Intelligent basket net type sewage screen monitoring platform

Abstract

The utility model discloses an intelligent basket net type sewage screen monitoring platform which comprises a containing groove area, a filtering device, an image obtaining device, light source equipment, local control equipment and a remote cloud platform, under the condition that extra illumination is provided through the light source equipment, image obtaining is carried out on the filtering device, and real-time images of the filtering device can be captured; and by combining with computer vision and an artificial intelligence algorithm, the loading condition of the filtering device is analyzed, working data are synchronized to a remote cloud platform, and when the load is close to full load, local control equipment triggers an alarm, so that an operator can arrange cleaning work as required.

Description

Technical Field

[0001] The utility model relates to the field of environmental protection, and more specifically, to an intelligent basket-net type sewage screening monitoring platform. Background Art

[0002] As the core component of the sewage pump station, the basket filter plays a key role. It is responsible for intercepting and removing suspended and floating objects in the water, ensuring that key facilities such as pumps and valves are not blocked or damaged, thereby maintaining the smooth operation of the entire sewage system.

[0003] However, the sewage pump room often receives sewage mixed with oil, which, once deposited, can easily cause the basket filter to become clogged. In the absence of real-time camera monitoring, it is difficult to accurately determine when the filter needs to be cleaned, and the frequency of cleaning is also extremely unstable.

[0004] The traditional solution is to rely on manual regular cleaning of the screen. However, this method has many disadvantages. First, it is quite passive. Regardless of whether the screen really needs to be cleaned, it must be done according to the established plan, which undoubtedly increases unnecessary costs. Secondly, during the manual handling and cleaning process, workers will inevitably be exposed to harmful gases and unpleasant odors such as hydrogen sulfide emitted by the filter, posing a threat to their health. The most important thing is that this method cannot detect the sudden blockage of the filter in time. Once the blockage is serious, it is very likely to cause sewage overflow and cause greater environmental problems.

[0005] Therefore, there is an urgent need for a more intelligent and efficient solution to replace the traditional manual cleaning method. Utility Model Content

[0006] The present invention aims to at least solve the above-mentioned technical problems. To this end, the present invention proposes an intelligent basket-type sewage screening monitoring platform, which integrates a camera with a floodlight application environment to capture real-time images of the basket screen and combines it with computer vision and artificial intelligence algorithms to analyze the condition of the basket screen. When the screen is close to full load, an alarm is triggered, allowing the operator to arrange cleaning work as needed.

[0007] According to the first aspect of the present invention, an intelligent basket-net type sewage screening monitoring platform includes: a accommodating tank area, which is located below the ground to form an accommodating cavity, and a sewage inlet is provided on a side wall of the accommodating cavity (pumping well);

[0008] A filter device, which is located at the sewage inlet and filters the sewage flowing into the accommodating chamber; the filter device can move in the vertical direction and can be lifted or lowered by a lifting rope according to installation or inspection needs;

[0009] An image acquisition device is arranged on the other side wall slide rail of the sewage inlet of the accommodating chamber, so that it can acquire images of the side or opposite direction of the screen; the image acquisition device can slide on the slide rail as needed to acquire images at different angles or move out of / into the accommodating chamber;

[0010] A light source device, which is installed on the inner wall of the accommodating cavity and can provide additional lighting for the entire accommodating cavity;

[0011] A local control device, which is installed on the ground. Preferably, the electric control can control the lifting and lowering of the screen rope and the power supply and control instructions given to the image acquisition device and the light source device; it sends the local working information to the remote receiving terminal through the communication module;

[0012] A remote cloud platform receives local working information acquired by a local control device.

[0013] Preferably, the accommodating groove area is in a closed state during operation, and is opened or closed only when the accommodating cavity operation is required.

[0014] Preferably, the top of the above-mentioned filtering device is fixedly connected to the lifting rope, and the rope is driven by external force (manually or by a motor). The filtering device is in the shape of a three-dimensional grid, but one of its outer walls is open and aligned with the sewage inlet, so that sewage can flow in from the open side and flow out from the other outer walls of the grid, and is used to filter large pieces of suspended or floating objects, such as plastic bags and submerged box-shaped objects, and even foam blocks.

[0015] Preferably, the image acquisition device can be a camera or a video camera, and a moisture-proof shield is provided on the outside for protection. The image acquisition device is installed on a slide rail, and the slide rail can be set in a vertical direction or a horizontal direction, or both, so that the image acquisition device can acquire pictures in different orientations in the accommodating cavity as needed. The image acquisition device takes images or videos regularly (or in real time) to capture the instantaneous state of the screen for artificial intelligence video analysis and processing. When it is not necessary to acquire images, there will be a protective cover to protect the lens and light source equipment of the image acquisition device from slight chemical erosion.

