Automatic cleaning device for membrane bioreactor
By combining image acquisition and control system in the automatic cleaning device of the membrane bioreactor, analyzing the changes in the membrane surface and adjusting the brush angle, the problem that the fixed angle brush cannot adapt to the irregular membrane surface is solved, and the cleaning efficiency and quality are significantly improved .
Patent Information
- Application Number
- CN202411232427.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-09-04
AI Technical Summary
In existing automatic cleaning devices for membrane bioreactors, fixed angle brushes are difficult to adapt to irregularities and specific inclinations of the membrane surface, resulting in limited cleaning effects, especially in the corners and edge areas of the membrane assembly that cannot effectively remove dirt.
An automatic cleaning device for membrane bioreactors including an image acquisition system and a control system is designed. By monitoring the changes in the surface of the video acquisition membrane, analyzing the grayscale difference in corner areas, and adjusting the inclination angle of the brush to achieve adaptive cleaning.
It improves the cleaning efficiency and quality of the membrane bioreactor, ensures in-depth and comprehensive cleaning of difficult-to-reach areas such as corners and edges, and improves the performance of the automatic cleaning device.
Smart Images

Figure CN118718751B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of sewage treatment, and in particular to an automatic cleaning device for a membrane bioreactor. Background Art
[0002] Membrane Bio-Reactor (MBR) is an efficient sewage treatment technology. The membrane surface, one of its core components, is prone to accumulate dirt and biofilm during long-term use. These accumulations may form irregular distribution patterns on the membrane surface, and due to the design and operating conditions of the membrane assembly, the membrane surface may have specific tilt angles or uneven areas. In the existing automatic cleaning device for membrane bioreactors (utility model patent with publication number CN220642765U), the brush brushes the membrane bioreactor at a fixed angle, but the brush is fixed at an angle, and it may not be able to adapt to these irregularities and tilts, resulting in limited cleaning effects. The fixed-angle brush may be effective in cleaning flat areas, but in tilted or hard-to-reach areas, such as corners or edges of membrane assemblies, it may not provide sufficient contact pressure or appropriate cleaning power, and thus cannot effectively remove all dirt from the membrane bioreactor. Summary of the invention
[0003] In order to solve the problem that some dirt areas cannot be effectively removed during the cleaning of membrane bioreactors using brushes with fixed slope angles, the present invention aims to provide an automatic cleaning device for membrane bioreactors, and the technical solutions adopted are as follows:
[0004] The present invention provides an automatic cleaning device for a membrane bioreactor, comprising a main housing of the membrane bioreactor, wherein the main housing of the membrane bioreactor comprises a cleaning component, wherein the cleaning component comprises a brush, and the main housing of the membrane bioreactor further comprises a driving component, wherein the driving component is used to adjust the inclination angle of the brush; the automatic cleaning device for the membrane bioreactor further comprises an image acquisition system and a control system;
[0005] The image acquisition system is used to collect monitoring video of a process in which a brush cleans a membrane bioreactor at a fixed angle, wherein the monitoring video is composed of multiple frames of monitoring images;
[0006] The control system is used to obtain a differential image corresponding to two adjacent frames of monitoring images, perform superpixel segmentation on each differential image to obtain superpixel blocks, match the superpixel blocks in the two adjacent differential images to obtain a plurality of superpixel block sequences, wherein adjacent elements in each superpixel block sequence have a matching relationship; determine the corresponding dirt consistency index based on the overlap of adjacent elements in each superpixel block sequence and the grayscale distribution of the corresponding area in the monitoring image; screen the corner area based on the continuity of the dirt consistency index; determine the brush angle adjustment parameter based on the grayscale difference of the corner area in all monitoring images; and control the slope angle of the brush based on the brush angle adjustment parameter.
[0007] Preferably, matching super pixel blocks in two adjacent differential images to obtain a plurality of super pixel block sequences includes:
[0008] The KM matching algorithm is used to match the superpixel blocks in each two adjacent differential images to obtain the corresponding matching relationship;
[0009] Based on the matching relationship, the superpixel blocks in all differential images are sorted in sequence according to the acquisition time to obtain a corresponding superpixel block sequence, wherein adjacent elements in the superpixel block sequence are pixel blocks in different differential images, and the number of elements in each superpixel block sequence is the number of frames of the differential image.
[0010] Preferably, the method of combining the overlap of adjacent elements in each superpixel block sequence and the grayscale distribution of the corresponding area in the monitoring image to determine the corresponding dirt consistency index includes:
[0011] Calculate the grayscale mean of each superpixel block in each frame of the monitoring image, and use the grayscale mean of each superpixel block as the corresponding dirtiness;
[0012] Respectively obtain three super-pixel blocks corresponding to each two adjacent elements in the candidate sequence in the corresponding three frames of monitoring images, and obtain the dirt consistency index of the three super-pixel blocks according to the discrete degree of dirtiness of the three super-pixel blocks, wherein the discrete degree is negatively correlated with the dirt consistency index;
[0013] The candidate sequence is any super pixel block sequence.
