A precise parking control method, system and storage medium based on the Internet of Things

Abstract

The present application relates to the technical field of the Internet of Things, and discloses a precise parking control method, system, and storage medium based on the Internet of Things. The method first obtains first image data and a marker position, uses a preset first algorithm to identify a first target graphic, and if identified, determines its target position. At the same time, the first target height data is obtained to adjust the height parameter of the marker position. The distance value between the target and the marker position is then calculated. Based on the distance value, if it is greater than or equal to a reference value, the difference and direction are calculated and output. If it is less than the reference value, the second target is driven, and a preset second algorithm is used to identify the second target graphic. Its moving distance value is calculated. If it is greater than a moving reference value, the contour overlap data is calculated. Whether parking is successful is determined based on the data, and the moving reference value is inversely correlated with the height parameter. This method can effectively improve parking accuracy, reduce manual operation errors, and improve parking efficiency.

Description

Technical Field

[0001] The present application relates to the technical field of the Internet of Things, and in particular to a precise parking control method, system, and storage medium based on the Internet of Things. Background Art

[0002] Slag powder, whose full name is granulated blast furnace slag powder, also known as ore powder or slag powder, is a fine powder made by drying and grinding the waste slag produced when smelting pig iron in steel plants.

[0003] Slag powder manufacturers use tank trucks and road transport to transport slag powder. The process involves pneumatically loading the slag powder into the tank truck, which is then unloaded to a designated location using the tank truck's built-in air compressor. Tank truck transport offers advantages such as high flexibility, easy loading and unloading, and direct delivery to the construction site. It is suitable for short-distance transportation and for applications requiring flexible delivery times and locations, such as urban construction sites.

[0004] The existing slag powder loading process involves the driver driving the vehicle into the micropowder storage silo, parking within a pre-marked area, climbing onto the vehicle roof to open the vehicle's tank opening. The delivery operator then controls the position of the delivery head so that it falls precisely into the vehicle's tank opening. Because the delivery operator can only observe the alignment of the delivery head with the vehicle's tank opening from a fixed direction, there are many blind spots. To ensure the delivery head is correctly positioned in the vehicle's tank opening, the operator needs to stand up and reconfirm. If the alignment is abnormal and requires fine-tuning, the operator must remotely communicate with the driver to manually adjust the delivery head. The existing manual process of aligning the delivery head with the vehicle's tank opening is time-consuming and inefficient. Summary of the Invention

[0005] In order to improve the efficiency of aligning the slag powder delivery head with the vehicle tank mouth, the present application provides a precise parking control method, system and storage medium based on the Internet of Things.

[0006] In the first aspect, the present application provides a precise parking control method based on the Internet of Things, which adopts the following technical solutions:

[0007] A precise parking control method based on the Internet of Things includes the following steps:

[0008] Acquire first image data and a marking position;

[0009] identifying a first target graphic corresponding to a first target from the first image data based on a preset first algorithm;

[0010] If the first target graphic is recognized, calculating the graphic position of the first target graphic as the target position;

[0011] Acquire height data of the first target, and adjust the height parameter of the marking position in a positive correlation manner according to the height data;

[0012] Calculating a position distance value between the target position and the mark position;

[0013] If the position distance value is greater than or equal to the distance reference value, calculating the distance difference between the position distance value and the distance reference value, calculating the position direction between the target position and the mark position, and outputting the distance difference and the position direction;

[0014] If the position distance value is less than a preset distance reference value, an instruction to drive a second target is issued; and a second target pattern is identified from the first image data based on a preset second algorithm;

[0015] If the second target graphic is identified, a moving distance value of the second target graphic is calculated; if the moving distance value is greater than a preset moving reference value, contour coincidence data of the second target graphic and the first target graphic is calculated;

[0016] If the contour overlap data is within a preset contour overlap range, it indicates that parking is successful; otherwise, it indicates that parking is unsuccessful, wherein the movement reference value is inversely correlated with the height parameter.

[0017] By employing the above technical solution, a preset first algorithm identifies the tank opening from on-site image data and calculates its position and height. This provides an accurate basis for subsequent precise adjustment of the feed tube position, improving positioning accuracy and reducing manual judgment errors. The height parameter of the preset feed tube position is adjusted in a positive correlation with the tank opening height data, allowing the feed tube to better adapt to vehicle tank openings of varying heights. This enhances the system's versatility and adaptability, avoids alignment difficulties caused by varying tank opening heights, and further improves alignment efficiency. The distance difference and direction between the tank opening and the preset feed tube position are calculated and output, providing clear guidance for the system to adjust the feed tube position in real time, enabling the feed tube to move quickly and accurately toward the tank opening, effectively reducing alignment time and improving work efficiency. After issuing a command to drive the feed tube, a second algorithm is used to identify the feed tube's shape and calculate its movement distance and overlap data with the tank opening's outline, ensuring precise alignment of the feed tube with the tank opening. Furthermore, the movement reference value is inversely correlated with the tank opening height parameter, enabling precise control of the feed tube's movement based on the tank opening's height, further ensuring accurate and efficient alignment. The system determines whether parking is successful based on the contour overlap data and provides corresponding prompts, allowing operators to promptly understand the alignment status. If parking fails, adjustments can be made in a timely manner, avoiding ineffective operations and wasted time, helping to continuously optimize the alignment process and improve overall work efficiency.

