting the preset time period according to the corrosiveness level of the aeration tank environment include:

[0076] S301 establishes an aeration tank corrosivity rating assessment model based on historical sewage data, wherein the historical sewage data includes sewage pH value, hydrogen sulfide concentration, chloride ion content, and ambient humidity data;

[0077] A model for assessing the corrosivity of aeration tanks was developed based on historical wastewater data. This data includes, but is not limited to, key parameters such as wastewater pH, hydrogen sulfide concentration, chloride ion content, and ambient humidity. These parameters are closely related to equipment corrosion rates and can be used to construct multi-factor regression models or machine learning prediction models to quantify the severity of environmental corrosion.

[0078] S302 calculates a corrosivity index based on currently collected sewage data and an aeration tank corrosivity grade assessment model;

[0079] Based on the currently collected wastewater data and the aeration tank corrosion rating assessment model, the corrosion index for the current operating conditions is calculated. This index is a normalized value that reflects the rate of contamination or corrosion of the camera housing in the current environment. For example, under conditions of high hydrogen sulfide, low pH, and high humidity, the corrosion index will increase significantly, indicating that the camera surface is more susceptible to dirt accumulation, affecting image quality.

[0080] S303 matches a corresponding preset time period in a mapping table based on the corrosivity index.

[0081] Based on the calculated corrosion index, a preset time period is matched to the corresponding time period in a pre-set mapping table. For example, when the corrosion index is low, the clarity test is set to be performed every 7 days; when the index is high, the test is shortened to once a day or even less frequently to ensure that image quality is always within a controllable range.

[0082] S203 triggers a camera self-cleaning mechanism when the detected clarity index is lower than a preset threshold;

[0083] When the detected clarity index falls below a preset threshold, the camera's self-cleaning mechanism is triggered. This self-cleaning mechanism can be a combination of physical cleaning methods such as a mechanical scraper, a spray cleaning system, or an ultrasonic defogging module. The system initiates the appropriate cleaning action to attempt to restore the camera's imaging performance.

[0084] If the clarity index still does not meet the requirement after self-cleaning, a maintenance alarm is issued and the abnormal state of the camera is recorded in S204.

[0085] If the clarity indicator still does not meet the requirements after self-cleaning, a maintenance alert is issued and the abnormal status of the camera is recorded. This indicates that the camera may be severely contaminated or has a hardware failure, requiring manual inspection or replacement. The system automatically records the time, location, relevant parameters, and handling results of the abnormality, facilitating subsequent data analysis and equipment maintenance decisions by maintenance personnel.

[0086] This method realizes the intelligent management and maintenance of aeration tank monitoring cameras by combining image quality assessment, environmental corrosivity analysis and automatic cleaning mechanism, effectively improving the automation level of the sewage treatment process and the stability of the monitoring system.

[0087] S102 uses an adaptive Retinex algorithm to eliminate water surface reflection interference from the real-time image of the aeration area, and reduces noise caused by water turbidity through a multi-frame weighted averaging method to obtain a processed image;

[0088] The specific steps include:

[0089] S401 performs multi-scale Gaussian filtering decomposition on the real-time image of the aeration area to obtain illumination components and reflection components;

[0090] Multi-scale Gaussian filtering is performed on the real-time image of the aeration area to obtain illumination and reflection components. This step, based on the fundamental principles of Retinex theory, decomposes the original image into an illumination component reflecting the scene's lighting conditions and a reflection component reflecting the true color and texture of the object's surface. By smoothing the image using multiple Gaussian kernels of varying scales and combining them with logarithmic domain operations, effective separation of illumination non-uniformity is achieved.

[0091] S402 enhances bubble edge information in the reflection component by a nonlinear mapping function, and compresses the illumination intensity of the overexposed area in the illumination component;

[0092] A nonlinear mapping function is used to enhance bubble edge information in the reflection component, while compressing the intensity of overexposed areas in the illumination component. For the reflection component, a gradient-enhanced nonlinear mapping function (such as gamma correction or contrast-limited adaptive histogram equalization (CLAHE)) is used to highlight key features such as bubble boundaries. For the illumination component, a dynamic range compression strategy is applied to suppress the influence of high-brightness areas (such as reflective spots on the water surface) to prevent them from obscuring surrounding valid image details.