[0016] Preferably, the above-mentioned light source device can be a fixed light source or a movable light source. If a fixed light source is selected, it means that it can be detachably installed on the inner wall of the accommodating cavity, and the installation point can be switched as needed. If a movable light source is selected, it can be moved in real time in the accommodating cavity. In this case, when using this method, it is necessary to set up vertical and / or horizontal slide rails for the light source device so that it can move in a linear direction or on a surface. The light source device is preferably a floodlight, and its illumination brightness can be adjusted by a local control device.

[0017] The local control device is located on the ground near the accommodating tank area, and a transparent cover is provided on its control panel for protection. Preferably, the control device can control the rope motor for lifting the screen, so that the screen can be raised and lowered, which is convenient for the installation and inspection of the screen. The control device is also provided with a conventional power switch and a circuit indicator light, and in particular, a display is provided, which can obtain the image captured by the image acquisition device. Preferably, the control device can control the movement of the image acquisition device and the light source device by button operation on its panel.

[0018] The local control device can record local operation records, especially the local control panel to issue an alarm (including but not limited to issuing an audible alarm or a visual alarm, such as a beep or flashing lights or vibration) for whether the screen is fully loaded or blocked, and also send the alarm information to several (customizable selected quantity and objects, preferably five objects) terminal recipients by SMS or email through the communication module. The selected object will have the right to confirm or reject the alarm detection result in order to carry out cyclic artificial intelligence training as an event source in the learning library. This can train the "reward model" directly from human feedback, and use the model as a reward function, and use reinforcement learning to optimize the agent strategy through optimization algorithms such as proximal strategy optimization. The reward model is pre-trained before the optimization strategy to predict whether a given output is good (high reward) or bad (low reward). Reinforcement learning from human feedback can improve the robustness and exploratory nature of reinforcement learning, especially when the reward function is sparse or noisy.

[0019] The above-mentioned working information refers to the lifting and lowering displacement information of the screen, the displacement information of the image acquisition device, the movement information of the light source equipment (if it is a movable light source), the image recognition screen and the alarm prompt information, etc.

[0020] Preferably, the cloud platform is a remote server, which is used to record the time of event occurrence and specific event trigger codes in real time, as well as remotely record real-time image screens, etc.

[0021] According to an embodiment of the utility model, an intelligent basket-net type sewage screening monitoring platform has at least the following features:

[0022] Beneficial effects:

[0023] 1. Real-time monitoring and timely response: Through real-time image detection, the sewage filtration process can be immediately understood. Once problems or abnormal situations occur, they can be quickly discovered and corresponding response measures can be taken to ensure the stability and efficiency of the sewage filtration process.

[0024] 2. Improve treatment efficiency: Real-time monitoring can assist operators in optimizing treatment processes, adjusting equipment operating parameters according to real-time data, and arranging staff to carry out corresponding cleaning treatments in real time to ensure the efficiency and stability of the sewage treatment process.

[0025] 3. Data recording and analysis: Real-time monitoring can continuously record various data, provide a basis for subsequent data analysis and processing, help to discover potential problems and improve the processing process.

[0026] 4. Strengthen supervision and law enforcement: For environmental protection departments, real-time monitoring can help them better supervise the operation of sewage treatment plants, ensure that companies strictly abide by environmental protection laws and regulations, and provide strong evidence for law enforcement.

[0027] 5. Improve public participation and environmental awareness: Real-time monitoring data and information can be screened and opened to the public through the network platform as needed, increasing public participation, improving public environmental awareness, and promoting sustainable development of society.

[0028] Additional aspects and advantages of the present invention will be given in part in the following description, and in part will become apparent from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0030] Figure 1 It is a schematic diagram of the overall structure of an embodiment of the utility model;

[0031] Figure 2 The working principle flow chart of the embodiment of the utility model;

[0032] Figure 3 A cloud dashboard for data logging and remote monitoring of a cloud platform in an embodiment of the present utility model;

[0033] 1-accommodating tank area, 2-filtering device, 3-image acquisition device, 4-light source device, 5-local control device, 6-remote cloud platform, 7-slide rail, 8-rope, 9-motor, 10-sewage inlet. DETAILED DESCRIPTION

[0034] The embodiments of the present invention are described in detail below. Figure 1-2 , an intelligent basket-net type sewage screening monitoring platform, comprising:

[0035] The receiving tank area 1 is located below the ground to form a receiving chamber, and a sewage inlet 10 is provided on one side wall of the receiving chamber (pumping well);

[0036] The filter device 2 is located at the sewage inlet 10 and filters the sewage flowing into the accommodating chamber; the filter device can move in the vertical direction and can be lifted or lowered by the lifting rope 8 according to the installation or inspection needs; preferably, the filter device 2 is made into a screen form.