[0018] Optionally, the method for obtaining the first image data includes the following steps:

[0019] Capturing a first image at a frequency of N images per unit time using a camera device tilted toward the first target;

[0020] Calculating the coefficient of dispersion of the height data within a set time period;

[0021] If the dispersion coefficient is greater than a set threshold, the size of N is adjusted according to the dispersion coefficient. The larger the dispersion coefficient is, the larger N is; the smaller the dispersion coefficient is, the smaller N is.

[0022] By adopting this technical solution, the camera and distance sensor located on industrial equipment are subject to jitter. Data optimization is performed to address this jitter. The camera captures N images—the number captured per unit time—and then performs an average filter. The value of N is adjusted based on the jitter of the distance sensor data. The camera is tilted toward the first target (the tank opening) and captures images at a frequency of N per unit time. Combined with an average filter, this process enables more stable image data despite equipment jitter. The acquisition of multiple images per unit time and the application of an average filter reduces image blur and distortion caused by jitter, improving image quality and providing a more reliable image foundation for accurate tank opening identification. This helps improve the accuracy of image recognition-based tank opening positioning and ensures the accuracy of precise parking control. Using a laser rangefinder tilted toward the second target (the feed pipe), a time-of-flight (TOF) algorithm is used to calculate height data based on the time difference between the transmitted and received laser pulses. This method accurately captures tank opening height information. This method is relatively unaffected by jitter in industrial equipment environments, providing accurate data support for subsequent adjustment of the feed pipe position based on the tank opening height, enhancing the accuracy of height adjustment in precise parking control. The system calculates the dispersion coefficient of height data within a set time period and adjusts the image acquisition frequency N accordingly. When the dispersion coefficient exceeds a set threshold, indicating significant height data jitter, N is increased, increasing the number of images collected per unit time. This allows for more images to be collected for averaging filtering, further reducing the impact of jitter on the image, resulting in clearer and more accurate images of the tank opening and improving the accuracy of tank opening recognition and positioning. Conversely, when the dispersion coefficient is low, N is reduced. This ensures image quality while avoiding resource waste caused by excessive image acquisition, thereby improving system efficiency. The camera's image acquisition frequency N is adjusted based on the jitter of the distance sensor data, achieving coordinated optimization of image acquisition and distance measurement. In the complex environment of industrial equipment jitter, the system can adaptively adjust operating parameters to ensure the accuracy of image and height data, improving the stability and reliability of the entire precision parking control system. This ensures more efficient and accurate alignment of the feed pipe with the vehicle tank opening, enhancing the efficiency of the slag micropowder delivery head.

[0023] Optionally, after acquiring the first image data, the method further includes the following steps:

[0024] Based on a preset dust and fog recognition algorithm, identifying target features from the first image;

[0025] If the target feature is identified, an early warning prompt is issued and an emergency stop command is triggered, and the positioning information of the target feature is sent to the background terminal.

[0026] By adopting this technical solution, if unexpected situations such as dust emission occur during the delivery process, panoramic cameras can capture images of the entire delivery site for identification. If dust emission is detected, the system will immediately stop without human intervention. This not only prevents further dust emission and dust pollution from spreading, but also effectively reduces the risk of equipment failure caused by accidents and protects the health of workers.

[0027] Optionally, based on a plurality of distance sensors facing horizontally toward the first target side wall, the method further comprises the following steps:

[0028] If the first target pattern is recognized, obtaining a set of lateral distance parameters measured by a plurality of distance sensors;

[0029] Fitting a lateral fluctuation curve based on a set of lateral distance parameters, matching the lateral fluctuation curve with a preset lateral distance curve by fluctuation amplitude, and calculating the lateral curve matching degree;

[0030] If the lateral curve matching degree is lower than a preset threshold, the target position is corrected according to a set of lateral distance parameters.

[0031] By implementing the above technical solution, during the slag fine powder delivery process, the position of the vehicle's tank opening (the first target) may deviate or become irregular. This technical solution utilizes multiple distance sensors positioned horizontally toward the sidewalls of the first target, significantly improving the accuracy of tank opening position determination and the stability of delivery operations. After identifying the first target pattern, a set of lateral distance parameters measured by the multiple distance sensors is acquired. These parameters accurately reflect the lateral distance information at different locations of the tank opening. Based on this fitted lateral fluctuation curve, the fluctuation amplitude is matched with a preset lateral distance curve, and the lateral curve matching degree is calculated, providing a quantitative basis for determining whether the actual tank opening position meets the standard. If the lateral curve matching degree falls below a preset threshold, indicating an abnormal deviation in the tank opening position, the target position is corrected based on the set of lateral distance parameters. This effectively corrects misalignment caused by image blur, reduces alignment errors due to mispositioning, mitigates the risk of material leakage, ensures smooth delivery operations, and avoids time and material loss caused by inaccurate tank opening position determination. This significantly improves the accuracy and reliability of the entire slag fine powder delivery process, ensuring efficient and stable delivery operations.

[0032] Optionally, the method for obtaining the height data of the first target includes the following steps:

[0033] Based on a plurality of distance sensors located on top of the first target and pointing vertically downward;

[0034] If the first target figure is recognized, obtaining a set of vertical distance parameters measured by multiple distance sensors;

[0035] Calculating the height data of the first target according to a set of vertical distance parameters;

[0036] A vertical fluctuation curve is fitted based on a set of vertical distance parameters, and the vertical fluctuation curve is matched with a preset vertical distance curve in terms of fluctuation amplitude, and the vertical curve matching degree is calculated;

[0037] If the vertical curve matching degree is lower than a preset threshold, the target position is corrected according to a set of vertical distance parameters.