[0014] Preferably, the continuous screening of corner areas based on the dirt consistency index includes:
[0015] Determine the intersection-and-union ratio corresponding to each two adjacent elements in the candidate sequence based on the coordinates of the pixel points in each two adjacent elements in the candidate sequence; sort all the intersection-and-union ratios corresponding to the candidate sequence in chronological order to obtain an intersection-and-union ratio sequence; the elements in the intersection-and-union ratio sequence whose element values are continuously greater than a preset first threshold constitute a first subsequence;
[0016] The dirt consistency indicators corresponding to all elements in the first subsequence constitute a dirt consistency indicator sequence; the elements in the dirt consistency indicator sequence whose element values are greater than a preset consistency threshold constitute a second subsequence; and the intersection area corresponding to the elements in the second subsequence is used as a corner area;
[0017] The candidate sequence is any super pixel block sequence.
[0018] Preferably, the step of determining the brush angle adjustment parameter according to the grayscale difference of the corner area in all monitoring images includes:
[0019] Calculate the grayscale mean of all pixels in each superpixel block in the sequence to be analyzed; the grayscale mean of all pixels in all superpixel blocks in the sequence to be analyzed constitutes a grayscale mean sequence; the difference between the latter element and the former element in two adjacent elements in the grayscale mean sequence constitutes a difference sequence; count the number of elements in the difference sequence whose element values are greater than 0, and if the number ratio is greater than a preset ratio threshold, determine that the corner area corresponding to the sequence to be analyzed is the area to be re-cleaned; the sequence to be analyzed is a superpixel block sequence where any corner area is located;
[0020] The brush angle adjustment parameter is obtained by combining the number of all areas to be re-cleaned.
[0021] Preferably, the step of obtaining the brush angle adjustment parameter in combination with the number of all areas to be re-cleaned includes:
[0022] The difference between the total number of corner areas and the number of all areas to be re-cleaned is calculated; and the ratio between the difference and the total number of corner areas is used as a brush angle adjustment parameter.
[0023] Preferably, the controlling the slope angle of the brush based on the brush angle adjustment parameter comprises:
[0024] If the brush adjustment parameter is greater than a preset adjustment threshold, the slope angle of the brush is adjusted;
[0025] If the brush adjustment parameter is less than or equal to the preset adjustment threshold, the slope angle of the brush is not adjusted.
[0026] Preferably, obtaining the discrete degrees of dirtiness of the three super-pixel blocks includes: taking the variance of the dirtiness of the three super-pixel blocks as the discrete degrees of dirtiness of the corresponding three super-pixel blocks.
[0027] Preferably, the acquisition of the differential image corresponding to two adjacent frames of surveillance images includes:
[0028] For any two adjacent surveillance images:
[0029] The grayscale difference between two pixels at corresponding positions in two adjacent surveillance images is calculated respectively; and the grayscale difference is used as the grayscale value of the pixel at corresponding position in the differential image corresponding to the two adjacent surveillance images.
[0030] Preferably, the driving component includes a crank, a connecting rod, a rocker, a drive shaft and a mounting plate, the rocker and the brush are connected via a universal joint, the universal joint and the drive shaft are mounted on the mounting plate, one end of the connecting rod is connected to the crank, and the other end is connected to the rocker, and the crank is connected to the drive shaft.
[0031] The present invention has at least the following beneficial effects:
[0032] The present invention improves the cleaning part on the basis of the general membrane bioreactor automatic cleaning device, adds a driving component for adjusting the tilt angle of the brush, and the image acquisition system first acquires the monitoring image of the brush cleaning the membrane bioreactor at a fixed angle, and the control system analyzes the acquired monitoring image, and selects the corner area in combination with the specific change characteristics of the membrane bioreactor surface during the cleaning process. The corner area generally hides a lot of dirt, and the corner area is often an area that is difficult to clean during the cleaning process. Further, according to the grayscale difference of the corner area in all monitoring images, the brush angle adjustment parameter is determined, and the cleaning condition of the membrane bioreactor is evaluated based on the brush angle adjustment parameter, and it is determined whether the brush tilt angle needs to be adjusted, and then the control system controls the slope angle of the brush, and realizes the dynamic adjustment of the contact angle between the brush and the membrane surface. This operation enables the brush to be adaptively adjusted according to the irregularity of the membrane surface and the specific tilt angle, so as to better fit the membrane surface and effectively cover the cleaning dead corner. By operating the control system, the angle change of the tilt module can be accurately controlled, so that the brush always maintains the best contact state with the membrane surface during rotation and reciprocating motion. The improvements in the present invention significantly improve the cleaning efficiency and quality of the membrane bioreactor, ensure that even in the hard-to-reach areas such as corners and edges of the membrane bioreactor, in-depth and comprehensive cleaning can be achieved, and improve the performance of the entire membrane bioreactor automatic cleaning device. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. 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 creative work.
