A method and system for detecting pollution under adverse weather conditions

Abstract

The application relates to the technical field of pollution discharge detection, and discloses a pollution discharge detection method and system under severe weather. The pollution discharge detection method under severe weather comprises the following steps: S1, acquiring video image information through a black light camera at a pollution discharge port; S2, classifying video images by adopting a convolutional neural network based on the collected video image information; S3, performing corresponding sharpening processing on the type of video images; and S4, putting the sharpened video images into a target detection neural network to perform pollution discharge detection. The application provides a pollution discharge detection video image enhancement technology based on video monitoring, which can sharpen the video images at the pollution discharge port according to different weather conditions. Whether the pollution discharge system has a pipeline or a rainwater well that is affected by rain can be judged by analyzing whether the pollution discharge amount is proportional to the rainfall.

Description

Technical Field

[0001] This invention relates to the field of sewage discharge detection technology, specifically to a sewage discharge detection method and system under severe weather conditions. Background Technology

[0002] Wastewater refers to wastewater discharged from domestic and industrial sources that has been polluted to some extent. Category 1: Industrial wastewater originates from manufacturing, mining, and industrial production activities, including runoff and leachate from industrial or commercial storage and processing, as well as other wastewater that is not domestic sewage. Category 2: Domestic sewage originates from residences, office buildings, government offices, or similar establishments; sanitary sewage; and sewer sewage, including industrial wastewater mixed with domestic sewage in the sewer system. Video surveillance of sewage outlets can provide a preliminary assessment of water quality and detect the presence of floating debris by distinguishing the color of the wastewater. However, existing video surveillance methods have drawbacks, especially in rainy and foggy weather. The images obtained by the cameras are not clear enough due to rain or fog, directly affecting the detection results.

[0003] In addition, during rainy weather, cities generally adopt a separate drainage system for rainwater and sewage, separating rainwater and sewage and using separate pipelines for each for discharge or subsequent treatment. Rainwater is discharged directly into rivers through rainwater pipe networks, while sewage is collected through sewage pipe networks and sent to sewage treatment plants for treatment, preventing sewage from directly entering rivers and causing pollution. Furthermore, the collection, utilization, and centralized management of rainwater discharge can reduce the impact of water volume on sewage treatment plants, ensuring their treatment efficiency. Clearly, connecting sewage to rainwater pipes poses a serious pollution problem. To avoid obstructing sewage pipes, some people still privately connect rainwater to them. This not only increases the workload of sewage treatment plants but also poses safety hazards. Heavy rainfall may cause pressure on sewage pipes, leading to overflows, damage, and other issues. From an environmental perspective, it also results in a waste of water resources. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for detecting sewage discharge under severe weather conditions, and to solve the following technical problems:

[0005] (1) How to improve the clarity of sewage monitoring videos for different weather conditions;

[0006] (2) How to determine the rainwater inflow situation in the sewage system through sewage discharge testing in response to rainy weather.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A method for detecting sewage discharge under severe weather conditions, the method comprising the following steps:

[0009] S1. Obtain video image information at the sewage outlet using a black light camera;

[0010] S2. Based on the acquired video image information, a convolutional neural network is used to classify the video images;

[0011] S3. Based on the type of video image, perform corresponding sharpening processing on that type of video image;

[0012] S4. The clarified video image is fed into the target detection neural network for sewage detection.

[0013] The above technical solution provides a method for enhancing video images of sewage discharge detection based on video surveillance, which can clarify video images at sewage outlets according to different weather conditions.

[0014] In one embodiment, the video image types include rainy days and foggy days, and the sharpening process for rainy days in step S3 includes:

[0015] A rough binary image is obtained from the original video image, and the binary image is displayed as a rain line layer and a background layer.

[0016] Principal component analysis is used to process the connected components of the binary image, removing high-frequency details and obtaining a refined binary image.

[0017] The thinned binary image is used as a mask, and the mask is removed from the original video image to obtain a rainless video image.

[0018] The above technical solution provides a specific method for sharpening video images in rainy weather, which can obtain rain-free video images.

[0019] In one embodiment, the clarity processing for foggy weather in step S3 includes:

[0020] Dehazing is performed based on a dark channel prior dehazing algorithm applied to the original video image, specifically as follows:

[0021] The input image undergoes adaptive filtering; then the dark channel value of each pixel is calculated.

[0022] Based on the characteristics of the dark channel, the location of the largest dark channel value in the image is determined, and the pixel value at that location is used as a reference value to determine the intensity of atmospheric light.