[0093] S403 fuses the processed illumination component and reflection component, and outputs an enhanced image with reflection interference eliminated;

[0094] The processed illumination and reflection components are fused to produce an enhanced image that eliminates reflection interference. The optimized illumination component and the enhanced reflection component are recombined in the logarithmic or linear domain to produce the final enhanced image. This image not only effectively reduces the visual interference caused by surface reflections but also significantly enhances the visibility of bubble edges, improving overall image contrast and detail clarity.

[0095] S404 continuously collects N frames of enhanced images and calculates the temporal stability weight of each pixel, using a first weight for high-fluctuation areas and a second weight for stable background areas;

[0096] N frames of enhanced images are continuously acquired and the temporal stability weight of each pixel is calculated. A first weight is applied to areas of high fluctuation, while a second weight is applied to areas with a stable background. To further reduce random noise caused by suspended particles in the water, the system continuously acquires N frames (e.g., 5 to 10 frames) of enhanced images and performs a time series analysis on each pixel in each frame. Different weight coefficients are assigned based on the magnitude of the change in pixel value over time: a lower first weight is assigned to areas with large fluctuations (e.g., areas with frequent bubble movement); a higher second weight is assigned to areas with a stable background with smaller changes.

[0097] S405 performs pixel-level fusion based on weight distribution, retaining the true bubble features while suppressing suspended particle noise;

[0098] Pixel-level fusion is performed based on a weighted distribution, preserving true bubble features while suppressing suspended particle noise. Using this weight matrix, a weighted average fusion of multiple frames is performed, preserving the true, stable bubble structure in the final output image while effectively reducing artifacts caused by suspended particles or water disturbances, thereby improving the image signal-to-noise ratio and the accuracy of feature extraction.

[0099] S406 detects the flow direction of the water body by using the optical flow method, performs multi-frame alignment compensation along the flow trajectory, and obtains a processed image.

[0100] The direction of water flow is detected by optical flow method, and multiple frames are aligned and compensated along the flow trajectory to obtain the processed image. Considering that the water flow in the aeration tank has certain dynamic characteristics, which may cause displacement offset between multiple frames, this step adopts the method based on dense optical flow estimation (such as Or TV-L1 optical flow algorithm) estimates the motion vector of the image sequence and performs spatial alignment correction on the image accordingly to ensure that multiple frames of image are in a consistent spatial coordinate system before fusion, thereby improving image fusion accuracy and final image quality.

[0101] In summary, the image processing process proposed by S102 combines multiple technical means, such as adaptive Retinex enhancement, multi-frame weighted fusion, and optical flow alignment compensation, to effectively solve problems such as reflection interference, noise pollution, and motion blur in aeration area images, significantly improve image quality, and provide a solid data foundation for subsequent intelligent analysis and process optimization.

[0102] S103 segments the bubble area in the processed image based on the U-Net network and calculates the bubble distribution parameters, wherein the bubble distribution parameters include bubble distribution density, average diameter, and rising speed;

[0103] The specific steps include:

[0104] S501 performs semantic segmentation on the preprocessed image through the U-Net network and outputs a binary mask of the bubble area;

[0105] The preprocessed image is semantically segmented using the U-Net network, and a binary mask of the bubble area is output. U-Net is a classic convolutional neural network with an encoder-decoder structure, which is particularly suitable for pixel-level segmentation tasks in medical and industrial images. In this step, a U-Net model that has been trained and optimized with a large number of labeled samples is used to perform end-to-end semantic segmentation on the enhanced image output by S102, accurately identifying and marking all bubble areas in the image, and generating a corresponding binary mask, where the bubble area is white (pixel value is 1) and the background area is black (pixel value is 0).

[0106] S502 marks the connected domains in the binary mask and extracts the contour coordinates and area attributes of individual bubbles;

[0107] Connected domains within the binary mask are labeled, and the contour coordinates and area attributes of individual bubbles are extracted. Image morphological operations and connected domain analysis algorithms (such as the findContours or label functions in OpenCV) are used to identify and label each individual bubble object within the binary mask image. Each bubble is assigned a unique label ID, and its boundary contour coordinates and pixel area are extracted for subsequent size calculation and statistical analysis.