[0037] An image acquisition device 3, which is arranged on the other side wall slide rail 7 of the sewage inlet 10 of the accommodating chamber, so that it can acquire images of the side or the opposite direction of the filtering device 2; the image acquisition device can slide on the slide rail 7 as needed to acquire images at different angles or move out of / into the accommodating chamber;

[0038] A light source device 4, which is mounted on the inner wall of the accommodating cavity and can provide additional lighting for the entire accommodating cavity;

[0039] A local control device 5, which is installed on the ground and can electrically control the lifting and lowering of the screen rope and provide power and control instructions to the image acquisition device and the light source device, and send local working information to the remote receiving terminal through the communication module;

[0040] The remote cloud platform 6 receives the local working information acquired by the local control device.

[0041] In other embodiments of the present solution, the above-mentioned accommodating groove area 1 is in a closed state during operation, and is opened or closed only when the accommodating cavity operation is required.

[0042] In other embodiments of the present scheme, the top of the above-mentioned filter device is fixedly connected to the lifting rope, the rope 8 is driven by an external force (motor 9 or manual drive), the filter device 2 is a three-dimensional grid shape, but one of its outer walls is open and aligned with the sewage inlet 10, so that sewage can flow in from the open side and flow out from the other outer walls of the grid, which is used to filter large pieces of suspended or floating matter, such as plastic bags and water-soaked box-shaped objects, and even foam blocks.

[0043] In other embodiments of the present scheme, the image acquisition device 3 can be a camera or a video camera, and a moisture-proof (physical moisture-proof or chemical moisture-proof) shield is provided on the outside for protection. The image acquisition device is installed on a slide rail 7, and the slide rail 7 can be set to a vertical direction or a horizontal direction, or both, so that the image acquisition device 3 can acquire pictures in different orientations in the accommodating cavity as needed. The image acquisition device 3 regularly takes images or videos (which can be connected to the local control device by wire or wireless electrical signals) to capture the instantaneous state of the filtering device 2 for artificial intelligence video analysis and processing. When it is not necessary to acquire an image, there will be a protective cover to protect the lens and light source equipment of the image acquisition device from slight chemical erosion.

[0044] Preferably, the camera is suitable for explosive gas and dust environments. For explosive gas environments, the maximum surface temperature is 85°C. For flammable dust environments, the maximum surface temperature is 80°C. The camera adopts stainless steel anti-corrosion spray treatment structure, or other anti-corrosion materials approved by the engineer to prevent damage caused by sparks or corrosive substances. The camera is suitable for outdoor use and the protection level is at least IP67.

[0045] In other embodiments of the present scheme, the above-mentioned light source device 4 can be a fixed light source or a movable light source. If a fixed light source is selected, it means that it can be detachably installed on the inner wall of the accommodating cavity, and the installation point can be switched as needed. If a movable light source is selected, it can be moved in real time in the accommodating cavity. In this case, when using this method, it is necessary to set up vertical and / or horizontal slide rails (not shown in the figure) for the light source device so that it can move in a linear direction or on a surface. The light source device is preferably a floodlight, and its illumination brightness can be adjusted by a local control device.

[0046] In other embodiments of the present scheme, the local control device 5 is located on the ground near the accommodating tank area 1, and a transparent cover is provided on its control panel for protection. The control device can select to control the rope motor for lifting the screen, so that the filter device can be raised and lowered, which is convenient for the installation and inspection of the screen. The control device is also provided with a control panel, which has a conventional power switch and a circuit indicator light, and in particular, is provided with a display, which can obtain the image captured by the image acquisition device. The control device can select the button operation on its panel to control the movement of the image acquisition device and the light source device.

[0047] The local control device 5 can record the image signal obtained by the image acquisition device into a local video or image record and operation record through an image signal processor, and analyze the condition of the basket screen by combining computer vision and artificial intelligence algorithms through a computer, especially the local control panel issues an alarm (including but not limited to issuing a sound alarm or a visual alarm, such as a buzzer or flashing lights or vibrations) for whether the screen is full or blocked, and also transmits the alarm information and the local network video or image recording data to the remote cloud platform after encryption through the communication module, and sends it to a number of (customizable selected number and objects, preferably five objects) terminal recipients in the form of text messages or emails. The local artificial intelligence processor or network video recorder includes a SIM card for wireless network connection, which can send detection information, images and at least 5 seconds of video to the remote cloud platform 6.

[0048] The artificial intelligence video analysis in the utility model is based on comparative analysis in an offline state, which is divided into initial state learning (initial sample photos) and on-site learning (the screen is fully loaded, blocked by large foreign objects, camera field of view obstacles, etc.), that is, the images of the screen under different filling states are collected, and the images of the screen in full and blocked conditions are compiled into video records and image training data sets, and then the real-time captured screen images are compared with the data sets.