[0034] Figure 1A schematic diagram of the mechanical structure of an automatic cleaning device for a membrane bioreactor provided by an embodiment of the present invention;
[0035] Figure 2 A schematic diagram of the mechanical structure of a driving component of an automatic cleaning device for a membrane bioreactor provided by an embodiment of the present invention;
[0036] Figure 3 A schematic diagram of a brush of an automatic cleaning device for a membrane bioreactor provided by an embodiment of the present invention after the tilt angle is adjusted;
[0037] Figure 4 A schematic diagram of a data processing flow corresponding to an automatic cleaning device for a membrane bioreactor provided in an embodiment of the present invention;
[0038] In the figure, 1 is the main box of the membrane bioreactor, 2 is the brush, 3 is the first motor, 4 is the driving sprocket, 5 is the driven sprocket, 6 is the transmission chain, 7 is the spray mechanism, 8 is the mounting plate, 9 is the crank, 10 is the connecting rod, 11 is the rocker, 12 is the universal joint, 13 is the second motor, 14 is the driving shaft, and 15 is the control system. DETAILED DESCRIPTION
[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, an automatic cleaning device for a membrane bioreactor proposed in accordance with the present invention is described in detail below in conjunction with the accompanying drawings and preferred embodiments.
[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0041] The specific scheme of the automatic cleaning device for a membrane bioreactor provided by the present invention is described in detail below with reference to the accompanying drawings.
[0042] See also Figure 1, which shows a structural block diagram of an automatic cleaning device for a membrane bioreactor provided by an embodiment of the present invention, the automatic cleaning device for the membrane bioreactor includes a membrane bioreactor main box 1, the membrane bioreactor main box 1 includes a cleaning component and a driving component, the cleaning component includes a brush 2, this embodiment adds a driving component to the membrane bioreactor main box 1 on the basis of a general automatic cleaning device for a membrane bioreactor, the driving component is used to adjust the inclination angle of the brush 2, in this embodiment, the driving component is mainly composed of a crank 9, a connecting rod 10, a rocker 11, a universal joint 12, a second motor 13, a drive shaft 14 and a mounting plate 8, so the membrane bioreactor main box 1 includes a first motor 3, a driving sprocket 4, a driven sprocket 5, a transmission chain 6, a spray mechanism 7, a mounting plate 8, a crank 9, a connecting rod 10, a rocker 11, a universal joint 12, a second motor 13, a drive shaft 14, a control system 15 and an image acquisition system. In the driving device of this embodiment, the rocker 11 and the brush 2 are connected via a universal joint 12, the universal joint 12 and the drive shaft 14 are mounted on the mounting plate 8, one end of the connecting rod 10 is connected to the crank 9, and the other end is connected to the rocker 11, and the crank 9 is connected to the drive shaft 14.
[0043] The signal output end of the image acquisition system is connected to the input end of the control system 15 . The image acquisition system is used to acquire monitoring video of the membrane bioreactor. The control system 15 is used to process the acquired monitoring video of the membrane bioreactor and control the angle of the brush 2 .
[0044] In this embodiment, the specific operation process of the membrane bioreactor automatic cleaning device is as follows: the membrane bioreactor is connected to the water production tank through a water production pipe, and the water in the water production tank is used as a cleaning water source; first, the brushes 2 under the hollow telescopic rods on both sides of the membrane bioreactor are driven by a motor to lift and lower, and at the same time, the nozzle inside the hollow telescopic rod obtains water flow through a connecting pipe connected to the water production tank, and the atomizing nozzle at the bottom of the nozzle atomizes the water flow and sprays it onto the surface of the membrane bioreactor for physical flushing, thereby achieving physical brushing of the surface of the membrane bioreactor; then, the image acquisition system is used to collect multi-frame monitoring images of the membrane bioreactor, and the image acquisition system inputs the collected monitoring images into the control system 15, and the control The control system 15 processes the monitoring image data, and judges whether the angle of the brush 2 needs to be adjusted in combination with the cleaning condition of the membrane bioreactor. If the membrane bioreactor is not cleaned clean enough, the second motor 13 rotates, driving the drive shaft 14 to make a rotating circular motion; the movement of the drive shaft 14 drives the crank 9 to make a circular motion; the movement of the crank 9 drives the connecting rod 10 to make a swinging motion; the connecting rod 10 drives the rocker 11 to make a swinging motion, and the rocker 11 drives the brush 2 to make a swinging motion, that is, the brush 2 under the hollow telescopic rods on both sides of the membrane bioreactor is driven by the motor to perform lifting and rotating actions, realizing the rotation of the brush tilt angle, completing the secondary cleaning of the membrane bioreactor, and achieving a better cleaning effect. In addition, the automatic cleaning device of the membrane bioreactor also includes a delivery pump and an electric regulating valve to control the flow and pressure of the cleaning water.