[0108] S503 fits a minimum circumscribed ellipse based on the contour coordinates of a single bubble, and calculates its major axis and minor axis as the bubble diameter feature;

[0109] A minimum circumscribed ellipse is fitted to the coordinates of a single bubble's outline, and its major and minor axes are calculated as the bubble's diameter. Because the projection of an actual bubble in water can appear irregular or elliptical, the minimum enclosing ellipse method is used to fit the outline, providing a more accurate estimate of the bubble's actual size. Based on the fitting results, the average of the major and minor axes is taken as the bubble's equivalent diameter. This is then used to calculate parameters such as the average diameter, maximum diameter, and diameter distribution of the bubbles within the entire image.

[0110] S504 counts the number of bubbles in a unit area to generate bubble distribution density data;

[0111] Counting bubbles per unit area generates bubble distribution density data. By counting all detected bubbles in the image and combining the image's actual physical dimensions (calculated from the camera's field of view and installation height), the number of bubbles per unit area (e.g., per square centimeter or square meter) is calculated to form a bubble distribution density metric. This metric can be used to reflect the aeration head's operating status and oxygen transfer efficiency.

[0112] S505 tracks the displacement of matching bubbles in the continuous frame images, calculates the pixel displacement using the Euler optical flow algorithm, and eliminates abnormal trajectories caused by water flow disturbances;

[0113] The specific steps include:

[0114] S601 creates a spatiotemporal cube data block for multiple consecutive frames of images, with a block size of 32×32 pixels;

[0115] A spatiotemporal cube data block of 32×32 pixels is created for consecutive frames. This step introduces the time dimension into image processing, stacking images of the same region from several consecutive frames (e.g., five frames) into a three-dimensional data structure, forming a spatiotemporal cube. This structure effectively captures the dynamic characteristics of pixels within the region over time, providing a good foundation for subsequent optical flow estimation.

[0116] S602 calculates the bubble displacement field using the Euler optical flow algorithm based on the space-time cube data block;

[0117] The Euler optical flow algorithm is used to calculate the bubble displacement field based on the space-time cube data block. This method is a motion estimation method based on the assumption of constant local pixel brightness and is suitable for describing motion trends in local areas of an image. By performing optical flow estimation on the space-time cube, the two-dimensional displacement vector of each pixel in the time series is obtained. This displacement field is then constructed for the entire image area, allowing the direction and magnitude of bubble movement to be identified.

[0118] S603 calculates the median and standard deviation of the displacement field of the current frame to obtain a judgment value;

[0119] The median and standard deviation of the displacement field of the current frame are calculated to obtain the judgment value. To distinguish normal bubble motion from abnormal disturbance trajectories, the system performs statistical analysis on the displacement field of the entire frame image, calculates the median and standard deviation of its displacement vector length, and sets a reasonable threshold range. For example, the median ± 2 times the standard deviation is used as the judgment boundary to identify abnormal displacements that exceed normal motion patterns.

[0120] S604: If the frame displacement field of the image is not within the judgment value range, mark it as abnormal;

[0121] Pixels in the image whose displacement field is outside the judgment value range are marked as abnormal. For each detected bubble object, if its displacement in the current frame exceeds the threshold range set above, it is considered that the bubble may be affected by unstable water flow or other interference factors, and its trajectory is marked as "abnormal".

[0122] In step S605 , bubble trajectories marked as abnormal for three consecutive frames are removed.

[0123] Bubble trajectories marked as abnormal for three consecutive frames are removed. To avoid misjudgments, the system uses a continuity judgment mechanism: only when a bubble is marked as an abnormal trajectory in three consecutive frames is it removed from the final bubble motion dataset. This ensures that only false trajectories caused by disturbances are filtered out, while retaining real and valid bubble motion information.

[0124] S506 converts the pixel displacement into a rising speed in combination with the camera calibration parameters.

[0125] After cleaning and optimizing the bubble displacement trajectory, the system uses pre-calibrated camera internal and external parameters (including focal length, pixel physical size, installation height, etc.) to convert pixel displacement into actual physical units (such as millimeters). Combined with the image acquisition time interval (such as 30 frames per second), the system further calculates the bubble's rising velocity (unit: mm / s or cm / s), thereby enabling quantitative analysis of the bubble dynamics during the aeration process.