[0038] The above technical solution, utilizing multiple distance sensors positioned vertically downward from the top of the first target, significantly improves height measurement accuracy and ensures operational stability. Upon identifying the first target pattern, the multiple distance sensors rapidly acquire a set of vertical distance parameters. These parameters comprehensively and accurately reflect the vertical distance from the tank opening top to the ground at various locations. Calculating these parameters yields height data, ensuring accurate height measurement. Furthermore, the vertical distance parameters are fitted into a vertical fluctuation curve, and the fluctuation amplitude is matched against a preset vertical distance curve to calculate the vertical curve matching degree. This provides a quantitative basis for determining the consistency and standardization of the tank opening height. If the vertical curve matching degree falls below a preset threshold, it indicates possible anomalies such as tilt or deformation of the tank opening. In this case, the target position is corrected based on the vertical distance parameters, effectively correcting any deviations in the delivery head positioning caused by this abnormality. This not only reduces the risk of material leakage during delivery, ensuring a safe and clean working environment, but also reduces the time wasted in repeatedly adjusting the delivery head position, significantly improving loading efficiency and ensuring efficient, accurate, and stable loading of slag powder.

[0039] Optionally, based on two pairs of laser ranging devices symmetrically arranged on both sides of the second target in different directions and tilted toward the first target, each pair of laser ranging devices includes two groups of mutually symmetrical laser ranging devices, the method includes the following steps:

[0040] Obtaining the time difference between the emission and reception of each laser pulse in each group of laser ranging devices, and calculating the height data of each laser ranging device from the first target based on the time difference based on a preset algorithm, so as to obtain two sets of height data for each pair;

[0041] The matching degree of the two sets of height data in each pair of laser ranging devices is calculated. If the matching degree is greater than a preset reference value, an instruction to drive the second target is issued.

[0042] The above-mentioned technical solution, utilizing two sets of laser ranging devices symmetrically positioned on either side of the delivery head and tilted toward the tank opening, provides a reliable guarantee for achieving this goal. By measuring the time difference between the emission and reception of each laser pulse in each set of laser ranging devices, a pre-set algorithm accurately calculates the height of each device relative to the tank opening, generating two sets of height data. This multi-set measurement method effectively avoids the measurement errors that can be introduced by a single device, significantly improving the accuracy of the height data. The matching degree of the two sets of height data is then calculated. When the matching degree exceeds a preset reference value, the measured data on both sides are highly consistent, reflecting the accuracy and stability of the tank opening height position. At this point, a command is issued to activate the delivery head. This process ensures that the tank opening height position is accurately determined and meets operational requirements before the delivery head moves, effectively reducing the risk of misalignment caused by blind movement of the delivery head. This not only ensures smooth material transfer during delivery and reduces environmental contamination from material leakage, but also avoids time wasted on repeated adjustments to the delivery head position, significantly improving the overall efficiency of slag powder loading operations.

[0043] Optionally, the method further comprises the following steps:

[0044] Acquiring noise parameters, and performing frequency analysis on the noise parameters;

[0045] If a characteristic frequency is extracted from the noise parameter, the energy value of the characteristic frequency is calculated;

[0046] The size of N is adjusted according to the positive correlation of the energy value. The larger the energy value, the larger the N, and the smaller the energy value, the smaller the N;

[0047] And the cleaning time is adjusted in positive correlation with the energy value. The greater the energy value is, the longer the cleaning time is, and the smaller the energy value is, the shorter the cleaning time is.

[0048] By employing this technical solution, the system adjusts N (image acquisition frequency) based on a positive correlation with energy values. When the vehicle brakes frequently, meaning the energy value is high, N is increased, increasing the number of images collected per unit time. This approach has the advantage of effectively removing image artifacts such as blur and jitter caused by frequent vehicle movement. By acquiring more images and combining them with processing methods like averaging filtering, image clarity and stability can be significantly improved, providing more reliable and accurate image data for subsequent precise identification of the tank opening (primary target). This effectively ensures the precise alignment of the delivery head (secondary target) with the tank opening, avoids misalignment caused by poor image quality, and improves loading accuracy. The system also adjusts the cleaning time based on a positive correlation with energy values. Frequent vehicle braking can raise more dust, leading to increased pollutants in the operating environment. Higher energy values, meaning more frequent braking, extend the cleaning time. This can timely and effectively clean dust and other pollutants attached to the surfaces and optical components of equipment such as laser rangefinders and camera devices, preventing pollutants from adversely affecting equipment performance. It ensures that the laser rangefinder can accurately measure the tank mouth height data and the camera can clearly capture image data, maintaining the accuracy of the entire system's height data measurement, image acquisition, and recognition, and ensuring that the equipment is always in good operating condition. By cleverly utilizing the characteristic frequency and energy value of vehicle braking sounds, dynamic and intelligent adjustment of image acquisition frequency and equipment cleaning time is achieved, enabling the entire slag micropowder loading system to better adapt to complex and changing operating scenarios, effectively improving the stability, accuracy, and efficiency of loading operations, reducing the risk of operational errors caused by difficulties in vehicle position adjustment and environmental interference, and comprehensively optimizing the loading operation process.

[0049] In a second aspect, the present application provides a precise parking control system based on the Internet of Things, which adopts the following technical solutions:

[0050] A precise parking control system based on the Internet of Things includes a processor, wherein the processor executes the steps of any one of the precise parking control methods based on the Internet of Things described above.