[0045] In this embodiment, the existing membrane bioreactor automatic cleaning device is improved, that is, a driving component is added, and the operation process of other components can be found in the utility model patent with publication number CN220642765U. The power transmission process of the cleaning component of the membrane bioreactor automatic cleaning device in this embodiment is specifically as follows:
[0046] 1. The circular motion of the input rocker is converted into the circular motion of the crank, and the circular motion of the crank is converted into the rotational motion of the connecting rod. The rotation of the connecting rod drives the rocker to rotate, and the rotation of the rocker drives one end of the brush to rotate, thereby realizing the rotation of the brush, that is, the adjustment of the angle;
[0047] 2. Shaft or hinge: installed on the frame of the membrane bioreactor and serves as the rotation fulcrum of the connecting rod;
[0048] 3. Driving device: installed at one end of the connecting rod, opposite to the rotating shaft or hinge, used to drive the connecting rod to rotate;
[0049] 4. Control system: integrated on the drive device, used to receive sensor signals and control the drive device;
[0050] 5. Support structure: installed on the frame of the cleaning component, used to fix the shaft and hinge to ensure the stability of the connecting rod mechanism.
[0051] During the process of cleaning the membrane bioreactor, some areas may be difficult to clean. Therefore, if a brush with a fixed tilt angle is used to clean the membrane bioreactor, the corner area often cannot achieve a good cleaning effect. Therefore, the automatic cleaning device of the membrane bioreactor provided in this embodiment evaluates the cleaning condition of the membrane bioreactor in combination with the image of the membrane bioreactor, and then determines whether the angle of the brush needs to be adjusted to ensure the cleaning quality of the membrane bioreactor.
[0052] See also Figure 4 , which shows a method flow chart of specific implementation steps of an image acquisition system and a control system in an automatic cleaning device for a membrane bioreactor provided by an embodiment of the present invention, specifically including:
[0053] The image acquisition system is used to collect monitoring video of the process in which the brush 2 cleans the membrane bioreactor at a fixed angle, and the monitoring video is composed of multiple frames of monitoring images.
[0054] First, the angle of the brush 2 is fixed, and the surface of the membrane bioreactor is cleaned. The image acquisition system is used to collect multiple monitoring images of the membrane bioreactor surface. It should be noted that all the monitoring images collected in this embodiment are monitoring images of the membrane bioreactor surface collected during the cleaning process of the membrane bioreactor at the same fixed angle. In order to prevent the influence of the water flow on the collected monitoring images during the cleaning process, this embodiment stops the membrane bioreactor every 20 seconds, and the stop time is 20 seconds, that is, when the water flow no longer changes, a monitoring image of the membrane bioreactor surface is collected, and then the membrane bioreactor is cleaned again. The cleaning process is repeated for multiple times, and multiple monitoring images of the membrane bioreactor surface are collected. In this embodiment, a total of 20 monitoring images are collected. In specific applications, the implementer can set the cleaning time, cleaning stop time and the number of monitoring images according to specific circumstances. Each frame of the monitoring image of the membrane bioreactor surface collected is grayed. The graying of the image is a prior art and will not be described in detail here. It should be noted that the monitoring images mentioned later are all grayed images.
[0055] The control system 15 is used to obtain a differential image corresponding to two adjacent frames of monitoring images, perform superpixel segmentation on each differential image to obtain superpixel blocks, match the superpixel blocks in the two adjacent differential images to obtain a plurality of superpixel block sequences, and adjacent elements in each superpixel block sequence have a matching relationship; determine the corresponding dirt consistency index based on the overlap of adjacent elements in each superpixel block sequence and the grayscale distribution of the corresponding area in the monitoring image; screen the corner area based on the continuity of the dirt consistency index; determine the brush angle adjustment parameter based on the grayscale difference of the corner area in all monitoring images; and control the slope angle of the brush based on the brush angle adjustment parameter.
[0056] Considering that there are darker and whiter areas on the surface of the membrane bioreactor itself, the area directly obtained by methods such as threshold segmentation or corner detection is not necessarily the corner area. Due to its shape, the corner area often becomes a retention area for water flow, causing water to gather in these areas and not easily flow away quickly. At the same time, there is more dirt at the corners that are difficult to clean, and the water flow there is generally black or gray.
