Control method and system of fog pile equipment

Abstract

The invention belongs to the technical field of dust fall, and relates to a control method and system for fog pile equipment, and the method comprises the steps: obtaining a road image shot by the fog pile equipment, and obtaining first operation intensity; obtaining local AQI data and a traffic flow trend, and obtaining second operation intensity; obtaining the single-time operation shortest duration, the single-time operation longest duration, the two-time operation interval shortest duration and the two-time operation interval longest duration of the fog pile equipment, and obtaining the single-time operation duration and the two-time operation interval duration of the fog pile equipment in combination with intermediate parameters of the first operation intensity, the second operation intensity and the operation time; a pressure response threshold value and a water drop atomization maximum threshold value of the fog pile equipment are obtained, and the water pressure of the fog pile equipment is obtained in combination with the intermediate parameter of the operation time and the single-time operation duration of the fog pile equipment; and the fog pile equipment is controlled to operate according to the single-time operation duration, the two-time operation interval duration and the water pressure. According to the invention, the efficiency and effect of dust control can be greatly improved.

Description

Technical Field

[0001] The present application relates to the field of dust suppression technology, and in particular to a control method and system for fog pile equipment. Background Art

[0002] In the field of road dust control, the use of mist piles that spray water mist is an important method. The mist pile equipment is deployed in the middle or side of the road, spraying water on the road where vehicles pass by to reduce dust. This method has been proven to be an effective method.

[0003] Existing fog pile systems generally have automated control capabilities, but they require manual control. Some systems offer remote control capabilities, but these are often limited to manual control and timed and quantitative control. Some fog pile systems that are linked to dust monitors are limited by wind direction and their ability to only detect specific areas, making them ineffective at addressing source control. Summary of the Invention

[0004] Based on this, it is necessary to provide a control method and system for fog pile equipment to address the above technical problems. The system can set the corresponding single operation time, the interval between two operations and the water pressure according to the road image, local AQI data and traffic flow trends to control the operation of the fog pile equipment, thereby greatly improving the efficiency and effect of dust control, and significantly reducing power consumption and water consumption.

[0005] A method for controlling a fog pile device, comprising: Obtaining a road image captured by the fog pile device and processing the road image to obtain vehicle type, number of vehicles, dust cloud size, and cumulative dust generation time; obtaining a first operating intensity of the fog pile device based on the vehicle type, number of vehicles, dust cloud size, and cumulative dust generation time; Obtain local AQI data and traffic flow trends, calculate dust fall intensity, and obtain the second operating intensity of the fog pile equipment; According to the first operating intensity and the second operating intensity, an intermediate parameter of the operating time is obtained; the shortest single operating time and the longest single operating time of the fog pile device are obtained, and the single operating time of the fog pile device is obtained by combining the intermediate parameter of the operating time; the shortest interval between two operations of the fog pile device and the longest interval between two operations are obtained, and the interval between two operations of the fog pile device is obtained by combining the intermediate parameter of the operating time and the single operating time of the fog pile device; the pressure response threshold value and the maximum water droplet atomization threshold value of the fog pile device are obtained, and the water pressure of the fog pile device is obtained by combining the intermediate parameter of the operating time and the single operating time of the fog pile device; The operation of the fog pile equipment is controlled according to the duration of a single operation, the interval between two operations and the water pressure.

[0006] In one embodiment, obtaining an intermediate parameter of the running time according to the first running intensity and the second running intensity includes:

[0007] Where, is the intermediate parameter of running time, The shortest duration of a single run. is the first running intensity, is the second running intensity; Obtain the shortest and longest single-running durations of the fog pile device, and combine them with the intermediate parameters of the running time to obtain the single-running duration of the fog pile device, including:

[0008] Where, is the duration of a single run, The shortest duration of a single run. The maximum duration of a single run.

[0009] In one embodiment, the shortest duration and the longest duration between two running times of the fog pile device are obtained, and the duration between two running times of the fog pile device is obtained by combining the intermediate parameters of the running time and the single running time of the fog pile device, including:

[0010] Where, is the time interval between two runs, The shortest time between two runs. The maximum time interval between two runs.

[0011] In one embodiment, the pressure response threshold and the maximum water droplet atomization threshold of the fog pile device are obtained, and the water pressure of the fog pile device is obtained by combining the intermediate parameters of the running time and the single running time of the fog pile device, including:

[0012] Where, The water pressure of the fog pile equipment, is the maximum threshold of water drop atomization, is the pressure response threshold.

[0013] In one embodiment, the first operation intensity of the fog pile device is obtained according to the vehicle type, the number of vehicles, the size of the dust cloud, and the cumulative dust generation time, including: The dust intensity within the identified range is obtained based on the vehicle type, number of vehicles, dust cluster size, and cumulative dust generation time:

[0014] Where, is the dust intensity, It is the integral obtained based on the vehicle type, number of vehicles, dust cloud size and cumulative dust generation time; According to the dust intensity, the first operating intensity of the fog pile equipment is obtained:

[0015] Where, This is the first operating intensity.

[0016] In one embodiment, local AQI data and traffic flow trends are obtained, and dust fall intensity is calculated to obtain a second operating intensity of the fog pile device, including: Obtain local AQI data and traffic flow trends, and calculate dust fall intensity:

[0017] Where, is the dust fall intensity, For local AQI data, For traffic flow trends, Minimum traffic volume; According to the dust fall intensity, the second operating intensity of the fog pile equipment is obtained:

[0018] Where, This is the second running intensity.

[0019] In one embodiment, a road image captured by a fog pile device is obtained and processed to obtain vehicle type, number of vehicles, dust cloud size, and accumulated dust generation time, including: Obtain the road image captured by the fog pile equipment, define the recognition range in the road image, and use the regional coordinate mapping algorithm to convert the recognition range into the image coordinate system to obtain the detection range; Using real-time target detection algorithm, target detection is performed within the detection range to obtain the vehicle position and dust cloud position; Using classification algorithm, the vehicle position is identified and the vehicle type is obtained; Using target tracking and counting algorithms, different vehicle types are counted to obtain the number of vehicles corresponding to the vehicle type; According to the position of the dust cluster, the size of the dust cluster and the cumulative time of dust generation are obtained.