[0051] In a third aspect, the present application provides a storage medium that adopts the following technical solution:

[0052] A storage medium stores a program, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for precise parking control based on the Internet of Things.

[0053] In summary, this application includes at least one of the following beneficial technical effects:

[0054] Precisely locate the tank mouth: The preset first algorithm is used to identify the tank mouth from the on-site image data and calculate its position and height, providing an accurate basis for the subsequent precise adjustment of the discharge pipe position, improving the positioning accuracy and reducing the error of manual judgment.

[0055] Adaptive height adjustment: The height parameters of the preset unloading position are adjusted in a positive correlation according to the tank mouth height data, so that the unloading pipe can better adapt to the tank mouths of vehicles with different heights, enhancing the versatility and adaptability of the system, avoiding the difficulty of alignment caused by the difference in tank mouth height, and further improving the alignment efficiency.

[0056] Real-time guidance for alignment: Calculates the distance difference and direction between the tank opening and the preset discharge position and outputs them, providing clear guidance for the system to adjust the discharge pipe position in real time, so that the discharge pipe can move quickly and accurately toward the tank opening, effectively shortening alignment time and improving work efficiency.

[0057] Precisely control the movement of the feed tube: After the command to drive the feed tube is issued, a second algorithm identifies the feed tube's geometry and calculates its movement distance and the overlap with the tank opening's contour, ensuring precise alignment of the feed tube with the tank opening. Furthermore, the movement reference value is inversely correlated with the tank opening height parameter, enabling precise control of the feed tube's movement based on the tank opening's height, further ensuring accurate and efficient alignment.

[0058] Prompt feedback on parking results: The system determines parking success based on contour overlap data and provides prompts, allowing operators to promptly understand the alignment status. If parking fails, adjustments can be made promptly, avoiding ineffective operations and wasted time, helping to continuously optimize the alignment process and improve overall work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a step diagram of a precise parking control method based on the Internet of Things.

[0060] Figure 2 It is a schematic diagram of the positional relationship between a plurality of distance sensors horizontally facing the first target side wall and the side of the vehicle.

[0061] Figure 3 This is a schematic diagram of the relationship between the positions of multiple distance sensors located on top of the first target and pointing vertically downward and the roof.

[0062] Figure 4 This is a schematic diagram of the positional relationship between two pairs of laser ranging devices symmetrically arranged on both sides of a second target in different directions and tilted toward a first target and a vehicle roof.

[0063] Figure 5 This is a schematic diagram of the positional relationship between two pairs of laser ranging devices symmetrically arranged on both sides of the second target in different directions and tilted towards the first target and the side of the vehicle. DETAILED DESCRIPTION

[0064] Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings.

[0065] Throughout this specification, reference to the terms "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0066] The present application discloses a precise parking control method based on the Internet of Things. Figure 1 , including the following steps:

[0067] The first image data and the marked position are acquired in real time. The first image data here refers to the on-site image captured by the camera. For example, at the slag powder loading site, the image will show information such as the vehicle and the surrounding environment. The marked position is the pre-set unloading location.

[0068] The first image data is processed based on the first algorithm to identify a first target graphic. In a real-world scenario, the first target graphic is the opening of a tank on the roof of a vehicle for unloading slag. Once this opening is successfully identified, its position is calculated and determined as the target location.

[0069] The distance sensor acquires the height data of the first target (the tank opening). For example, the tank opening height may vary between different vehicle models, and the distance sensor accurately measures this specific height. Based on this height data, the height parameters of the marking position are then adjusted in a positively correlated manner. For example, if the tank opening is higher, the height of the marking position is raised accordingly.

[0070] The distance between the target position (tank opening) and the marked position (discharging position) is calculated. This distance represents the actual distance between the tank opening and the discharging position. If this distance is greater than or equal to the reference distance, the difference between the distance and the reference distance is calculated. The direction between the target position and the marked position is also calculated and output. This allows the system to determine the distance and direction of the discharging tube that needs to be adjusted.

[0071] If the position distance value is less than the preset distance reference value, a command is issued to drive the second target (the feeding tube). Next, based on a preset second algorithm, the second target pattern (the feeding tube) is identified from the first image data. Once the feeding tube pattern is identified, its movement distance is calculated. If the movement distance value is greater than the preset movement reference value, the contour overlap data of the second target pattern and the first target pattern is calculated.

[0072] The parking success is determined based on the contour overlap data. If the contour overlap data is within the preset contour overlap range, the parking is successful; otherwise, the parking is unsuccessful. It is important to note that the movement reference value and the height parameter are inversely correlated.

[0073] A preset first algorithm identifies the tank opening from on-site image data and calculates its position and height, providing an accurate basis for subsequent precise adjustment of the discharge pipe position. This greatly improves positioning accuracy and reduces errors that may occur due to manual judgment. The height parameters of the preset discharge position are adjusted in a positive correlation with the tank opening height data, allowing the discharge pipe to better adapt to vehicle tank openings of different heights. This enhances the system's versatility and adaptability, avoids alignment difficulties caused by differences in tank opening heights, and further improves alignment efficiency. The distance difference and direction between the tank opening position and the preset discharge position are calculated and output, providing clear guidance for the system to adjust the discharge pipe position in real time, allowing the discharge pipe to move quickly and accurately toward the tank opening, effectively shortening alignment time and improving work efficiency. After issuing a command to drive the discharge pipe, the discharge pipe graphics are identified based on the second algorithm, and its movement distance value and data on overlap with the tank opening's contour are calculated to ensure that the discharge pipe can be accurately aligned with the tank opening. Furthermore, the movement reference value is inversely correlated with the tank opening height parameter, enabling precise control of the feed tube's movement based on the tank opening height, further ensuring accurate and efficient alignment. The system determines parking success based on contour overlap data and provides prompts, allowing operators to quickly understand the alignment status. If parking fails, adjustments can be made promptly, avoiding ineffective operations and wasted time. This helps continuously optimize the alignment process and improves overall work efficiency.