[0057] For any two adjacent surveillance images, obtain the coordinates of each pixel in the two surveillance images. Since there is a one-to-one correspondence between the positions of the pixels in the two surveillance images, for any pixel in any of the two surveillance images, there is a pixel corresponding to its position in the other surveillance image. For any position in the two surveillance images: calculate the absolute value of the difference between the grayscale values of the two pixels at that position, as the grayscale value of that position in the differential image. By using this method, the grayscale value of each position in the differential image can be obtained, that is, the differential image corresponding to the two surveillance images can be obtained. By using the above method, the differential image corresponding to every two adjacent surveillance images can be obtained.
[0058] For each frame of differential image, a superpixel segmentation algorithm is used to perform superpixel segmentation processing on each frame of differential image to obtain multiple superpixel blocks in each frame of differential image. Next, the KM matching algorithm will be used to match the superpixel blocks in two adjacent frames of differential image. The superpixel segmentation algorithm is a prior art and will not be described in detail here.
[0059] The KM matching algorithm is used to match the superpixel blocks in each two adjacent differential images. The existing KM matching is to match the nodes on the left and right sides one-to-one. Initially, each node on the left has an edge with all nodes on the right. In this embodiment, all superpixel blocks in the previous differential image of the two adjacent differential images are used as left nodes, and each superpixel block corresponds to a node. All superpixel blocks in the next differential image of the two adjacent differential images are used as right nodes. The node value is the center coordinate of each superpixel block. Each node on the left has an edge with each node on the right. The edge value is the Euclidean distance of the corresponding two node coordinates. Following the minimum matching principle, a one-to-one matching of the left node and the right node is obtained through KM matching calculation, that is, a one-to-one matching relationship between each superpixel block in the previous differential image and each superpixel block in the next differential image is obtained. Based on the matching relationship, the superpixel blocks in all differential images are sorted in order of acquisition time to obtain the corresponding superpixel block sequence. , that is, a plurality of superpixel block sequences are obtained, wherein adjacent elements in the superpixel block sequence are pixel blocks in different differential images, and the number of elements in each superpixel block sequence is the number of frames of the differential image; for example: for any superpixel block sequence: the first element in the superpixel block sequence is a superpixel block in the first frame of the monitoring image, the second element in the superpixel block sequence is a superpixel block in the second frame of the monitoring image that successfully matches the first element in the superpixel block sequence, the third element in the superpixel block sequence is a superpixel block in the third frame of the monitoring image that successfully matches the second element in the superpixel block sequence, the fourth element in the superpixel block sequence is a superpixel block in the fourth frame of the monitoring image that successfully matches the third element in the superpixel block sequence, and so on. The total number of elements in the superpixel block sequence is equal to the total number of monitoring images, that is, different elements in the superpixel block sequence are superpixel blocks in different monitoring images, and two adjacent superpixel blocks in the superpixel block sequence are successfully matched superpixel blocks. The KM matching algorithm is an existing technology and will not be described in detail here.
[0060] If the area on the reactor corresponding to the adjacent elements in the superpixel block sequence is a dirty area, the color of the water flow at the corresponding position is darker. Therefore, the grayscale similarity is combined to further analyze the difficult-to-clean areas such as corners.
[0061] First, the grayscale mean of each superpixel block in each frame of the monitoring image is calculated, and the grayscale mean of each superpixel block is used as the corresponding dirtiness, that is, the dirtiness corresponding to each superpixel block in each frame of the monitoring image is obtained; for any superpixel block, the grayscale mean of the superpixel block is obtained as follows: according to the grayscale value of each pixel in the superpixel block, the average grayscale value of all pixels in the superpixel block is calculated, and the average grayscale value is used as the grayscale mean of the superpixel block.
[0062] Next, any superpixel block sequence is taken as an example for explanation, and other superpixel block sequences can be processed by the method provided in this embodiment. Specifically, any superpixel block sequence is recorded as a candidate sequence. Since any element in the superpixel block sequence corresponds to a local area in two frames of monitoring images, every two adjacent elements in the candidate sequence correspond to three local areas in three consecutive frames of monitoring images, that is, correspond to three superpixel blocks. The three superpixel blocks corresponding to every two adjacent elements in the candidate sequence in the corresponding three frames of monitoring images are obtained respectively. According to the discrete degree of dirtiness of the three superpixel blocks, the dirtiness consistency index of the three superpixel blocks is obtained, and the discrete degree is negatively correlated with the dirtiness consistency index.