[0020] In one embodiment, after obtaining the vehicle type, the number of vehicles, the dust mass size, and the accumulated dust generation time, the method further includes: Add dynamic occlusion to the training set images and perform occlusion enhancement training to suppress background interference in the image; Dynamic lighting is added to the training set images for visual enhancement training to improve image transmittance.

[0021] A control system for a fog pile device adopts a control method for the fog pile device, comprising: a pile, a nozzle assembly, a camera, an Internet of Things component, and a main control cabinet; wherein the nozzle assembly and the camera are both arranged on the pile; The nozzle assembly is connected to the main control cabinet and is used to perform fixed-point dust reduction on the dust-generating location according to the spray signal of the main control cabinet; The camera is connected to the main control cabinet and is used to obtain road conditions, process the road conditions, obtain road information, and send the road information to the main control cabinet so that the main control cabinet generates a spray signal; The Internet of Things component is connected to the main control cabinet, and is used to obtain environmental conditions, process the environmental conditions, obtain environmental information, and send the environmental information to the main control cabinet so that the main control cabinet generates a spray signal; The main control cabinet includes: a controller, a frequency converter and a water pump; the controller is used to receive road information from the camera and environmental information of the Internet of Things component, and generate a spray signal based on the road information and the environmental information to control the water spraying action of the sprinkler assembly; the spray signal includes: the duration of a single operation, the duration of the interval between two operations and the water pressure; the frequency converter is connected to the controller and the water pump, and is used to control the water spraying volume by adjusting the power frequency of the water pump according to the spray signal of the controller, and set the water pressure threshold to achieve water and energy saving while meeting the water spraying distance and atomization capacity.

[0022] In one embodiment, the nozzle assembly includes: a servo motor, a transmission mechanism, a rotary nozzle, a sensing mechanism, and a calibration mechanism; The servo motor is connected to the rotary nozzle through the transmission mechanism, so as to control the rotary nozzle through the transmission mechanism according to the spray signal of the controller to perform fixed-point dust reduction at the dust-generating location; the rotary nozzle has a 360-degree motion range; The sensing mechanism is connected to the controller and is spaced apart from the rotary nozzle so as to send a trigger signal to the controller after sensing the position probe of the rotary nozzle; the controller receives the trigger signal of the sensing mechanism and generates a calibration signal and sends it to the calibration mechanism; the calibration mechanism is connected to the controller and the rotary nozzle so as to calibrate the position of the rotary nozzle according to the calibration signal of the controller.

[0023] The control method and system of the above-mentioned fog pile equipment can set the corresponding single operation time, two-operation interval time and water pressure according to road images, local AQI data and traffic flow trends to control the operation of the fog pile equipment. That is, it integrates road conditions, traffic conditions and weather conditions to effectively control dust in real time, greatly improves the efficiency and effectiveness of dust control, and significantly reduces power consumption and water consumption. It also integrates a variety of new technologies to empower fog piles, turning the fog pile equipment into an intelligent terminal, greatly expanding its application scenarios, and greatly expanding the application boundaries of fog piles, becoming an application node in multiple links such as environmental protection, municipal administration, transportation, and urban management. While improving the air quality of cities and parks, it creates other value, reduces the management burden on collaborative management departments to a certain extent, and reduces the overall cost of urban management. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 1 is a flow chart of a method for controlling a fog pile device according to an embodiment; Figure 2 A schematic diagram of the overall structure of a control system of a fog pile device in one embodiment; Figure 3 A schematic structural diagram of a nozzle assembly of a control system of a mist pile device according to one embodiment; Figure 4 is a diagram showing the relationship between water pressure and inverter output frequency in one embodiment; Figure 5 The figure is a schematic diagram of the architecture of a control system of a fog pile device in one embodiment.

[0025] Description of reference numerals: 1. Stake erection; 2 nozzle assembly; servo motor 21, calibration mechanism 22; 3 cameras; 4 IoT components; 5 main control cabinet. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for explaining this application and are not intended to limit this application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in this application without creative work are within the scope of protection of this application.

[0027] It should be noted that all directional indications in the embodiments of the present application (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0028] In addition, the terms "first," "second," and so on, used in this application are for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, features specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "multiple groups" means at least two groups, such as two groups, three groups, and so on, unless otherwise specifically defined.

[0029] In this application, unless otherwise specified or limited, the terms "connect," "fix," etc. should be understood in a broad sense. For example, "fix" can mean a fixed connection, a detachable connection, or an integral connection; it can mean a mechanical connection, an electrical connection, a physical connection, or a wireless communication connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean internal communication between two elements or an interaction between two elements, unless otherwise specified. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0030] In addition, the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in this field can implement it. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0031] This application provides a control method for a fog pile device, such as Figure 1 As shown, in one embodiment, it includes: Step 101: Acquire a road image captured by a fog pile device and process the road image to obtain vehicle type, number of vehicles, dust ball size, and cumulative dust generation time; obtain a first operating intensity of the fog pile device based on the vehicle type, number of vehicles, dust ball size, and cumulative dust generation time.

[0032] Specifically: Obtain the road image captured by the fog pile equipment, define the recognition range in the road image, and use the regional coordinate mapping algorithm to convert the recognition range into the image coordinate system to obtain the detection range; Perform target detection in the detection range to obtain the vehicle position and dust ball position; According to the vehicle location, obtain the vehicle type and number of vehicles; According to the position of the dust cluster, the size of the dust cluster and the cumulative time of dust generation are obtained; The first operation intensity of the fog pile equipment is obtained according to the vehicle type, the number of vehicles, the size of the dust cluster and the cumulative time of dust generation.