[0023] The intensity of fog in the image is determined by using the least squares algorithm and the intensity of atmospheric light and the dark channel curve.

[0024] Based on the atmospheric light and fog intensity obtained above, the original video image can be calculated, thereby achieving image dehazing.

[0025] Through the above technical solution, this embodiment provides a specific method for sharpening video images in foggy weather, which can obtain clear video images after defogging.

[0026] In one embodiment, the sewage discharge detection method under severe weather conditions further includes:

[0027] S5. Analyze the original video images of the rainy day to obtain the function of rainfall changing over time. Simultaneously, acquire the state parameters of the flowing sewage in the sewage pipe; analyze the state parameters of the flowing sewage within a preset time period and obtain the discharge flow function, then combine the flow function with the function... Conduct comparative analysis and establish a sewage system detection strategy based on the analysis results.

[0028] The above technical solution provides a method for formulating a sewage system detection strategy, which can determine whether there is rain entering the sewage system through pipelines or sewage wells by analyzing whether the sewage discharge volume and rainfall volume are directly proportional.

[0029] In one embodiment, the method for obtaining the state parameters of the flowing sewage in the sewage pipe is specifically as follows: obtaining the real-time flow rate of the sewage in the sewage pipe based on a flow meter. The real-time deepest water level of sewage in the pipeline is obtained based on the water level detector. .

[0030] In one embodiment, the process of analyzing the state parameters of flowing sewage within a preset time period is as follows:

[0031] Through formula Obtain the function of the cross-sectional area of ​​the flowing sewage in the sewage pipe as a function of time. , where R is the inner radius of the sewage pipe;

[0032] Through formula Obtain wastewater flow function .

[0033] The above technical solution provides a method for measuring the flow rate of sewage in a sewage pipe.

[0034] In one embodiment, the flow function AND function The process of analysis and comparison is as follows:

[0035] Take multiple intervals with lengths of interval ,in This represents the i-th time point taken in chronological order, and , , All are preset constants;

[0036] According to the formula and Calculate the total rainfall per unit time. Total amount of sewage discharged per unit time .

[0037] According to the formula Obtain the ratio of total pollution discharge to total rainfall per unit time. ,in , Using preset coefficients, n sets of proportional coefficients are measured, where n > 10. The status of the sewage system is determined based on M sets of proportional coefficients.

[0038] In one embodiment, the process of establishing a sewage system detection strategy is as follows:

[0039] Based on the status of the sewage system, according to the formula Calculate the variance of the proportionality coefficient for each group. ,in For n groups of proportionality coefficients The average;

[0040] like Then the sewage system needs to be inspected and repaired;

[0041] like If so, there is no need to inspect or repair the sewage system; among which As a preset standard value, and .

[0042] Through the above technical solution, this embodiment provides a function for wastewater discharge and a function for wastewater discharge. The specific process of analysis and comparison, and the specific process of establishing a sewage system detection strategy.

[0043] A sewage discharge detection system under severe weather conditions, the system comprising:

[0044] The information acquisition module is used to collect video image information of sewage at the sewage outlet;

[0045] Convolutional neural networks are used to classify acquired video image information;

[0046] The image processing module is used to perform sharpening processing on video images according to the type of video image information;

[0047] The rainfall detection module is used to obtain rainfall information based on video images of rainy weather types.

[0048] The wastewater detection module is used to detect and acquire the status parameters of the flowing wastewater in the wastewater pipe;

[0049] The analysis module is used to comprehensively analyze the obtained rainfall information and the state parameters of flowing sewage, and to determine the status of the sewage system.

[0050] The beneficial effects of this invention are:

[0051] (1) This invention provides a video image enhancement technology for sewage discharge detection based on video surveillance. Utilizing the night vision capability of a black light camera, it can clearly capture the original video image information at the sewage outlet regardless of day or night. The original video image information is then classified, and different video image processing techniques are applied to clarify it according to different weather conditions. Finally, the clarified video image is fed into a target detection neural network for sewage discharge detection. Therefore, this invention achieves the effect of clarifying video images at the sewage outlet according to different weather conditions.

[0052] (2) The present invention also provides a method for formulating a sewage system detection strategy. The rainfall function is obtained by video capture, and the sewage flow function is obtained by analyzing the state parameters of the flowing sewage in the sewage pipe. The flow function is compared with the rainfall function. By analyzing whether the sewage discharge and rainfall are proportional, it is determined whether there is rain entering the sewage system through the pipe or sewage well. Based on the analysis results, a sewage system detection strategy can be established.