[0063] Among them, the discrete degree of dirtiness of the three super-pixel blocks can be characterized by the standard deviation of the dirtiness of the three super-pixel blocks, or by the variance of the dirtiness of the three super-pixel blocks, or by the extreme difference of the dirtiness of the three super-pixel blocks, etc. In this embodiment, the variance of the dirtiness of the three super-pixel blocks is used to characterize the discrete degree of dirtiness of the three super-pixel blocks, that is, the variance of the dirtiness of the three super-pixel blocks is used as the corresponding discrete degree. The larger the variance, the more discrete the dirtiness of the three super-pixel blocks. The negative correlation relationship indicates that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtraction relationship, a division relationship, etc., which is determined by actual application. In this embodiment, the value of the exponential function with a natural constant as the base and the negative dirtiness of the three super-pixel blocks as the exponent is used as the dirtiness consistency index of the three super-pixel blocks.
[0064] Next, the intersection-and-union ratio corresponding to each two adjacent elements in the candidate sequence is determined based on the coordinates of the pixels in each two adjacent elements in the candidate sequence; for any two adjacent elements in the candidate sequence, the specific process of obtaining the intersection-and-union ratio corresponding to the two adjacent elements is: respectively counting the number of intersection areas of the pixels in the areas corresponding to the two elements and the number of union areas of the pixels in the areas corresponding to the two elements, and taking the ratio between the number of intersection areas of the pixels in the areas corresponding to the two elements and the number of union areas of the pixels in the areas corresponding to the two elements as the intersection-and-union ratio corresponding to the two adjacent elements. By adopting the above method, the intersection-and-union ratio corresponding to each two adjacent elements in the candidate sequence can be obtained. In chronological order, all the intersection-and-union ratios corresponding to the candidate sequence are sorted to obtain an intersection-and-union ratio sequence; the elements in the intersection-and-union ratio sequence whose element values are continuously greater than the preset first threshold constitute a first subsequence. The dirt consistency indicators corresponding to all elements in the first subsequence constitute a dirt consistency indicator sequence; the elements in the dirt consistency indicator sequence whose element values are greater than the preset consistency threshold constitute a second subsequence; the intersection area in the monitoring image corresponding to the elements in the second subsequence is taken as the corner area. In this embodiment, the preset first threshold and the preset consistent threshold are both 0.7. In specific applications, the implementer can set them according to specific circumstances.
[0065] If an area is not easy to clean, the color of the water flow at that location will not gradually become lighter during the cleaning process, but will show an overall trend of becoming lighter. This is because during the cleaning process, for the corners that are difficult to clean, the more superficial dirt will be cleaned away first, and the difficult-to-clean dirt will begin to be cleaned out when the water flushing time reaches a certain level. The repeated cleaning process will cause the water flow color at that location to become lighter overall, but there will be a darkening process in the middle.
[0066] Next, any superpixel block sequence is used as an example for explanation. The method provided in this embodiment can be used to process other superpixel block sequences. Specifically, any superpixel block sequence is recorded as a sequence to be analyzed, and the grayscale mean of all pixels in each superpixel block in the sequence to be analyzed is calculated; the grayscale mean of all pixels in all superpixel blocks in the sequence to be analyzed constitutes a grayscale mean sequence; the difference between the latter element and the former element in the two adjacent elements in the grayscale mean sequence constitutes a difference sequence; the number of elements with element values greater than 0 in the difference sequence is counted, and if the number ratio is greater than the preset ratio threshold, the corner area corresponding to the sequence to be analyzed is determined to be the area to be re-cleaned; in this embodiment, the preset ratio threshold is 0.3, and in specific applications, the implementer can set it according to the specific situation. The specific calculation method for counting the number of elements with element values greater than 0 in the difference sequence is: the ratio between the number of elements with element values greater than 0 in the difference sequence and the total number of elements in the difference sequence is taken as the number ratio of elements with element values greater than 0 in the difference sequence. The difference between the total number of corner areas and the number of all areas to be re-cleaned is calculated; it should be noted that the method for obtaining the difference between the total number of corner areas and the number of all areas to be re-cleaned is: the difference obtained by subtracting the number of all areas to be re-cleaned from the total number of corner areas is used as the difference between the total number of corner areas and the number of all areas to be re-cleaned. The ratio between the difference and the total number of corner areas is used as the brush angle adjustment parameter.
[0067] If the brush adjustment parameter is greater than the preset adjustment threshold, it means that the effect of using the brush angle at this time to clean the membrane bioreactor is not ideal, and some dirt areas cannot be cleaned. At this time, the slope angle of the brush is adjusted, and the specific adjustment angle of the brush is set by the implementer according to the specific situation; if the brush adjustment parameter is less than or equal to the preset adjustment threshold, it means that the effect of using the brush angle at this time to clean the membrane bioreactor is good and can meet the cleaning requirements, so the slope angle of the brush is not adjusted. In this embodiment, the preset adjustment threshold is 0.3. In specific applications, the implementer can set it according to the specific situation.