[0033] More specifically: Obtain road images captured by fog pile equipment, delineate polygonal or rectangular areas in the road images as the recognition range, and use a regional coordinate mapping algorithm to convert the recognition range into an image coordinate system. Use a spatial geometric transformation algorithm to match the virtual fence (i.e., the recognition range) with the image pixels to obtain the detection range. A real-time target detection algorithm is used to detect targets within the detection range, and the positions are marked by bounding boxes to obtain the vehicle position and dust cloud position. Using classification algorithms, the vehicle position is identified and the vehicle type (such as car, truck, bus) is obtained; Using target tracking and counting algorithms, different vehicle types are counted to obtain the number of vehicles corresponding to the vehicle type; According to the position of the dust cluster, the size of the dust cluster and the cumulative time of dust generation are obtained; The dust intensity within the identified range is obtained based on the vehicle type, number of vehicles, dust cluster size, and cumulative dust generation time:

[0034] Where, is the dust intensity, It is the integral obtained based on the vehicle type, number of vehicles, dust cloud size and cumulative dust generation time; According to the dust intensity, the first operating intensity of the fog pile equipment is obtained:

[0035] Where, It is the first operation intensity, which mainly affects the duration of a single operation (T) and the amount of spray.

[0036] In this step, after obtaining the vehicle type, number of vehicles, dust group size and cumulative dust generation time, it also includes: adding dynamic occlusion to the training set images and performing occlusion enhancement training to suppress background interference in the image; adding dynamic lighting to the training set images and performing visual enhancement training to improve image transmittance.

[0037] Note: Dust grade points are assigned to identified targets, such as 1 for dump trucks, 0.6 for box trucks, 0.4 for buses / coaches, 0.2 for cars, and 0.03 for small motor vehicles. Dust grade points are assigned from 5 to 20 based on the cumulative time of dust generation and the size of dust clusters.

[0038] For example: Y ai= Every time the level score is greater than 20, the dust intensity is accumulated every 5 seconds, that is: Y ai =Y ai +1; if the grade score is less than 2 for 5 consecutive seconds, the dust intensity will be reduced by one, that is: Y ai =Y ai -1.

[0039] I ai (Running Intensity) = , within a unit of time (10 minutes), each time Y appears ai >10 minutes / 5 seconds, accumulate the treatment intensity once, i.e.: I ai =I ai +1; Level points Y for every 5 seconds ai <12 minutes / 5 seconds, the running intensity is reduced once, that is: I ai =I ai -1.

[0040] Step 102: Obtain local AQI data and traffic flow trends, calculate dust fall intensity, and obtain a second operating intensity of the fog pile device.

[0041] Specifically: Obtain local AQI data and traffic flow trends, and calculate dust fall intensity:

[0042] Where, is the dust fall intensity, For local AQI data, For traffic flow trends, is the minimum traffic volume, which is a fixed constant; According to the dust fall intensity, the second operating intensity of the fog pile equipment is obtained:

[0043] Where, This is the second running intensity.

[0044] In this step, the local AQI data includes: local AQI index, local real-time rainfall intensity, rainfall trend, air humidity, and ambient temperature.

[0045] Step 103: obtain the intermediate parameters of the operating time according to the first operating intensity and the second operating intensity; obtain the shortest single operating time and the longest single operating time of the fog pile device, and obtain the single operating time of the fog pile device in combination with the intermediate parameters of the operating time; obtain the shortest time interval between two operations of the fog pile device and the longest time interval between two operations, and obtain the time interval between two operations of the fog pile device in combination with the intermediate parameters of the operating time and the single operating time of the fog pile device; obtain the pressure response threshold value and the maximum water droplet atomization threshold value of the fog pile device, and obtain the water pressure of the fog pile device in combination with the intermediate parameters of the operating time and the single operating time of the fog pile device.

[0046] Specifically: According to the first running intensity and the second running intensity, the intermediate parameter of the running time is obtained: ; Where, is the intermediate parameter of running time, The shortest duration of a single run. is the first running intensity, is the second running intensity; Get the shortest and longest single-run durations of the fog pile device, and combine them with the intermediate parameters of the running time to get the single-run duration of the fog pile device:

[0047] Where, is the duration of a single run, The shortest duration of a single run (the initial setting constant), The maximum duration of a single run (the initial setting constant); Get the shortest and longest intervals between two runs of the fog pile device, and combine the intermediate parameters of the run time and the single run time of the fog pile device to get the interval between two runs of the fog pile device:

[0048] Where, is the time interval between two runs, is the shortest time between two runs (it is the initial setting constant), The longest time interval between two runs (the initial setting constant), for multiplication; Obtain the pressure response threshold and the maximum threshold of water droplet atomization of the fog pile device. Combined with the intermediate parameters of the running time and the single running time of the fog pile device, the water pressure of the fog pile device is obtained:

[0049] Where, The water pressure of the fog pile equipment, is the maximum threshold of water drop atomization, is the pressure response threshold.

[0050] In this step, the obtained single operation duration, the interval between two operations and the water pressure form a spray signal.

[0051] Step 104 , controlling the operation of the fog pile device according to the single operation duration, the interval between two operations, and the water pressure.

[0052] In this step, after obtaining the single operation time, the interval between two operations and the water pressure, automatic time comparison, operation frequency output, and injection position and area calculation are performed, thereby realizing the time control, start and stop control, and opening water pressure control of the fog pile equipment; and then combined with the positioning cruise function, the operation of the fog pile equipment is controlled.

[0053] In this embodiment, road images captured by the fog pile device are acquired and processed to determine vehicle type, number of vehicles, dust cloud size, and accumulated dust generation time. This can be accomplished by either a camera or a main control cabinet. Local AQI data and traffic flow trends can be acquired using IoT components or a main control cabinet. The main control cabinet also calculates the first and second operating intensities, intermediate parameters, single operating duration, interval between two operations, water pressure, and controls the operation of the fog pile device.

[0054] The control method of the above-mentioned fog pile equipment can set the corresponding single operation time, the interval between two operations and the water pressure according to the road image, local AQI data and traffic flow trends to control the operation of the fog pile equipment. That is, it comprehensively integrates road conditions, traffic conditions and weather conditions to carry out real-time and effective dust control, realizes dynamic response to dust reduction needs, and lays the basic conditions for implementing targeted control in the set area, greatly improves the dust reduction capacity, improves the efficiency and effectiveness of dust control, and significantly reduces electricity consumption and water consumption.