[0053] (3) The present invention also solves the problem that it is difficult to detect the real-time flow in large sewage pipes because flow meters cannot be installed. By measuring the sewage flow velocity and water level, the cross-sectional area function is established and the final flow function is obtained. This method can monitor the real-time flow of sewage in non-full-flow sewage pipes. Attached Figure Description

[0054] The invention will now be further described with reference to the accompanying drawings.

[0055] Figure 1 This is a flowchart illustrating the steps of the wastewater discharge detection method for severe weather conditions according to the present invention.

[0056] Figure 2 This is a schematic block diagram of the sewage discharge detection system for severe weather conditions according to the present invention; Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Please refer to the attached diagram. Figure 1 As shown, this invention provides a method for detecting sewage discharge under severe weather conditions, the method comprising the following steps:

[0059] S1. Obtain video image information at the sewage outlet using a black light camera;

[0060] S2. Based on the acquired video image information, a convolutional neural network is used to classify the video images;

[0061] S3. Based on the type of video image, perform corresponding sharpening processing on that type of video image;

[0062] S4. The clarified video image is fed into the target detection neural network for sewage detection.

[0063] Through the above technical solution, this embodiment provides a video image enhancement technology for sewage discharge detection based on video surveillance. It can enhance the clarity of video images at sewage outlets under different weather conditions. Specifically, utilizing the night vision capability of a black light camera, it can clearly capture the original video image information at the sewage outlet regardless of day or night. Then, a convolutional neural network (ResNet18+Softmax) is used to classify the original video image information. Based on the different weather types identified, different video image processing techniques are applied for clarity enhancement. Finally, the enhanced video image is fed into a target detection neural network for sewage discharge detection. It should be noted that the enhanced video image can observe floating debris and the color of the sewage. The color of the sewage at night can be captured by the black light camera.

[0064] The video images include rainy and foggy days. Step S3, specifically the sharpening process for rainy days, includes:

[0065] A rough binary image is obtained from the original video image, and the binary image is displayed as a rain line layer and a background layer.

[0066] Principal component analysis is used to process the connected components of the binary image, removing high-frequency details and obtaining a refined binary image.

[0067] The thinned binary image is used as a mask, and the mask is removed from the original video image to obtain a rainless video image.

[0068] Through the above technical solution, this embodiment provides a method for sharpening video images of rainy weather. Specifically, since raindrops typically have a strong reflective effect on light, a coarse binary image can be obtained from the original video image. The binary image is displayed as a rain line layer and a background layer. The rain line layer of the binary image includes raindrops and other high-frequency details of the image. Since the raindrops are of similar size, principal component analysis (PCA) can be used to process the connected components of the binary image, removing high-frequency details and obtaining a refined binary image. PCA involves transforming a set of potentially correlated variables into a set of linearly uncorrelated variables through orthogonal transformation. This transformed set of variables is the principal component. PCA is an existing analytical method for processing the connected components of binary images, which will not be detailed here. The original video image includes a rain line map and a background map. The refined binary image can be used as a mask to remove the mask from the original video image, ultimately obtaining a rain-free video image.

[0069] The clarity enhancement process for foggy weather in step S3 includes:

[0070] Dehazing is performed based on a dark channel prior dehazing algorithm applied to the original video image, specifically as follows:

[0071] The input image undergoes adaptive filtering; then the dark channel value of each pixel is calculated.

[0072] Based on the characteristics of the dark channel, the location of the largest dark channel value in the image is determined, and the pixel value at that location is used as a reference value to determine the intensity of atmospheric light.

[0073] The intensity of fog in the image is determined by using the least squares algorithm and the intensity of atmospheric light and the dark channel curve.

[0074] Based on the atmospheric light and fog intensity obtained above, the original video image can be calculated, thereby achieving image dehazing.

[0075] Through the above technical solution, this embodiment provides a method for sharpening video images in foggy weather. Specifically, a dark channel prior dehazing algorithm is used. First, the input image undergoes adaptive filtering to reduce the impact of noise on the dehazing result. Then, the dark channel value of each pixel is calculated. This dark channel value reflects the darkest pixel value in the image; the smaller the value, the stronger the fog effect. Next, based on the characteristics of the dark channel, the position of the largest dark channel value in the image is determined. Using the pixel value at this position as a reference value, the intensity of atmospheric light is determined. Then, using the least squares algorithm, the intensity of fog in the image is calculated using the intensity of atmospheric light and the dark channel curve. Based on the obtained atmospheric light and fog intensities, the original video image is calculated, achieving the dehazing and sharpening process.