[0126] S104 compares the bubble distribution parameter with the bubble distribution threshold and generates an aeration optimization instruction based on the average dissolved oxygen data obtained from the dissolved oxygen sensor;

[0127] The specific steps include:

[0128] S701 establishes a standard database of bubble distribution parameters and stores reference DO concentration benchmark values ​​under different sewage treatment conditions;

[0129] A standard database of bubble distribution parameters is established, storing reference DO concentration baseline values ​​for different sewage treatment conditions. This database, built based on historical operating data and expert experience, contains ideal bubble distribution characteristics (such as bubble density, average diameter, and rise velocity) and their corresponding optimal dissolved oxygen concentration ranges under various typical operating conditions (such as influent COD concentration, water temperature, and sludge concentration (MLSS)). By comparing the current operating status with matching conditions in the database, the system can obtain a desired DO concentration target value, which serves as a reference for subsequent control.

[0130] S702: If the difference between the current DO concentration and the reference DO concentration is lower than the first threshold, an instruction to increase the aeration power is generated; if the difference between the current DO concentration and the reference DO concentration is higher than the second threshold, an instruction to decrease the aeration power is generated;

[0131] If the difference between the current DO concentration and the reference DO concentration is below a set first threshold, it indicates insufficient aeration intensity and the oxygen supply cannot meet process requirements. The system then generates a control instruction to increase aeration power. Conversely, if the current DO concentration is above the reference DO concentration and exceeds a set second threshold, it indicates excessive aeration and the risk of energy waste. The system then generates a control instruction to reduce aeration power. The first and second thresholds can be set to, for example, ±0.5 mg / L based on process accuracy requirements, or can be dynamically adjusted based on actual conditions to achieve more precise process control.

[0132] S703 quantifies the instruction into a blower speed adjustment value through a PID controller.

[0133] The PID controller quantifies these control commands into blower speed adjustment values. The PID (Proportional-Integral-Derivative) control algorithm is a widely used closed-loop feedback control method. The system inputs the DO concentration deviation signal into a pre-set PID controller. After comprehensive calculation of the proportional, integral, and differential terms, it outputs a precise blower speed adjustment (ΔRPM), which drives the inverter to adjust the blower's operating frequency, thereby achieving continuous and smooth regulation of the aeration volume. The PID parameters (Kp, Ki, Kd) can be automatically adjusted online based on the system's response characteristics to meet control requirements under different operating conditions.

[0134] S105 controls the blower based on the aeration volume optimization instruction.

[0135] After completing image recognition and fusion analysis of dissolved oxygen data, the system generates aeration optimization instructions (such as increasing or decreasing aeration power) that are transmitted to the blower control system. This instruction adjusts the inverter output frequency, thereby changing the blower speed, achieving precise control of air supply. The control process uses a closed-loop feedback mechanism to ensure that the actual aeration volume quickly responds to changes in the target value, maintaining the dissolved oxygen concentration in the aeration tank within the set range while avoiding equipment loss and energy waste caused by frequent starts and stops or large fluctuations.

[0136] S106 establishes a bubble distribution energy consumption correlation model and dynamically optimizes the bubble distribution threshold using the blower current value as feedback.

[0137] The specific steps include:

[0138] S801 collects the three-phase current value of the blower in real time and calculates the effective value of the three-phase current value;

[0139] The system uses current transformers or smart meters installed in the power distribution cabinet to collect the three-phase current signals (A, B, and C) of the blower in real time and calculate their effective current values ​​(RMS values). This RMS value accurately reflects the blower's real-time load status and is a key indicator for evaluating energy consumption levels.

[0140] S802 establishes a multivariate regression model of bubble distribution density, average diameter and rising speed and effective value of three-phase current;

[0141] Based on historical operating data, the system constructs a multivariate regression model, using key bubble distribution parameters (such as bubble density, average diameter, and rise velocity) as input variables and the effective value of the blower current as the output variable, to fit the mathematical relationship between the two. This model can be used to predict energy consumption levels under different bubble distribution conditions and provide a theoretical basis for subsequent threshold adjustments.

[0142] S803 increases the first threshold value and triggers the quality mode when the actual current is within the first range, thereby improving the aeration quality;

[0143] When the blower current is detected to be in the low range (e.g., 40% to 60% of rated current), it indicates that the current aeration intensity is low, and there may be a risk of insufficient oxygen mass transfer efficiency. In this case, the system determines that it is in "quality priority" mode and automatically increases the first threshold of bubble distribution density or bubble diameter, prompting the control system to actively increase the aeration volume, improve aeration quality, and ensure the stability of the biochemical reaction process.