[0055] This application also provides a fog pile device, such as Figure 2 and Figure 3 As shown, in one embodiment, it includes: a stake, a sprinkler assembly, a camera, an Internet of Things assembly and a main control cabinet.

[0056] Among them, the nozzle assembly and the camera are arranged on the upper part of the pile, and the Internet of Things assembly and the main control cabinet are arranged on the lower part of the pile, and can also be arranged at intervals from the pile.

[0057] 1. Stake The pile is the main body of the fog pile equipment and provides a bearing position for other components.

[0058] 2. Nozzle assembly The nozzle assembly is connected to the main control cabinet and is used to perform targeted spraying to reduce dust at dust-generating locations according to the spray signal from the main control cabinet.

[0059] The nozzle assembly includes: a servo motor, a transmission mechanism, a rotating nozzle, a sensing mechanism and a calibration mechanism.

[0060] The servo motor is connected to the rotary nozzle through a transmission mechanism, and controls the rotary nozzle through the transmission mechanism according to the spray signal of the controller, so as to carry out targeted dust reduction at the dust-producing location; the servo motor contains a position closed-loop function, and has the ability to hit wherever it is pointed, so that together with the sensing mechanism and calibration mechanism, it can achieve a full closed-loop to ensure long-term precise control of the rotation position of the rotary nozzle.

[0061] The transmission mechanism is connected to the rotary sprinkler head to control the movement of the rotary sprinkler head under the instruction of the servo motor.

[0062] The rotating nozzle has a 360-degree motion range and can be directed to the dust-generating location or area, achieving fixed-point dust reduction and positioning cruise function at any angle of 360 degrees (360°).

[0063] The sensing mechanism is connected to the controller and is spaced apart from the rotating nozzle so as to send a trigger signal to the controller after sensing the position probe of the rotating nozzle. The specific sensing mechanism may be a magnetic sensing mechanism.

[0064] The calibration mechanism is connected to the controller and the rotary nozzle to calibrate the position of the rotary nozzle according to the calibration signal of the controller. Specifically, the calibration mechanism controls the unidirectional rotation of the rotary nozzle according to the calibration signal of the controller and sends a completion signal to the controller.

[0065] It should be noted that the setting of the sensing mechanism and the calibration mechanism can avoid the problem of position deviation caused by the gap of the transmission mechanism of the rotating nozzle during long-term operation, ensure the accuracy of the control angle during long-term operation, and accurately strike when visible dust occurs (visible dust has a higher priority than road vehicles), providing effective guarantee for the efficient control of dust by fog piles.

[0066] 3. Camera The camera is connected to the main control cabinet and is used to obtain road conditions, process the road conditions, obtain road information, and send the road information to the controller of the main control cabinet so that the controller of the main control cabinet generates a spray signal.

[0067] Specifically: Obtain a road image captured by the fog pile device, delineate a polygonal or rectangular area in the road image as the recognition range, and use a regional coordinate mapping algorithm to convert the recognition range into an image coordinate system. A spatial geometric transformation algorithm is used to match the virtual fence (i.e., the recognition range) with the pixels of the image to obtain the detection range. (For example, establish a correspondence between the 0° angle of the AI camera and the 0° angle of the fog pile nozzle. Delineate a visual recognition area in the image (multiple non-overlapping areas can be defined as needed). The recognition area is slightly larger than the coverage area of the fog pile spraying (approximately 15% larger). Convert the delineated area into an image coordinate system, and use a spatial set transformation algorithm to match the virtual fence with the pixels. Perform regional spraying by matching the target coordinate system with the spray angle of the fog pile nozzle. For example, define area 1, and the corresponding nozzle angle is -10°~20°.) A real-time target detection algorithm is used to detect targets within the detection range, and the positions are marked by bounding boxes to obtain the vehicle position and dust cloud position. Use classification algorithms to identify the vehicle type (such as car, truck, bus / coach, van, muck truck, small motor vehicle) based on the vehicle location. Using target tracking and counting algorithms, different vehicle types are counted to obtain the number of vehicles corresponding to the vehicle type; According to the position of the dust cluster, the size of the dust cluster and the cumulative time of dust generation are obtained; The road information is composed of vehicle type, number of vehicles, dust cluster size and accumulated dust generation time; The road information is sent to the controller of the main control cabinet so that the controller of the main control cabinet generates a spray signal.

[0068] Preferably, after obtaining the vehicle type, number of vehicles, dust group size and cumulative dust generation time, it also includes: adding dynamic occlusion to the training set images and performing occlusion enhancement training to suppress background interference in the image; adding dynamic lighting to the training set images and performing visual enhancement training to improve image transmittance.

[0069] Specifically: Add dynamic occlusion to the training set images and perform occlusion enhancement training to suppress background interference in the image, including: 1) A lightweight detection model is used with an input resolution adjusted to 640×384 (to accommodate edge devices): This model enhances feature expression capabilities by designing channel operations with a lightweight network structure while maintaining low computational complexity. This model simplifies multi-scale features, reduces the computational overhead of cross-layer connections, and lowers model complexity.

[0070] 2) Low-bit quantization: Compresses 32-bit floating-point weights to 8-bit integers (INT8), combines dynamic range calibration to reduce precision loss, and adapts to edge device deployment.

[0071] 3) Simulate complex environments during training: randomly add foliage occlusions (no more than 30% of the occlusion area in each image) and strong light reflection effects to road images.

[0072] 4) Collect 200 road images covering different lighting and occlusion conditions and use INT8 precision for calculation.

[0073] 5) Using a dynamic occlusion generator: This randomly overlays gray occlusion blocks ranging from 15×15 to 60×60 pixels onto 200 collected road photos covering different lighting and occlusion conditions, with an occlusion ratio of 20%-40%. Unlike conventional static occlusion recognition, dynamic occlusion recognition shifts from passive to active, significantly improving the accuracy of identifying obscured objects.