[0076] The sewage discharge detection method under severe weather conditions also includes:

[0077] S5. Analyze the original video images of the rainy day to obtain the function of rainfall changing over time. Simultaneously, acquire the state parameters of the flowing sewage in the sewage pipe; analyze the state parameters of the flowing sewage within a preset time period and obtain the discharge flow function, then combine the flow function with the function... Conduct comparative analysis and establish a sewage system detection strategy based on the analysis results.

[0078] The above technical solution provides a method for formulating a sewage system detection strategy. Specifically, it first requires obtaining a function of rainfall changing over time. It should be noted that this function was obtained through the video capture method described above. The rainfall represented is generally limited to localized rainfall within the depth of field and shot area of ​​the video recording. Then, the sewage flow rate function is obtained by analyzing the state parameters of the flowing sewage in the sewage pipe. This flow rate function is then compared with the function... The analysis and comparison are conducted, and a sewage system detection strategy is established based on the analysis results. It should be noted that the purpose of the above comparison is to determine whether there is a situation where rainwater enters the sewage system through pipelines or sewage wells by analyzing whether there is a positive relationship between the amount of sewage discharged and the amount of rainfall. If there is a positive relationship, it indicates that there may be rainwater entering the pipelines or sewage wells, and therefore the sewage system must be investigated and repaired.

[0079] The specific method for obtaining the state parameters of flowing sewage in the sewage pipe is as follows: Real-time flow rate of the sewage in the sewage pipe is obtained based on a flow meter. The real-time deepest water level of sewage in the pipeline is obtained based on the water level detector. .

[0080] The process of analyzing the state parameters of flowing sewage within a preset time period is as follows:

[0081] Through formula Obtain the function of the cross-sectional area of ​​the flowing sewage in the sewage pipe as a function of time. , where R is the inner radius of the sewage pipe;

[0082] Through formula Obtain wastewater flow function .

[0083] Through the above technical solution, this embodiment provides a method for measuring sewage flow rate in a sewage pipe, specifically:

[0084] First, we need to obtain the function of how the cross-sectional area of ​​the flowing sewage in the sewage pipe changes over time. It should be noted that sewage pipes are usually circular and typically lie horizontally at the sewage outlet. Therefore, a rectangular coordinate system can be established with the lowest point of the pipe's vertical cross-section as the origin, the horizontal radial direction as the x-axis, and the vertical direction as the y-axis. Through integration, according to the formula... Obtain the function of the cross-sectional area of ​​the flowing sewage in the sewage pipe as a function of time. Where R is the inner radius of the sewage pipe. Then, using the formula... The function of wastewater discharge volume changing over time was calculated. It should be noted that if the sewage pipe is square, the formula can be used... b Calculate the function of the corresponding sewage cross-sectional area changing with time. This embodiment solves the problem that it is difficult to detect the real-time flow rate in large sewage pipes because flow meters cannot be installed. At the same time, the above solution can also detect the real-time flow rate of sewage in sewage pipes that are not at full flow.

[0085] flow function AND function The process of analysis and comparison is as follows:

[0086] Take multiple intervals with lengths of interval ,in This represents the i-th time point taken in chronological order, and , , All are preset constants;

[0087] According to the formula and Calculate the total rainfall per unit time. Total amount of sewage discharged per unit time .

[0088] According to the formula Obtain the ratio of total pollution discharge to total rainfall per unit time. ,in , Eleven sets of proportional coefficients were measured using preset coefficients, and the status of the sewage system was determined based on these eleven sets of proportional coefficients.

[0089] The process of establishing a sewage system detection strategy is as follows:

[0090] Based on the status of the sewage system, according to the formula Calculate the variance of the proportionality coefficient for each group. ,in There are 11 sets of proportionality coefficients. The average;

[0091] like Then the sewage system needs to be inspected and repaired;

[0092] like If so, there is no need to inspect or repair the sewage system; among which As a preset standard value, and .

[0093] Through the above technical solution, this embodiment provides a function for wastewater discharge and a function for wastewater discharge. The specific process of analysis and comparison, as well as the establishment process of the sewage system detection strategy, firstly yielded 11 sets of proportional coefficients, specifically: taking multiple interval lengths as... interval ,in This represents the i-th time point taken in chronological order, and , , All are preset constants;

[0094] According to the formula and Calculate the total rainfall per unit time. Total amount of sewage discharged per unit time It should be noted that, assuming rainwater enters the sewage pipes, there is a time error between the rainfall captured by the video and the actual amount of rainwater entering the sewage pipes and affecting the sewage flow. This error is what is described above. ,and .