[0074] Given that industrial equipment on which cameras and distance sensors are based often experience jitter, the method for obtaining first image data includes the following steps:

[0075] Image acquisition is performed using a camera device tilted toward the first target (the tank opening). The camera captures the first image at a rate of N images per unit time. In actual slag powder loading sites, equipment vibrations are complex and variable. For example, the engine of the transport vehicle generates continuous low-frequency vibrations, while mechanical collisions during loading and unloading can cause transient high-frequency vibrations. To address this complex vibration environment, the camera captures N images per unit time and then applies an averaging filter. The core principle behind this process is that the degree and direction of vibration influence vary among the multiple images captured per unit time. For example, one image may be blurred and displaced on the left edge of the tank opening due to momentary leftward vibration of the equipment; while another image, captured during rightward vibration, may exhibit similar effects on the right edge. Averaging filtering, by comprehensively averaging the pixel information of these images, can offset the adverse effects of vibration, such as blurring and distortion. This process results in an image that accurately reproduces the true shape of the tank opening, significantly improving image quality. High-quality images provide a solid and reliable foundation for subsequent accurate identification of the tank opening, helping to significantly improve the accuracy of tank opening positioning based on image recognition, thereby effectively ensuring the accuracy of precise parking control.

[0076] To acquire height data, a laser ranging device tilted toward the secondary target (the feed tube) utilizes a Time of Flight (TOF) algorithm to precisely measure the time difference between the transmitted and received laser pulses to calculate height data. This measurement method demonstrates excellent interference resistance in the harsh environment of industrial equipment vibration. Unlike traditional measurement methods that rely on physical contact or are significantly affected by ambient light and air movement, the TOF algorithm exploits the constant propagation velocity of laser pulses in a vacuum and is virtually unaffected by common environmental factors. Even when industrial equipment experiences significant vibration during operation, the laser pulses emitted by the laser ranging device can still reach the tank opening at a stable speed and be accurately reflected back to the receiver. This high-precision measurement of the time difference enables highly accurate acquisition of tank opening height information. For example, during slag powder loading, the laser pulses consistently complete their round trip regardless of equipment vibration, ensuring that the measured height data is relatively unaffected by vibration. This provides critical data support for subsequent precise adjustment of the feed tube position based on tank opening height, significantly enhancing the accuracy of height adjustment for precision parking control.

[0077] To further optimize the image acquisition process, the system also incorporates a calculation step for the coefficient of dispersion of height data within a set time period. As an important indicator of data dispersion, the coefficient of dispersion plays an indispensable role in this system. When the calculated coefficient of dispersion exceeds a set threshold, it clearly indicates significant height data jitter within the set time period. To obtain more stable and clear images of the can opening in complex jittery environments, the system dynamically adjusts the value of N based on the coefficient of dispersion. Specifically, a larger coefficient of dispersion corresponds to a larger value of N. For example, if the coefficient of dispersion far exceeds the set threshold, this indicates extremely severe device jitter, significantly impacting image acquisition quality. In this case, a significant increase in the value of N significantly increases the number of images acquired per unit time. More images participate in the subsequent averaging and filtering process, effectively filtering out the information that best reflects the true shape of the can opening from a larger number of samples. Through multiple averaging and filtering steps, the impact of jitter on the image is further reduced, resulting in clearer and more accurate images of the can opening, significantly improving the accuracy of can opening recognition and location. Conversely, when the coefficient of dispersion is small, it indicates that the height data is relatively stable and the impact of equipment jitter on image acquisition is within an acceptable range. In this case, the value of N can be reduced. This ensures that image quality meets the basic requirements of can mouth recognition while avoiding a series of problems caused by excessive image acquisition, such as excessive storage resource usage, limited data transmission bandwidth, and slowed data processing speed, thereby effectively improving the overall operating efficiency of the system.

[0078] The camera's image acquisition frequency N is dynamically adjusted based on the jitter of the distance sensor data, achieving deep collaborative optimization of the two key links of image acquisition and distance measurement. In the complex environment of industrial equipment vibration, the system can perceive jitter changes in the distance sensor data in real time and respond quickly, adaptively adjusting the camera's image acquisition frequency. When equipment vibration is severe, the system detects an increase in the dispersion coefficient of the distance sensor data and automatically increases the N value, increasing the image acquisition frequency to capture more images and counteract the impact of vibration on image quality. When equipment vibration is mild, the dispersion coefficient decreases, and the system correspondingly decreases the N value and reduces the image acquisition frequency. This adaptive adjustment mechanism ensures that the system consistently maintains the accuracy of image and height data under various vibration conditions. This collaborative optimization significantly improves the stability and reliability of the entire precision parking control system, laying a solid foundation for ensuring more efficient and accurate alignment of the feed pipe with the vehicle tank opening. Ultimately, this improves the efficiency of slag micropowder delivery heads, contributing to more efficient and stable industrial production operations.

[0079] During the slag powder delivery process, the on-site environment is complex and ever-changing, and various unexpected situations may occur. After successfully acquiring the first image data, the system will further conduct an in-depth and detailed analysis of the first image based on the preset dust and fog recognition algorithm, aiming to accurately identify specific target features.