[0144] S804: When the actual current is in the second range, the second threshold is reduced and the energy-saving mode is triggered, with priority given to maintaining the lower limit of the DO concentration.

[0145] If the blower current has entered a high range (for example, above 70% of the rated current), it indicates that current energy consumption is too high and there is potential for energy conservation. The system switches to "energy conservation priority" mode, appropriately lowering the judgment threshold of bubble distribution parameters, allowing a certain degree of bubble sparseness or size variation, thereby reducing unnecessary aeration volume. Under the premise of ensuring that the dissolved oxygen does not fall below the process lower limit, the purpose of energy conservation and consumption reduction is achieved.

[0146] Second embodiment

[0147] The present invention also provides an image data processing system for digital management of environmental protection equipment, comprising an image acquisition module, an image processing module, a bubble parameter calculation module, an aeration instruction optimization module, and a control module; the image acquisition module is used to collect real-time images of the aeration area through a waterproof camera deployed above the aeration tank, and regularly detect the clarity of the camera, and clean it when the clarity is lower than a preset value;

[0148] The image processing module is used to use an adaptive Retinex algorithm to eliminate water surface reflection interference from the real-time image of the aeration area, and to reduce noise caused by water turbidity through a multi-frame weighted averaging method to obtain a processed image; the bubble parameter calculation module is used to segment the bubble area in the processing image based on a U-Net network and calculate bubble distribution parameters, wherein the bubble distribution parameters include bubble distribution density, average diameter and rising speed; the aeration instruction optimization module is used to compare the bubble distribution parameters with the bubble distribution threshold and generate an aeration volume optimization instruction based on the average dissolved oxygen data obtained from the dissolved oxygen sensor; and the control module is used to control the blower based on the aeration volume optimization instruction.

[0149] In this embodiment, the image acquisition module is used to collect image information of the aeration area in real time through a waterproof camera deployed above the aeration tank. The module has the functions of timed acquisition, clarity detection and automatic maintenance. Specifically, the camera collects images at a fixed frequency (such as every 5 minutes) and regularly evaluates the image quality. When the system detects that the image clarity is lower than the preset threshold (for example, the edge blur increases and the contrast decreases), the camera surface self-cleaning mechanism will be triggered; if the clarity cannot be restored after cleaning, a maintenance alarm will be generated to prompt manual intervention. In addition, the module also supports dynamic adjustment of the clarity detection cycle according to the environmental corrosivity level, thereby improving the adaptability and stability of the system.

[0150] The image processing module enhances and denoises the collected raw images to eliminate reflections from the water surface and suppress the noise caused by turbid water. This module uses an adaptive Retinex algorithm to perform multi-scale Gaussian filtering on the image, separating the illumination and reflection components. It then applies a nonlinear mapping function to enhance bubble edge features and compress the brightness of overexposed areas.

[0151] The bubble parameter calculation module is used to segment and quantitatively analyze bubble regions in processed images based on a deep learning model. This module uses a U-Net network structure to perform pixel-level semantic segmentation on the image, extract a binary mask image of the bubble region, and identify each individual bubble object through connected domain analysis. The system further performs elliptical fitting on the contour of each bubble, calculating its major and minor axes to derive its equivalent diameter. The system counts the number of bubbles per unit area to determine the bubble distribution density. Simultaneously, the system uses the Euler optical flow algorithm to estimate the bubble rise velocity based on the bubble displacement trajectories in consecutive frame images, eliminating anomalous trajectories caused by water flow disturbances. The module ultimately outputs complete bubble distribution parameters, including bubble density, average diameter, and rise velocity, providing key data support for aeration efficiency assessment.

[0152] The aeration command optimization module compares the bubble distribution parameters acquired through image recognition with the set bubble distribution threshold. Combined with the dissolved oxygen (DO) concentration data collected by the dissolved oxygen sensor, it comprehensively determines whether the current aeration effect meets process requirements and generates corresponding aeration rate optimization commands. The system includes a standard database of bubble distribution parameters, containing reference DO concentration baseline values ​​for different operating conditions. When the actual DO concentration deviates from the reference value by more than the set threshold, the system generates a control command to increase or decrease the aeration power based on the deviation direction. This command is then quantified into a specific blower speed adjustment value through a PID controller, enabling fine-grained regulation of the air supply.