[0068] The working process of automatic cleaning of membrane bioreactor is:
[0069] The water in the water production tank is pumped to the nozzle through the connecting pipe; the atomizing nozzle atomizes the water and sprays it onto the surface of the membrane bioreactor for physical flushing. At the same time, the motor drives the hollow shaft to rotate, driving the hollow telescopic rod and the brush to rotate and lift at high speed, thereby achieving comprehensive cleaning of the membrane bioreactor. During the cleaning process, the multiple frames of monitoring images collected by the image acquisition system are used as input. The control system analyzes and processes the monitoring images to determine whether to adjust the slope angle of the brush. When the brush angle needs to be adjusted, the brush tilting module works as follows: [1] The drive device starts: the motor starts and rotates to generate input motion; [2] The crank rotates: the rotation of the input connecting rod causes the crank to rotate around a fixed axis; [3] The connecting rod swings: the rotation of the crank is transmitted to the connecting rod, causing the connecting rod to swing under the drive of the crank; [4] The rocker swings: the swing of the connecting rod is transmitted to the rocker, and the rocker swings on the fixed frame; [5] The brush angle is adjusted: the swing of the rocker causes the brush support structure to move, thereby adjusting the tilt angle of the brush; such as Figure 2 and Figure 3 As shown in the figure, these two figures show the schematic diagrams before and after the brush angle is adjusted. The reactor is cleaned by the tilted brush. After the cleaning is completed, the brush automatically rises to the top of the membrane area without affecting the water production function of the membrane.
[0070] Thus, the membrane bioreactor can be cleaned by the method provided in this embodiment, and the cleaning quality of the membrane bioreactor can be ensured by adjusting the angle of the brush.
[0071] This embodiment is improved on the basis of the general membrane bioreactor automatic cleaning device, and a driving component for adjusting the tilt angle of the brush is added. The image acquisition system first acquires the monitoring image of the brush cleaning the membrane bioreactor at a fixed angle. The control system analyzes the acquired monitoring image, and selects the corner area in combination with the specific change characteristics of the membrane bioreactor surface during the cleaning process. The corner area generally hides a lot of dirt, and the corner area is often an area that is difficult to clean during the cleaning process. Further, according to the grayscale difference of the corner area in all monitoring images, the brush angle adjustment parameter is determined. The cleaning condition of the membrane bioreactor is evaluated based on the brush angle adjustment parameter, and it is determined whether the brush tilt angle needs to be adjusted. Then, the control system controls the slope angle of the brush, and realizes the dynamic adjustment of the contact angle between the brush and the membrane surface. This operation enables the brush to be adaptively adjusted according to the irregularity of the membrane surface and the specific tilt angle, so as to better fit the membrane surface and effectively cover the cleaning dead corner. By operating the control system, the angle change of the tilt module can be accurately controlled, so that the brush always maintains the best contact state with the membrane surface during rotation and reciprocating motion. The improvements in this embodiment significantly improve the cleaning efficiency and quality of the membrane bioreactor, ensuring that in-depth and comprehensive cleaning can be achieved even in hard-to-reach areas such as corners and edges of the membrane bioreactor, thereby improving the performance of the entire membrane bioreactor automatic cleaning device.