[0074] 6) Adversarial training strategy: Each batch consists of 50% real occluded samples (collected samples) and 50% generated samples (generated using the dynamic occlusion generator). The update amplitude of each discriminative training sample (50% real occluded samples + 50% generated samples) is 1 / 3.

[0075] This setting can dynamically focus on the features of the unobstructed areas of the target and suppress background interference; fuse the target's motion trajectory and appearance features, and compensate for the feature loss caused by occlusion through spatiotemporal information; randomly add occluders (such as rectangular blocks, natural object maps) in the data enhancement stage to simulate real occlusion scenes, greatly improve the recognition accuracy of occluded objects, and enhance the generalization ability of the model; identify and locate vehicles and dust clouds, and improve the accuracy of governance and scene adaptability; adopt a lightweight detection model and INT8 precision calculation method to effectively reduce the system's requirements for video accuracy and edge computing power, improve the computing power contradiction between the algorithm and the edge computing power, and effectively improve the system's adaptability.

[0076] Dynamic lighting is added to the training set images to perform visual enhancement training to improve image transmittance, including: 1) Dynamic Illumination Preprocessing: Combining visible light and infrared sensor data, cross-modal feature alignment enhances target outline information in low light or strong backlight conditions. Self-supervised learning is introduced to extract illumination-invariant features from unlabeled data, reducing the model's sensitivity to lighting conditions.

[0077] 2) Enhanced Bright and Dark Details with Automatic Contrast Adjustment: The image is divided into 8×8 grids. When the image is too dark (e.g., at night when lux is less than 20), the brightness of each grid is adjusted individually (similar to Photoshop's local dimming). The system automatically enhances dark details, similar to the night scene mode on mobile phones. Similarly, if an area in the image is too bright (e.g., with a brightness value greater than 50,000), the brightness of that area will be automatically reduced by 5%-15%.

[0078] 3) Environmental Adaptive Optimization: When the camera is facing the sun (e.g., during sunrise and sunset), data from both the regular camera and the infrared camera are simultaneously used: the regular camera captures color information, the infrared camera provides contour information, and a compensated image is generated through data fusion (similar to the three-image synthesis in the HDR mode on mobile phones). A mapping table is established between light intensity (lux) and model accuracy (i.e., the following mode): 50000 lux (strong light) → Enable HDR enhanced mode; 1000-50000 lux → standard detection mode; <1000lux → activate infrared fusion mode; 4) Learning Mechanism: A dual-mode database is established: During the day, it prioritizes learning the characteristics of strong natural light reflections, while at night, it focuses on learning the characteristics of vehicle and streetlight illumination. Automatic updates every 72 hours: Lighting processing parameters are fine-tuned based on 1,000 recently collected images of typical scenes. (Field tests have verified that the vehicle detection rate in nighttime, without streetlights, remains stable at over 92%, and the false alarm rate in backlit scenes has dropped to 4.7%.) 5) The actual application is divided into three steps: ‌Detecting dirt‌ (detecting dirt): The system compares the image clarity once a second, just like checking an eye chart; when it finds that key information such as vehicles has become blurred (for example, a car that was originally visible 50 meters away can now only be seen within 30 meters).

[0079] Smart Cleaning (Automatic Cleaning): Slightly dirty: Automatically enhance the image (similar to increasing the contrast of mobile phone photos); Severely blocked: Activate the AI camera's cleaning wipers to automatically clean the lens.

[0080] Dirt Compensation (temporary remedy): Before cleaning is completed: Use the previous 5 seconds of clean footage as a reference to intelligently fill in the current missing parts (similar to using PS to repair scratches on old photos).

[0081] Actual case: When a truck passes by and raises dust to block the camera lens: within 1 second, the image is detected to be blurry → within 2 seconds, the image is enhanced to ensure temporary monitoring → within 3 seconds, the AI camera self-cleaning function is activated → after 10 seconds, the image becomes clear again, without any human intervention.

[0082] This setting can improve the light transmittance of the camera image, avoiding the camera lens from getting dirty or the image from being unclear due to poor control quality after long-term use. It is like putting "smart glasses" on the camera. When the lens is blocked by dust, raindrops or mud spots, the corresponding output is executed: automatically detecting that the lens is dirty, enhancing the image recognition (enhancing contrast, etc., and "guessing" the blocked image through calculation without cleaning), and starting the automatic cleaning function of the AI camera when necessary (like the automatic wipers on a car).

[0083] It should be noted that the regional coordinate mapping algorithm, spatial geometric transformation algorithm, real-time target detection algorithm, classification algorithm, target tracking and counting algorithm are all existing technologies.

[0084] Further preferably, the functions of the camera can be expanded to achieve multifunctional applications, including: 1) The camera images are connected to the video surveillance system of the competent unit, which can increase the density of regional video surveillance at an ultra-low cost; 2) Additional algorithms can be embedded in the edge of AI cameras in various fog pile areas to enhance municipal / park management capabilities. These embedded algorithms target scenarios such as falls, car accidents, vehicles driving the wrong way, unusual intrusions, and security. 3) The data of traffic flow monitoring by fog piles can be reported to the traffic authorities in real time to actively improve the level of municipal management; 4) With self-cleaning function; 5) Edge computing is used to implement edge intelligent computing, and the specific algorithm belongs to the existing technology; 6) More functions can be integrated and derived according to needs.

[0085] It should also be noted that the camera is installed at a height of 4-5 meters and adjusted to a suitable angle to ensure that the vehicle at the farthest position to be analyzed can be distinguished in the picture.

[0086] 4. IoT components The Internet of Things component is connected to the controller of the main control cabinet, and is used to obtain environmental conditions, process the environmental conditions, obtain environmental information, and send the environmental information to the controller so that the controller generates a spray signal according to the road information and the environmental information.

[0087] Specifically: Based on historical data (including environmental parameters and dust reduction effect records), the machine learning model is trained and optimized to obtain an environmental model; Real-time environmental conditions are acquired through a distributed sensor network, and edge computing nodes are used to pre-process the environmental conditions to obtain environmental data. Data security is ensured through encrypted transmission protocols. Environmental data is used as input to the environmental model to obtain the output of the environmental model; the output of the environmental model includes: local AQI data and traffic flow trends; The output of the environmental model (i.e., local AQI data and traffic flow trends) is used as environmental information; The environmental information is sent to the controller so that the controller generates a spray signal according to the road information and the environmental information.