[0095] According to the formula Obtain the ratio of total pollution discharge to total rainfall per unit time. ,in , Eleven sets of proportionality coefficients were measured based on preset coefficients and fitted from experimental data. If these eleven sets of proportionality coefficients are relatively uniform, it indicates a relationship between sewage discharge and rainfall, suggesting a risk of rainwater continuously flowing into the sewage pipes. Therefore, this can be addressed using the formula... Calculate the variance of the proportionality coefficient for each group. ,in There are 11 sets of proportionality coefficients. The average; if Then the sewage system needs to be inspected to check if there is a long-term problem of rainwater flowing into the sewage pipes; if If so, there is no need to inspect or repair the sewage system; among which As a preset standard value, and .

[0096] Please refer to the attached diagram. Figure 2 As shown, a sewage discharge detection system under severe weather conditions is provided, the system comprising:

[0097] The information acquisition module is used to collect video image information of sewage at the sewage outlet;

[0098] Convolutional neural networks are used to classify acquired video image information;

[0099] The image processing module is used to perform sharpening processing on video images according to the type of video image information;

[0100] The rainfall detection module is used to obtain rainfall information based on video images of rainy weather types.

[0101] The wastewater detection module is used to detect and acquire the status parameters of the flowing wastewater in the wastewater pipe;

[0102] The analysis module is used to comprehensively analyze the obtained rainfall information and the state parameters of flowing sewage, and to determine the status of the sewage system.

[0103] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for detecting a pollution under adverse weather conditions, characterized in that, The method comprises the following steps: S1, obtaining video image information through a black light camera at a sewage outlet; S2, classifying the video image based on the collected video image information using a convolutional neural network; S3, performing corresponding sharpening processing on the type of video image; S4, placing the sharpened video image into a target detection neural network for sewage detection; The method further comprises: S5, analyzing the original video image in rainy day to obtain a function of rainfall change with time ; simultaneously obtaining the state parameters of the flowing sewage in the sewage pipe; analyzing the state parameters of the flowing sewage in a preset time period, and obtaining a sewage discharge flow function according to the analysis; comparing the flow function with the function , and establishing a sewage discharge system detection strategy according to the comparison result. The manner for acquiring the state parameters of the flowing sewage in the sewage pipe is specifically: acquiring the real-time flow rate of the sewage in the sewage pipe based on a flow rate meter , and acquiring the real-time deepest water level of the sewage in the pipe based on a water level detector . The process of analyzing the state parameters of the flowing sewage in the preset time period is: The function of the cross-sectional area of the flowing sewage in the sewage pipe with time is obtained by the formula where R is the inner radius of the sewage pipe;​ The wastewater flow function is obtained by the formula .​ 2. The method of claim 1, wherein, The types of video images include rainy days and foggy days, and the sharpening processing for the rainy day type in step S3 comprises: Based on the original video image, a rough binary image is obtained, and the binary image includes a rain line layer and a background layer; Using principal component analysis to process the binary image connected domain, remove the high-frequency details of the image, and obtain a refined binary image; The refined binary image is used as a mask, and the mask is removed from the original video image to obtain a video image without rain.

3. The method of claim 2, wherein, The sharpening processing for the foggy day type in step S3 comprises: Based on the original video image, a dark channel prior dehazing algorithm is used to dehaze, specifically: The input image is subjected to adaptive filtering processing; then the dark channel value of each pixel point is calculated; Determine the position of the largest dark channel value in the image, and use the pixel value at this position as a reference value to determine the intensity of the atmospheric light; According to the least square algorithm, the intensity of the atmospheric light and the dark channel curve are used to solve the intensity of the fog in the image; According to the atmospheric light and the intensity of the fog obtained above, the original video image is calculated, thereby realizing image dehazing.

4. The method of claim 1, wherein, The flow function with the function The process of analysis comparison is: Take multiple interval lengths interval , wherein represents the i-th time point taken in time order, and , , are all preset constants; The total amount of rainfall per unit time and The total amount of pollution per unit time and are calculated according to the formula According to the formula The proportional coefficient of the total amount of pollution and the total amount of rainfall in a unit of time is obtained Wherein , Both are preset coefficients, n groups of proportional coefficients are measured, n>10, and the state of the pollution discharge system is determined according to the n groups of proportional coefficients.

5. The method of claim 4, wherein, The process of establishing the sewage system detection strategy is: For the state of the sewage system, according to the formula The variance value of each group of proportional coefficients is calculated , wherein is the average of n groups of proportional coefficients ; If then the waste system needs to be serviced; If , then no maintenance is required for the sewage system; wherein is a preset standard value, and .

Images