[0080] The dust and fog recognition algorithm has been designed and trained with extensive data, resulting in powerful image analysis capabilities. It can keenly capture subtle features in images associated with abnormal phenomena like dust emission, such as the shape of dust dispersion, grayscale variations caused by concentration changes, and specific texture structures. By accurately analyzing and comparing these features, the algorithm can quickly determine whether the image contains target features corresponding to unexpected conditions like dust emission.

[0081] Once the algorithm successfully identifies the target features, the system immediately initiates a series of efficient and automated countermeasures. First, a sharp and eye-catching warning alert is issued, using various means such as audible and visual alarms to ensure that workers on site are immediately aware of any abnormalities. Simultaneously, an emergency stop command is quickly triggered, immediately halting the ongoing shipping process. This emergency stop command is transmitted to the control systems of the relevant equipment within milliseconds, rapidly braking the equipment's moving parts, preventing further exacerbation of dust emissions caused by continued operation and effectively curbing the spread of dust pollution.

[0082] The system also accurately transmits the target feature's location information to the backend terminal. This location information is precise to the exact coordinates of the dust emission within the shipping site image. Backend staff can visually see the exact location of the dust emission through specialized monitoring software. With this precise location, staff can quickly determine the source of the dust emission and promptly dispatch professional maintenance personnel to the site for investigation and resolution, significantly shortening troubleshooting time and improving problem-solving efficiency.

[0083] During the shipping process, unexpected events such as dust emission not only cause severe pollution to the work environment but can also trigger a series of chain reactions. Large amounts of diffused dust can easily enter delicate equipment components, accelerating wear and increasing the risk of equipment failure. However, panoramic cameras capture comprehensive images of the entire shipping site and intelligently identify dust emission using a dust and fog recognition algorithm. Once dust emission is detected, the system automatically stops the process without manual intervention. This automated mechanism acts like a vigilant guardian, reacting swiftly the moment an incident occurs. This not only prevents dust emission from escalating and preventing the spread of dust pollution within the work area, creating a cleaner and safer working environment, but also effectively reduces the risk of equipment failure caused by accidents, extending equipment life, and reducing maintenance costs. More importantly, it significantly protects the health of workers, reducing the risk of respiratory diseases, pneumoconiosis, and other occupational diseases caused by excessive dust inhalation. This fully demonstrates our concern and attention to the health of our workers and provides a solid foundation for the safe, stable, and efficient operation of slag micropowder shipping.

[0084] Reference Figure 2 In the actual operation of slag powder delivery, the position of the vehicle tank mouth (i.e., the first target) is often complex and changeable. Due to factors such as angle deviation when the vehicle is parked and uneven ground, the tank mouth position may deviate. This application is based on multiple distance sensors facing the side wall of the first target horizontally, and the method also includes the following steps:

[0085] Once the system identifies the first target pattern—the image of the vehicle's tank opening—it acquires a set of lateral distance parameters measured by multiple distance sensors. These sensors are evenly distributed around the tank opening, facing horizontally toward the sidewalls. They accurately measure the lateral distance from the opening at various locations. For example, at a large slag fine powder shipping yard, a transport vehicle's tank opening is equipped with a 3×9 matrix of distance sensors. Each sensor provides real-time feedback on the lateral distance to its corresponding location.

[0086] Based on this set of lateral distance parameters, a lateral fluctuation curve is fitted. This curve reflects the changing trend of the lateral distance of the can opening. This curve intuitively illustrates the shape characteristics and position changes of the can opening in the lateral direction. The fluctuation amplitude of this lateral fluctuation curve is matched with the preset lateral distance curve, and the lateral curve matching degree is calculated. The preset lateral distance curve is an ideal curve pre-set based on the standard can opening shape and position, representing the standard state that the can opening should present. By comparing the fluctuation amplitudes of the two, a specific value, namely the lateral curve matching degree, can be obtained. This value can be used to determine whether the actual position of the can opening is consistent with the position detected by image recognition.

[0087] If the calculated lateral curve matching degree is lower than the preset threshold, it indicates that there is an abnormal deviation in the position of the tank mouth. This indicates that the result of image recognition deviates from the data of the distance sensor, which may be due to problems such as image jitter when the vehicle is parked. At this time, the system will correct the target position based on this set of lateral distance parameters. For example, if the lateral distance parameters show that the tank mouth is offset to the right by a certain distance in the horizontal direction, the system will move the target position determined by image recognition to the right by the same distance in the horizontal direction to match it with the actual position of the tank mouth. The data measured by the distance sensor is generally faster than the result of image recognition. Compared with the time delay of image processing, the results of the sensor are also more accurate and more timely. Therefore, combining the data of the two is more conducive to improving the accuracy of the data.

[0088] The system will correct the target position based on this set of lateral distance parameters. The specific steps are as follows:

[0089] Preprocess the lateral distance parameters measured by multiple distance sensors. Because sensors may be subject to measurement errors due to environmental interference and other factors, data filtering is required. For example, methods such as mean filtering and median filtering are used to remove outliers and make the distance parameters more accurate and reliable. Furthermore, coordinate transformation is performed on the measurement data from each sensor based on their installation location and distribution, unifying them into a standard coordinate system for subsequent analysis.