[0088] in, The machine learning model utilizes a hybrid machine learning framework, integrating a random forest (for feature importance analysis) and an LSTM neural network (for time series trend modeling). The input layer integrates multidimensional feature vectors (dimensionality ≥ 8), and the output layer generates fog pile control parameters (dust fall sensitivity, intensity, and frequency). During training and optimization, supervised learning is performed using historical data, and hyperparameters are optimized through cross-validation and grid search to enhance model generalization.

[0089] Environmental conditions include multi-dimensional environmental and traffic parameters, specifically local AQI data and traffic flow trends. Local AQI data includes the local AQI index, local real-time rainfall intensity, rainfall trends, air humidity, and ambient temperature. The local AQI index is an air quality index calculated based on six pollutants: sulfur dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO), ozone (O3), inhalable particulate matter (PM10), and fine particulate matter (PM2.5). Rainfall trends are linear / nonlinear predictions based on time series. Traffic flow trends are based on forecast data provided by traffic management departments, for example, predicting the possible traffic flow within the next hour.

[0090] Preprocessing includes data cleaning and feature fusion. Data cleaning involves removing sensor outliers (e.g., data outside the physical range), identifying and filling missing values using boxplots, and smoothing time series data (e.g., traffic flow trends) to eliminate noise. Feature fusion involves constructing composite feature vectors. For example, using "rainfall intensity × air humidity" as a derivative indicator of humidity saturation, and standardizing using the Z-score method to unify dimensions and ensure consistency in model inputs.

[0091] It should be noted that the IoT components may include rainfall detectors to obtain rainfall trends in real time, providing more complete data perception capabilities for the system's intelligent dust control capabilities. For example, when the rainfall is greater than 5 mm, dust will no longer fall due to environmental dust factors.

[0092] 5. Main control cabinet The main control cabinet is connected to the nozzle assembly, camera and Internet of Things assembly. It is used to receive road information from the camera assembly and environmental information from the Internet of Things assembly, generate a spray signal based on the road information and environmental information, and send the spray signal to the nozzle assembly to carry out targeted dust reduction at the dust-producing location and achieve water and energy saving.

[0093] The main control cabinet includes: controller, inverter and water pump.

[0094] The controller is used to receive road information from the camera component and environmental information from the Internet of Things component, generate a spray signal based on the road information and environmental information (including: the duration of a single water spray operation, the duration of the interval between two operations, and the water pressure), and send the spray signal to the nozzle component and the frequency converter, so that the servo motor controls the water spraying action of the rotating nozzle through the transmission mechanism, thereby achieving fixed-point dust reduction at the dust-producing location, and enables the frequency converter to control the water spraying volume of the water pump, thereby achieving water and energy saving; during the rotation of the rotary nozzle, the controller also receives the trigger signal of the calibration mechanism and compares it with the rotation direction of the rotary nozzle (only the agreed single direction is valid), and then generates a calibration signal to achieve dynamic position soft calibration of the current position of the rotary nozzle.

[0095] The frequency converter is connected to the water pump and the controller, and is used to effectively control the water spray volume by adjusting the power frequency of the water pump according to the spray signal of the controller, and to set the water pressure threshold to achieve water and energy saving while meeting the water spray distance and atomization capacity. Specifically: a water pressure value that is simultaneously greater than the maximum pressure value of the distance response and the maximum pressure value of the water droplet atomization is used as the water pressure threshold to achieve water and energy saving while meeting the water spray distance and atomization capacity.

[0096] The water pump includes an AC asynchronous motor to achieve variable frequency control.

[0097] It should be noted that the spray distance and atomization capacity of the water droplets are closely related to water pressure. The spray distance determines the dust reduction range of the spray pile. The longer the spray distance, the greater the dust reduction range. A sufficient spray distance is required to effectively cover the road area. In theory, smaller water droplets are more likely to combine with and settle in the airborne dust. Therefore, the atomization capacity of the water droplets determines their effective dust reduction capacity to a certain extent. Water pressure is closely related to motor rotation. Within the normal rotation range of the water pump motor, water pressure and water volume are almost directly proportional.

[0098] Specifically, the water quantity and pressure control model of the water pump includes: By analyzing the relationship between water pressure, spray distance, and atomization capacity, we can determine the water pressure required to meet target requirements. The experimental environment is: ambient wind speed ≤ 1 meter (all directions); the nozzle height is the same as in actual conditions, specifically 6 meters.

[0099] 1) Relationship between water spray distance and water pressure:

[0100] in, is the water spray distance; is a constant value, indicating the distance / pressure coefficient; is the actual water pressure; is the pressure response threshold, which can be obtained according to the existing technology; is the lowest 0-point threshold of the pressure response, which can be obtained according to the existing technology.

[0101] From the above formula, it can be seen that before the water pressure reaches the minimum 0-point threshold of the pressure response ( ), no effective spray can be formed; after the water pressure reaches above the pressure threshold ( ), there is no direct relationship between the spray distance of water droplets and water pressure. It can be judged that as long as the water pressure is higher than the threshold ( ), the output water pressure of the water pump controlled by the inverter can be adjusted arbitrarily without affecting the spray distance of the water droplets, that is, without affecting the dust reduction range of the fog pile. Therefore, when the actual water pressure is greater than or equal to the pressure response threshold ( ), you can adjust P as needed without affecting the dust reduction ability of water droplets.

[0102] 2) Relationship between atomization capacity and water pressure:

[0103] in, is the index value of atomization ability; is a constant value, indicating the atomization capacity coefficient; is the atomized particle size; is the actual water pressure; is the minimum threshold of water drop atomization, which can be obtained according to the existing technology; is the maximum threshold of water drop atomization, which can be obtained according to the existing technology.