[0049] For model training, the first step is to obtain the data set, and then process the data including filtering, enhancement and labeling, and then convert it into AI deep learning computing resources. During the training process, the validation data set will be used for verification; after the training is completed, manual inspection will be carried out to conduct preliminary verification of the training. In order to obtain satisfactory results, several rounds of training and verification are required, and the image and video data sets are updated during this process.

[0050] Selected end recipients will have the power to confirm or reject alert detections for recurrent AI training as event sources in the learning library. This allows training a “reward model” directly from human feedback and using that model as a reward function to optimize the agent policy using reinforcement learning through optimization algorithms such as proximal policy optimization. The reward model is pre-trained before optimizing the policy to predict whether a given output is good (high reward) or bad (low reward). Reinforcement learning from human feedback can improve the robustness of reinforcement learning, especially when the reward function is sparse or noisy.

[0051] The working information mentioned in the embodiment refers to the lifting and displacement information of the screen, the displacement information of the image acquisition device, the movement information of the light source equipment (if it is a movable light source), the image recognition screen and the alarm prompt information, etc.

[0052] In other embodiments of this solution, see the attached Figure 3 The remote cloud platform 6 has an image database for data storage, and a cloud dashboard for data recording and displaying remote monitoring, real-time recording of the time of event occurrence and specific event trigger codes, as well as remote recording of real-time image screens, etc.

[0053] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. An intelligent basket-type sewage screening monitoring platform, characterized in that: It includes: A receiving tank area (1) is located below the ground to form a receiving cavity, and a sewage inlet (10) is provided on a side wall of the receiving cavity; A filter device (2), the filter device (2) being located at the sewage inlet (10) and filtering the sewage flowing into the accommodating chamber; the filter device (2) being movable in the vertical direction and being able to be lifted or lowered by a lifting rope (8) according to installation or inspection requirements; An image acquisition device (3), the image acquisition device (3) being arranged on the other side wall slide rail (7) of the sewage inlet (10) of the accommodating chamber, so that it can acquire a picture image of the side or the opposite direction of the filtering device (2); the image acquisition device (3) can slide on the slide rail (7) as required; A light source device (4), wherein the light source device (4) is mounted on the inner wall of the accommodating cavity; A local control device (5), the local control device (5) is installed on the ground and sends local working information to a remote receiving terminal through a communication module; A remote cloud platform (6), wherein the remote cloud platform receives local working information acquired by the local control device (5).

2. The intelligent basket-type sewage screening monitoring platform according to claim 1 is characterized by: The top of the filter device (2) is fixedly connected to a lifting rope (8), and the rope (8) is driven by an external force. The filter device (2) is in the shape of a three-dimensional grid, and has an outer wall that is open and aligned with the sewage inlet (10).

3. The intelligent basket-type sewage screening monitoring platform according to claim 2 is characterized by: The image acquisition device (3) is a camera or a video camera, and is provided with a moisture-proof shield on the outside for protection. The image acquisition device (3) is installed on a slide rail (7), and the slide rail (7) is arranged in a vertical and / or horizontal direction.

4. The intelligent basket-type sewage screening monitoring platform according to claim 3 is characterized by: The light source device (4) is a fixed light source or a movable light source. If a fixed light source is selected, it means that it can be detachably installed on the inner wall of the accommodating cavity, and the installation point can be switched as needed. If a movable light source is selected, it can be moved in real time in the accommodating cavity. It is necessary to set up vertical and / or horizontal slide rails on the light source device (4) so ​​that it can move in a linear direction or on a surface.

5. The intelligent basket-net type sewage screening monitoring platform according to claim 4 is characterized by: The local control device (5) is located on the ground near the accommodating tank area (1), and a transparent cover is provided on its control panel for protection. The control device is also provided with a conventional power switch and a circuit indicator light, and a display is provided. The display can obtain the image captured by the image acquisition device (3).

6. The intelligent basket-type sewage screening monitoring platform according to claim 5 is characterized by: The local control device (5) is capable of recording local video or image records, issuing an alarm on the local control panel when the filtering device (2) is fully loaded or blocked, and also sending the alarm information to several terminal recipients in the form of text messages or emails through the communication module.

7. The intelligent basket-type sewage screening monitoring platform according to claim 4 is characterized by: The light source device (4) is a floodlight.

8. The intelligent basket-type sewage screening monitoring platform according to claim 6 is characterized by: The working information includes image recognition screen and alarm prompt information.

9. The intelligent basket-net type sewage screening monitoring platform according to claim 8 is characterized by: The remote cloud platform (6) is a remote server, which is used to record the time of event occurrence and the specific event trigger code in real time, and remotely record real-time image screens.

10. The intelligent basket-type sewage screening monitoring platform according to claim 9 is characterized in that: The analysis after the image recognition is to use computer vision and artificial intelligence algorithms to analyze and determine whether to issue an alarm.

Images