[0090] The preprocessed lateral distance parameters are compared with the preset standard distance parameters (corresponding to the preset lateral distance curve). For each distance sensor measurement, the difference between it and the standard value is calculated to obtain the deviation value for each position point. These deviation values reflect the difference between the actual position of the tank opening and the standard position at different points. By comprehensively analyzing all deviation values, the direction and approximate magnitude of the overall position deviation of the tank opening can be determined.

[0091] Based on the calculated deviations, the system determines the direction and magnitude of the target position correction. If the majority of the deviations indicate a horizontal shift to the left, the correction direction is right; the average of the deviations determines the magnitude of the correction. To more accurately determine the magnitude of the correction, the system may use a weighted average, assigning different weights to deviations at different locations, as deviations at different locations on the can mouth may have different impacts on the overall position.

[0092] After determining the correction direction and magnitude, the system performs corrections to the target position. This may involve adjusting the target position coordinates within the image recognition system or modifying control parameters related to the target position. For example, in a system that automatically controls the alignment of a slag powder conveying pipeline with a tank opening, the horizontal and vertical position parameters of the conveying pipeline are adjusted based on the corrected target position information to ensure accurate alignment with the tank opening.

[0093] After correcting the target position, the system monitors the distance sensor's measurement data in real time and recalculates the lateral curve matching. If the matching degree remains below the preset threshold, indicating unsatisfactory corrections, the system will recalculate the deviation and make corrections based on the new measurement data until the lateral curve matching degree meets the preset requirements. Throughout this process, the system continuously provides feedback and makes adjustments to ensure the accuracy of the target position.

[0094] The system records relevant information for each correction, including the target position before the correction, the lateral distance parameters used for the correction, the direction and magnitude of the correction, and the final corrected target position. This information can be used for subsequent data analysis and system optimization to continuously improve the accuracy and efficiency of target position corrections.

[0095] This application is based on multiple distance sensors facing horizontally toward the side wall of the first target, which improves the accuracy of the tank mouth position judgment and the stability of the delivery operation. When the system recognizes the first target graphic, a set of lateral distance parameters is obtained, which can accurately reflect the lateral distance information of different positions of the tank mouth, providing a solid data foundation for subsequent accurate judgment and correction. The lateral fluctuation curve fitted based on these parameters is matched with the preset curve to obtain the lateral curve matching degree, so that it is possible to accurately determine whether the tank mouth position is accurately identified in a quantitative manner. This makes it possible to complete the alignment operation quickly and accurately, improving work efficiency. In addition, reducing material loss means reducing production costs and improving the economic benefits of the enterprise.

[0096] Reference Figure 3 In slag powder loading operations, accurately acquiring the height data of the vehicle's tank opening (the first target) is crucial for achieving precise alignment between the shipping head and the tank opening and ensuring smooth operation. The solution uses multiple distance sensors located on top of the first target and pointing vertically downward to complete the acquisition and related processing of height data. The method for acquiring the height data of the first target includes the following steps:

[0097] In actual operations, such as a large-scale slag powder loading facility, multiple distance sensors are installed on the top of each vehicle's tank opening, distributed in a regular pattern. Consider a typical tank truck with five distance sensors evenly distributed on the top of its opening: at the front, rear, left, right, and center. Once the system recognizes the first target pattern—the vehicle's tank opening—the five distance sensors quickly begin operating. Using their own measurement principles, such as laser or ultrasonic ranging technology, they transmit signals downward and measure the time between signal transmission and reception, quickly acquiring a set of vertical distance parameters. For example, the front sensor measures 3.5 meters above the ground, the rear 3.48 meters, the left 3.51 meters, the right 3.52 meters, and the center 4.5 meters. This high height indicates that the light penetrates deep into the tank opening. This set of vertical distance parameters comprehensively and accurately reflects the vertical distance from the different locations on the tank opening to the ground.

[0098] This set of vertical distance parameters is calculated using a simple average method. The data is filtered first, and the distance values measured by the four sensors are added together and divided by 4. This is (3.5 + 3.48 + 3.51 + 3.52) ÷ 4 = 3.502 meters. This result is the calculated height data of the first target, ensuring the accuracy of the height measurement.

[0099] In actual calculations, the number of distance sensors is not simply 5, but more, mainly concentrated near the feed pipe, and the data from the inside of the tank mouth is not used in the calculation of the tank mouth height data. The example of 5 in this embodiment is only for the convenience of the example calculation.

[0100] At the same time, the system fits this set of vertical distance parameters into a vertical fluctuation curve, which visually displays the height variations at different locations on the tank mouth. This vertical fluctuation curve is then matched against a preset vertical distance curve for fluctuation amplitude. The preset vertical distance curve is pre-set based on the height characteristics of a standard tank mouth and represents the ideal tank mouth height. By comparing the fluctuation amplitudes of the two curves, the system uses a specific mathematical formula to calculate the degree of vertical curve matching. For example, after calculation, the vertical curve matching degree is 98%.

[0101] If the vertical curve match falls below a preset threshold—for example, if the preset threshold is 96% and the calculated match is only 85%—this indicates that the image recognition of the tank mouth may deviate from the actual sensor measurement. In this case, the system will correct the target position based on the vertical distance parameter.

[0102] This solution effectively reduces the risk of material leakage during the delivery process. Previously, due to irregularities in the tank opening, the delivery head would be mispositioned, causing material to leak into the work environment, resulting in both material waste and contamination. With this solution, material leakage has been significantly reduced, ensuring a safe and clean work environment. It also reduces the time wasted on repeatedly adjusting the delivery head position.