[0104] From the above formula, it can be seen that the water pressure is less than When the water droplet atomization capacity is 0, the atomization capacity cannot be achieved; when the water pressure is greater than When the water droplet atomization capacity is increased, it is difficult to improve it. That is, when the actual water pressure reaches the maximum threshold of water droplet atomization ( ), P can be adjusted as needed without affecting the dust reduction ability of water droplets.

[0105] Therefore, when P>MAX (P j ,P f ), the water output can be changed by adjusting the inverter output frequency to effectively optimize the water spray volume without affecting the dust reduction capacity and dust reduction coverage per unit water volume. Furthermore, by arbitrarily adjusting the spray pressure of the fog pile (synchronously adjusting the water volume), different water outputs can be achieved under different dust reduction intensity requirements. This can ensure the dust reduction effect of the fog pile equipment while significantly reducing water and power consumption, providing a quantifiable and effective basis for frequency conversion control, system energy saving, and water conservation control goals. For example: during the off-peak period, the required spray volume per unit time is less than that during the peak period. At this time, the spray volume can be adjusted by adjusting the inverter output frequency, achieving a relatively small spray volume per unit time during the off-peak period and a relatively large spray volume per unit time during the peak period.

[0106] The control model of the inverter includes: There is a positive correlation between the water pressure P and the inverter output frequency F, but not a proportional relationship. At the same time, because the water spraying capacity of the equipment is also limited by the water flow limit of the nozzle, a mapping relationship is established through the comparison between P and the inverter output frequency F, such as Figure 4 As shown. Among them: F a =MAX(P j0 ,P f0 ) F b =MAX(P j ,P f ) When F<Fa, no effective water pressure can be formed. When Fa≤F≤Fb, effective water pressure Pa≤P≤Pb is generated. When F>Fb, the generated water pressure no longer changes.

[0107] During the control process, the input end provides the target water pressure and water volume requirements, and provides a reasonable inverter frequency output (i.e., any selection between Fa and Fb).

[0108] Generally speaking, during off-peak periods, the duration, volume, and frequency of water spraying for dust suppression are all lower, and the intervals between sprayings are longer, compared to peak periods. Depending on the spray volume requirements for different time periods or scenarios, the spray volume can be adjusted by adjusting the inverter output frequency (how to do this is known technology), achieving a relatively low spray volume per unit time during off-peak periods and a relatively high spray volume per unit time during peak periods. This feature significantly reduces water and electricity consumption while maintaining the dust suppression effectiveness of the fog pile equipment.

[0109] Therefore, Fa≤F≤Fb and P>MAX (P j ,P f ) when the water pressure is adjusted arbitrarily without affecting the dust reduction capacity and dust reduction range per unit water volume, and then different water outputs under different dust reduction intensity requirements can be achieved by adjusting the water pressure of the fog pile (synchronously adjusting the water volume).

[0110] like Figure 5 As shown, the water pressure / flow control model, fluid control model, visual AI algorithm model, IoT control information model, and positioning cruise control module act on the equipment control assembly (main control cabinet) to control equipment operation.

[0111] The above-mentioned fog pile equipment has the following beneficial effects: 1) The nozzle assembly is designed to achieve 360-degree arbitrary angle positioning and cruising function, and the servo motor and position calibration mechanism ensure long-term and accurate position control.

[0112] 2) The designed camera implements visual AI functions, can monitor left and right lanes, and realize intelligent spraying; it integrates real-time target detection, vehicle classification, regional coordinate mapping, target tracking and counting, dust penetration and light balance algorithms to accurately identify and monitor road vehicle traffic, and can also expand capabilities to achieve functions such as increasing regional video surveillance density, enhancing municipal / park management capabilities, and reporting traffic flow data.

[0113] 3) Designed IoT components collaborate to analyze environmental, meteorological, and traffic data, integrating multi-source data to improve the accuracy of dust suppression decisions. An online learning mechanism updates parameters in real time, allowing for timely adjustments to spray strategies and dynamic adaptation to environmental changes. This intelligent dust suppression system, suitable for urban roads, industrial parks, and other scenarios, reduces PM2.5 concentrations and conserves water resources.

[0114] 4) The designed main control cabinet uses a frequency converter to adjust the water pump power frequency, thereby achieving effective automatic control of the water spray volume of the fog pile and reasonable water saving; and determines the adjustable water pressure range, within which the water pressure can be arbitrarily adjusted without affecting the dust reduction capacity and dust reduction range per unit water volume.

[0115] 5) This application integrates the control with the scene in which the equipment is located, and comprehensively considers the traffic conditions and weather conditions to effectively control dust, realizes the quantitative control, directional attack, environmental perception and judgment of fog pile spray, solves the problem of source control of dust, greatly improves the efficiency and effect of dust control, and significantly reduces power consumption and water consumption. It has good compatibility and scalability, integrates a variety of new technologies to give new capabilities to fog piles, and turns fog pile equipment into an intelligent terminal, greatly expanding its application scenarios and greatly expanding the application boundaries of fog piles, becoming an application node in multiple links such as environmental protection, municipal administration, transportation, and urban management. While improving the air quality of cities and parks, it creates other value, reduces the management burden on collaborative management departments to a certain extent, and reduces the overall cost of urban management.

[0116] It should also be noted that there are no restrictions on the specific appearance design, material selection, specific models and installation methods of the fog pile equipment (such as rainfall detectors).

[0117] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

[0118] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0119] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for controlling a fog pile device, characterized in that: include: Obtaining a road image captured by the fog pile device and processing the road image to obtain vehicle type, number of vehicles, dust cloud size, and cumulative dust generation time; obtaining a first operating intensity of the fog pile device based on the vehicle type, number of vehicles, dust cloud size, and cumulative dust generation time; Obtain local AQI data and traffic flow trends, calculate dust fall intensity, and obtain the second operating intensity of the fog pile equipment; Obtaining an intermediate parameter of the running time according to the first running intensity and the second running intensity; Obtain the shortest single-run duration and the longest single-run duration of the fog pile device, and combine them with the intermediate parameters of the run time to obtain the single-run duration of the fog pile device; obtain the shortest interval between two runs of the fog pile device and the longest interval between two runs, and combine them with the intermediate parameters of the run time and the single-run duration of the fog pile device to obtain the interval between two runs of the fog pile device; Obtain the pressure response threshold and the maximum threshold of water droplet atomization of the fog pile device, and combine the intermediate parameters of the running time and the single running time of the fog pile device to obtain the water pressure of the fog pile device; The operation of the fog pile equipment is controlled according to the duration of a single operation, the interval between two operations and the water pressure.