[0072] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. An automatic cleaning device for a membrane bioreactor, comprising a membrane bioreactor main body (1), the membrane bioreactor main body (1) comprising a cleaning component, the cleaning component comprising a brush (2), characterized in that: The membrane bioreactor main box (1) further comprises a driving component, wherein the driving component is used to adjust the inclination angle of the brush (2); the membrane bioreactor automatic cleaning device further comprises an image acquisition system and a control system (15); The image acquisition system is used to collect monitoring video of a process in which a brush (2) cleans a membrane bioreactor at a fixed angle, wherein the monitoring video is composed of a plurality of monitoring image frames; The control system (15) is used to obtain a differential image corresponding to two adjacent frames of monitoring images, perform superpixel segmentation on each differential image to obtain a superpixel block, match the superpixel blocks in the two adjacent differential images to obtain a plurality of superpixel block sequences, wherein adjacent elements in each superpixel block sequence have a matching relationship; determine a corresponding dirt consistency index based on the overlap of adjacent elements in each superpixel block sequence and the grayscale distribution of the corresponding area in the monitoring image; and screen the corner area based on the continuity of the dirt consistency index; Determine brush angle adjustment parameters according to the grayscale differences of the corner area in all monitoring images; and control the slope angle of the brush based on the brush angle adjustment parameters; The super pixel blocks in two adjacent differential images are matched to obtain a plurality of super pixel block sequences, including: The KM matching algorithm is used to match the superpixel blocks in each two adjacent differential images to obtain the corresponding matching relationship; Based on the matching relationship, the superpixel blocks in all the differential images are sorted in the order of acquisition time to obtain a corresponding superpixel block sequence, wherein adjacent elements in the superpixel block sequence are pixel blocks in different differential images, and the number of elements in each superpixel block sequence is the number of frames of the differential image; The step of determining the brush angle adjustment parameter according to the grayscale difference of the corner area in all monitoring images includes: Calculate the grayscale mean of all pixels in each superpixel block in the sequence to be analyzed; the grayscale mean of all pixels in all superpixel blocks in the sequence to be analyzed constitutes a grayscale mean sequence; the difference between the latter element and the former element in two adjacent elements in the grayscale mean sequence constitutes a difference sequence; count the number of elements in the difference sequence whose element values are greater than 0, and if the number ratio is greater than a preset ratio threshold, determine that the corner area corresponding to the sequence to be analyzed is the area to be re-cleaned; the sequence to be analyzed is a superpixel block sequence where any corner area is located; The brush angle adjustment parameter is obtained by combining the number of all areas to be re-cleaned; The driving component comprises a crank (9), a connecting rod (10), a rocking arm (11), a driving shaft (14) and a mounting plate (8); the rocking arm (11) and the brush (2) are connected via a universal joint (12); the universal joint (12) and the driving shaft (14) are mounted on the mounting plate (8); one end of the connecting rod (10) is connected to the crank (9) and the other end is connected to the rocking arm (11); the crank (9) is connected to the driving shaft (14).
2. The automatic cleaning device for a membrane bioreactor according to claim 1, characterized in that: The method of combining the overlapping of adjacent elements in each superpixel block sequence and the grayscale distribution of the corresponding area in the monitoring image to determine the corresponding dirt consistency index includes: Calculate the grayscale mean of each superpixel block in each frame of the monitoring image, and use the grayscale mean of each superpixel block as the corresponding dirtiness; Respectively obtain three super-pixel blocks corresponding to each two adjacent elements in the candidate sequence in the corresponding three frames of monitoring images, and obtain the dirt consistency index of the three super-pixel blocks according to the discrete degree of dirtiness of the three super-pixel blocks, wherein the discrete degree is negatively correlated with the dirt consistency index; The candidate sequence is any super pixel block sequence.
3. The automatic cleaning device for a membrane bioreactor according to claim 1, characterized in that: The continuous screening of corner areas based on the dirt consistency index includes: Determine the intersection-and-union ratio corresponding to each two adjacent elements in the candidate sequence based on the coordinates of the pixel points in each two adjacent elements in the candidate sequence; sort all the intersection-and-union ratios corresponding to the candidate sequence in chronological order to obtain an intersection-and-union ratio sequence; the elements in the intersection-and-union ratio sequence whose element values are continuously greater than a preset first threshold constitute a first subsequence; The dirt consistency indicators corresponding to all elements in the first subsequence constitute a dirt consistency indicator sequence; the elements in the dirt consistency indicator sequence whose element values are greater than a preset consistency threshold constitute a second subsequence; and the intersection area corresponding to the elements in the second subsequence is used as a corner area; The candidate sequence is any super pixel block sequence.
4. The automatic cleaning device for a membrane bioreactor according to claim 1, characterized in that: The step of obtaining the brush angle adjustment parameter by combining the number of all areas to be re-cleaned includes: The difference between the total number of corner areas and the number of all areas to be re-cleaned is calculated; and the ratio between the difference and the total number of corner areas is used as a brush angle adjustment parameter.
5. The automatic cleaning device for a membrane bioreactor according to claim 1, characterized in that: The controlling the slope angle of the brush based on the brush angle adjustment parameter comprises: If the brush angle adjustment parameter is greater than a preset adjustment threshold, the slope angle of the brush is adjusted; If the brush angle adjustment parameter is less than or equal to a preset adjustment threshold, the slope angle of the brush is not adjusted.
6. The automatic cleaning device for a membrane bioreactor according to claim 2, characterized in that: The acquisition of the discrete degrees of dirtiness of the three super-pixel blocks includes: taking the variance of the dirtiness of the three super-pixel blocks as the discrete degrees of dirtiness of the corresponding three super-pixel blocks.
7. The automatic cleaning device for a membrane bioreactor according to claim 1, characterized in that: The acquisition of the differential image corresponding to the two adjacent frames of monitoring images includes: For any two adjacent surveillance images: The grayscale difference between two pixels at corresponding positions in two adjacent surveillance images is calculated respectively; and the grayscale difference is used as the grayscale value of the pixel at corresponding position in the differential image corresponding to the two adjacent surveillance images.