[0153] The control module is used to receive control signals from the aeration instruction optimization module and perform closed-loop control of the blower accordingly, ensuring that the aeration intensity responds to process requirements in real time. This module supports variable frequency control and can adjust the operating frequency and speed of the blower according to instructions, thereby accurately controlling the air supply. In addition, the system also introduces an energy consumption feedback mechanism to establish a multivariate regression model between bubble distribution characteristics and blower current. The bubble distribution threshold is dynamically optimized according to current changes, achieving energy conservation and consumption reduction while ensuring treatment results. For example, when the current is in the low load range, the system enters "quality priority" mode to improve aeration quality; when the current is high, it switches to "energy saving priority" mode, prioritizing the maintenance of the lower limit of dissolved oxygen concentration, taking into account both process stability and energy efficiency.

[0154] In summary, the image data processing system for digital management of environmental protection equipment provided by this invention integrates image acquisition, processing, analysis, and control, creating a closed-loop management system from visual perception to action feedback. This system not only enables intelligent monitoring and dynamic adjustment of the aeration process, but also significantly improves the operational efficiency and energy utilization of sewage treatment systems through multi-source data fusion and energy consumption optimization modeling.

[0155] The above disclosure is only a preferred embodiment of the present invention, and certainly cannot be used to limit the scope of the rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. An image data processing method for digital management of environmental protection equipment, It is characterized in that This includes: using a waterproof camera deployed above the aeration tank to collect real-time images of the aeration area, regularly testing the camera's clarity, and cleaning it when the clarity falls below a preset value; The real-time image of the aeration area is subjected to an adaptive Retinex algorithm to eliminate the interference of water surface reflections, and a multi-frame weighted averaging method is used to reduce the noise caused by water turbidity to obtain a processed image; Based on the U-Net network, the bubble area in the image is segmented and the bubble distribution parameters are calculated, including the bubble distribution density, average diameter and rising speed; Comparing the bubble distribution parameter with the bubble distribution threshold and generating an aeration optimization instruction based on average dissolved oxygen data obtained from a dissolved oxygen sensor; The blower is controlled based on the aeration rate optimization instruction.

2. The image data processing method for digital management of environmental protection equipment according to claim 1, characterized in that: After the blower is controlled based on the aeration volume optimization instruction, the method further includes establishing a bubble distribution energy consumption correlation model and dynamically optimizing the bubble distribution threshold using the blower current value as feedback.

3. The image data processing method for digital management of environmental protection equipment according to claim 2, characterized in that: The specific steps of collecting real-time images of the aeration area by deploying a waterproof camera above the aeration tank and regularly testing the clarity of the camera and cleaning when the clarity is lower than a preset value include: A waterproof camera is deployed above the aeration tank to collect image data of the aeration area at a fixed frequency; Calculating clarity indicators of the aeration area image data at every preset time period, the clarity indicators including edge sharpness, contrast, and blur; the preset time period is dynamically adjusted according to the corrosiveness level of the aeration tank environment; When the detected clarity index falls below the preset threshold, the camera self-cleaning mechanism is triggered; If the clarity index still does not meet the requirements after self-cleaning, a maintenance alarm will be issued and the abnormal status of the camera will be recorded.

4. The image data processing method for digital management of environmental protection equipment according to claim 3, characterized in that: The specific steps of dynamically adjusting the preset time period according to the corrosiveness level of the aeration tank environment include: Establishing an aeration tank corrosivity rating assessment model based on historical sewage data, wherein the historical sewage data includes sewage pH value, hydrogen sulfide concentration, chloride ion content, and ambient humidity data; Calculate the corrosivity index based on the currently collected sewage data and the aeration tank corrosivity rating assessment model; A corresponding preset time period is matched in a mapping table based on the corrosivity index.

5. The image data processing method for digital management of environmental protection equipment according to claim 4, characterized in that: The specific steps of using the adaptive Retinex algorithm to eliminate the interference of water surface reflections on the real-time image of the aeration area and reducing the noise caused by water turbidity by the multi-frame weighted averaging method to obtain the processed image include: Perform multi-scale Gaussian filtering decomposition on the real-time image of the aeration area to obtain illumination component and reflection component; The bubble edge information in the reflection component is enhanced by a nonlinear mapping function, while the illumination intensity of the overexposed area in the illumination component is compressed. The processed illumination component and reflection component are fused to output an enhanced image with reflection interference eliminated. Continuously collect N frames of enhanced images and calculate the temporal stability weight of each pixel, using the first weight for high-fluctuation areas and the second weight for stable background areas; Pixel-level fusion based on weight distribution, retaining the real bubble features while suppressing the suspended particle noise; The optical flow method is used to detect the flow direction of the water body, and multi-frame alignment compensation is performed along the flow trajectory to obtain the processed image.