2. The control method of a fog pile device according to claim 1, characterized in that: According to the first operating intensity and the second operating intensity, intermediate parameters of the operating time are obtained, including: Where, is the intermediate parameter of running time, The shortest duration of a single run. is the first running intensity, is the second running intensity; Obtain the shortest and longest single-running durations of the fog pile device, and combine them with the intermediate parameters of the running time to obtain the single-running duration of the fog pile device, including: Where, is the duration of a single run, The shortest duration of a single run. The maximum duration of a single run.

3. The control method of a fog pile device according to claim 2, characterized in that: Obtain the shortest and longest intervals between two runs of the fog pile device, and combine the intermediate parameters of the run time and the single run time of the fog pile device to obtain the interval between two runs of the fog pile device, including: Where, is the time interval between two runs, The shortest time between two runs. The maximum time interval between two runs.

4. The method for controlling a fog pile device according to claim 3, characterized in that: Obtain the pressure response threshold and the maximum threshold of water droplet atomization of the fog pile device. Combined with the intermediate parameters of the running time and the single running time of the fog pile device, the water pressure of the fog pile device is obtained, including: Where, is the water pressure of the fog pile equipment, is the maximum threshold of water drop atomization, is the pressure response threshold.

5. A method for controlling a fog pile device according to any one of claims 1 to 4, characterized in that: According to the vehicle type, number of vehicles, dust cloud size and accumulated dust generation time, the first operation intensity of the fog pile equipment is obtained, including: The dust intensity within the identified range is obtained based on the vehicle type, number of vehicles, dust cluster size, and cumulative dust generation time: Where, is the dust intensity, It is the integral obtained based on the vehicle type, number of vehicles, dust cloud size and cumulative dust generation time; According to the dust intensity, the first operating intensity of the fog pile equipment is obtained: Where, This is the first operating intensity.

6. The method for controlling a fog pile device according to claim 5, characterized in that: Obtain local AQI data and traffic flow trends, calculate dust fall intensity, and obtain the second operating intensity of the fog pile equipment, including: Obtain local AQI data and traffic flow trends, and calculate dust fall intensity: Where, is the dust fall intensity, For local AQI data, For traffic flow trends, Minimum traffic volume; According to the dust fall intensity, the second operating intensity of the fog pile equipment is obtained: Where, This is the second running intensity.

7. A method for controlling a fog pile device according to any one of claims 1 to 4, characterized in that: Obtain road images taken by fog pile equipment and process them to obtain vehicle type, number of vehicles, dust cloud size, and accumulated dust generation time, including: Obtain the road image captured by the fog pile equipment, define the recognition range in the road image, and use the regional coordinate mapping algorithm to convert the recognition range into the image coordinate system to obtain the detection range; Using real-time target detection algorithm, target detection is performed within the detection range to obtain the vehicle position and dust cloud position; Using classification algorithm, the vehicle position is identified and the vehicle type is obtained; Using target tracking and counting algorithms, different vehicle types are counted to obtain the number of vehicles corresponding to the vehicle type; According to the position of the dust cluster, the size of the dust cluster and the cumulative time of dust generation are obtained.

8. The method for controlling a fog pile device according to claim 7, characterized in that: After obtaining the vehicle type, number of vehicles, dust cloud size and cumulative dust generation time, it also includes: Add dynamic occlusion to the training set images and perform occlusion enhancement training to suppress background interference in the image; Dynamic lighting is added to the training set images for visual enhancement training to improve image transmittance.

9. A control system for a fog pile device, characterized in that: A control method for a fog pile device according to any one of claims 1 to 8, comprising: a pile, a nozzle assembly, a camera, an Internet of Things component, and a main control cabinet; wherein the nozzle assembly and the camera are both arranged on the pile; The nozzle assembly is connected to the main control cabinet and is used to perform fixed-point dust reduction on the dust-generating location according to the spray signal of the main control cabinet; The camera is connected to the main control cabinet and is used to obtain road conditions, process the road conditions, obtain road information, and send the road information to the main control cabinet so that the main control cabinet generates a spray signal; The Internet of Things component is connected to the main control cabinet, and is used to obtain environmental conditions, process the environmental conditions, obtain environmental information, and send the environmental information to the main control cabinet so that the main control cabinet generates a spray signal; The main control cabinet includes: a controller, a frequency converter and a water pump; the controller is used to receive road information from the camera and environmental information of the Internet of Things component, and generate a spray signal based on the road information and the environmental information to control the water spraying action of the sprinkler assembly; the spray signal includes: the duration of a single operation, the duration of the interval between two operations and the water pressure; the frequency converter is connected to the controller and the water pump, and is used to control the water spraying volume by adjusting the power frequency of the water pump according to the spray signal of the controller, and set the water pressure threshold to achieve water and energy saving while meeting the water spraying distance and atomization capacity.

10. The control system of the fog pile equipment according to claim 9, characterized in that: The nozzle assembly includes: a servo motor, a transmission mechanism, a rotary nozzle, a sensing mechanism and a calibration mechanism; The servo motor is connected to the rotary nozzle through the transmission mechanism, so as to control the rotary nozzle through the transmission mechanism according to the spray signal of the controller to perform fixed-point dust reduction at the dust-generating location; the rotary nozzle has a 360-degree motion range; The sensing mechanism is connected to the controller and is spaced apart from the rotary nozzle so as to send a trigger signal to the controller after sensing the position probe of the rotary nozzle; the controller receives the trigger signal of the sensing mechanism and generates a calibration signal and sends it to the calibration mechanism; the calibration mechanism is connected to the controller and the rotary nozzle so as to calibrate the position of the rotary nozzle according to the calibration signal of the controller.

Images