6. The image data processing method for digital management of environmental protection equipment according to claim 5, characterized in that: The specific steps of segmenting the bubble area in the image based on the U-Net network and calculating the bubble distribution parameters include: Perform semantic segmentation on the preprocessed image through the U-Net network and output the binary mask of the bubble area; Mark the connected domains in the binary mask and extract the contour coordinates and area attributes of individual bubbles; The minimum circumscribed ellipse is fitted based on the contour coordinates of a single bubble, and its major axis and minor axis are calculated as the bubble diameter characteristics; Count the number of bubbles per unit area and generate bubble distribution density data; Track the displacement of matching bubbles in continuous frame images, calculate the pixel displacement using the Euler optical flow algorithm, and eliminate abnormal trajectories caused by water flow disturbances; The pixel displacement is converted into rising speed by combining the camera calibration parameters.

7. The image data processing method for digital management of environmental protection equipment according to claim 6, characterized in that: The specific steps of tracking the displacement of matching bubbles in continuous frame images, calculating pixel displacement using the Euler optical flow algorithm, and eliminating abnormal trajectories caused by water flow disturbances include: Create a spatiotemporal cube data block for consecutive multi-frame images, with a block size of 32×32 pixels; The bubble displacement field is calculated using the Euler optical flow algorithm based on the space-time cube data block; Calculate the median and standard deviation of the displacement field of the current frame to obtain the judgment value; Mark the image frame displacement field that is not within the judgment value range as abnormal; The bubble trajectories marked as abnormal for three consecutive frames are removed.

8. The image data processing method for digital management of environmental protection equipment according to claim 7, characterized in that: The specific steps of comparing the bubble distribution parameter with the bubble distribution threshold and generating an aeration optimization instruction based on the average dissolved oxygen data obtained from the dissolved oxygen sensor include: Establish a standard database of bubble distribution parameters and store reference DO concentration benchmark values ​​under different sewage treatment conditions; If the difference between the current DO concentration and the reference DO concentration is lower than the first threshold, an instruction to increase the aeration power is generated; if the current DO concentration and the reference DO concentration are higher than the second threshold, an instruction to reduce the aeration power is generated; The command is quantified into the blower speed adjustment value through the PID controller.

9. The image data processing method for digital management of environmental protection equipment according to claim 8, characterized in that: The specific steps of establishing the bubble distribution energy consumption correlation model and dynamically optimizing the bubble distribution threshold using the blower current value as feedback include: Collect the three-phase current value of the blower in real time and calculate the effective value of the three-phase current value; Establish a multivariate regression model between bubble distribution density, average diameter, rising speed and effective value of three-phase current; When the actual current is in the first range, the first threshold is increased and the quality mode is triggered to improve the aeration quality; When the actual current is within the second range, the second threshold is reduced and the energy-saving mode is triggered, with priority given to maintaining the lower limit of the DO concentration.

10. An image data processing system for digital management of environmental protection equipment, using the image data processing method for digital management of environmental protection equipment according to claim 1, characterized in that: The system includes an image acquisition module, an image processing module, a bubble parameter calculation module, an aeration instruction optimization module, and a control module. The image acquisition module is used to collect real-time images of the aeration area through a waterproof camera deployed above the aeration tank, and regularly detect the clarity of the camera. When the clarity is lower than a preset value, the system cleans the image. The image processing module is used to use an adaptive Retinex algorithm to eliminate water surface reflection interference from the real-time image of the aeration area, and to reduce noise caused by water turbidity through a multi-frame weighted averaging method to obtain a processed image; The bubble parameter calculation module is used to segment the bubble area in the image based on the U-Net network and calculate the bubble distribution parameters, which include bubble distribution density, average diameter and rising speed; The aeration instruction optimization module is used to compare the bubble distribution parameter with the bubble distribution threshold and generate an aeration amount optimization instruction based on the average dissolved oxygen data obtained from the dissolved oxygen sensor; The control module is used to control the blower based on the aeration amount optimization instruction.

Images