[0103] Reference Figure 4 and Figure 5 During slag powder loading operations, ensuring the delivery head (second target) accurately aligns with the vehicle's tank opening (first target) is crucial for efficient and safe material loading. To address this, this solution utilizes two pairs of laser rangefinders, symmetrically positioned on either side of the delivery head and tilted toward the tank opening. Each pair contains multiple laser rangefinders, concentrated near the discharge pipe. This creates a highly accurate position measurement and operation control system. The specific steps are as follows:

[0104] The time difference between the emission and reception of each laser pulse in each group of laser ranging devices is obtained, and the height data of each laser ranging device from the first target is calculated based on the time difference based on a preset algorithm, so that each pair of laser ranging devices obtains two sets of height data.

[0105] The matching degree of the two sets of height data in each pair of laser ranging devices is calculated. If the matching degree is greater than the preset reference value, it means that the two sets of data in each pair of data are symmetrical, which means that the data in different directions symmetrical to the direction of the tank mouth are consistent. It can be concluded that the position of the tank mouth is directly below the discharge pipe. Then, an instruction to drive the second target is issued to drive the discharge pipe downward.

[0106] In the complex operation scenario of loading slag powder, environmental factors have a significant impact on the accuracy and stability of the operation process. To effectively address these challenges, this system introduces an intelligent adjustment mechanism based on noise parameter analysis. Its specific steps and functions are as follows:

[0107] Real-time acquisition of noise parameters at the worksite. At the slag powder loading site, various types of machinery operate, and vehicles travel back and forth, creating a complex noise source. The system uses specialized noise acquisition equipment, such as a highly sensitive microphone array, to comprehensively collect these noise signals and convert them into electrical signal parameters for analysis.

[0108] An in-depth frequency analysis of the acquired noise parameters is performed. Using advanced Fourier transform algorithms, the noise signal in the time domain is converted to the frequency domain, clearly displaying the distribution of noise at different frequencies. During this process, the noise spectrum is carefully identified, attempting to capture characteristic frequencies of particular significance. For example, a characteristic frequency is the sound of a vehicle braking when it stops. Because truck brakes use air pressure to push the brake pads, air is sprayed when the vehicle stops. This causes air flow in the airspace, increasing the frequency of dust particle disturbances.

[0109] Once the characteristic frequency is successfully extracted from the noise parameters, the system immediately calculates the energy value of the characteristic frequency. In real-world scenarios, for example, the braking sound generated when a vehicle repeatedly adjusts its position will exhibit specific frequency characteristics in the noise spectrum. When braking, the brake pads and brake discs rub violently, and the sound generated has a relatively concentrated frequency range. The signal corresponding to this frequency range will show a higher energy value in the frequency domain. The system calculates the energy value of this characteristic frequency through precise mathematical operations. The size of this energy value directly reflects the frequency of braking behavior.

[0110] Based on the calculated energy value, the system positively correlates and adjusts two key parameters. The first is N, the image acquisition frequency. Frequent vehicle braking indicates a higher energy value. For example, in a busy loading area, vehicles frequently brake and adjust their positions to accurately park under the shipping head. This significantly increases the energy value of the characteristic frequency of the brake sound in the noise. In this case, the system automatically increases N, increasing the number of images collected per unit time. For example, if 10 images were originally collected per unit time, N can be increased to 20 when the energy value reaches a certain level. This approach has the advantage of fully utilizing redundant information between images when frequent vehicle movement can easily introduce blur and jitter. By acquiring more images and combining them with image processing techniques such as averaging filtering, the system can fully utilize the redundant information between images. Images captured at different times may exhibit varying directions and degrees of blur and jitter. Averaging these images effectively offsets these interferences, significantly improving image clarity and stability. High-quality images provide a more reliable and accurate image data basis for the subsequent precise identification of the position of the vehicle tank mouth (the first target), effectively ensuring the accuracy of the alignment between the shipping head (the second target) and the tank mouth, effectively avoiding alignment errors caused by poor image quality, and greatly improving the accuracy of loading operations.

[0111] The system adjusts cleaning times based on a positive correlation with energy values. Frequent vehicle braking raises large amounts of dust, leading to a dramatic increase in pollutants in the work environment. As energy values increase, meaning more frequent braking occurs, dust concentration at the work site rises significantly. At this point, the system automatically extends cleaning times. For example, under normal circumstances, laser rangefinders and cameras are cleaned every hour. However, if the energy value of the brake sound characteristic frequency remains persistently high, the cleaning interval may be shortened to 30 minutes. This practice ensures timely and effective removal of dust and other contaminants adhering to equipment surfaces and optical components. Prolonged attachment of dust and other contaminants to critical components, such as the transmitting and receiving windows of laser rangefinders and the camera lens, can severely impact equipment performance. For laser rangefinders, dust can cause laser signal scattering and attenuation, significantly compromising the accuracy of tank height measurements. For cameras, dust can reduce lens transmittance, resulting in blurred images and reduced contrast. By extending the cleaning time, it is ensured that the laser ranging device can accurately measure the height data of the tank mouth, and the camera device can clearly collect image data, maintaining the accuracy of the height data measurement and image acquisition and recognition of the entire system, and ensuring that the equipment is always in good operating condition.

[0112] By cleverly utilizing the characteristic frequency and energy of vehicle braking sounds, the system achieves dynamic, intelligent adjustment of image acquisition frequency and equipment cleaning time. This intelligent adjustment mechanism enables the entire slag powder loading system to better adapt to complex and changing operating scenarios, effectively improving the stability, accuracy, and efficiency of loading operations. It also reduces the risk of operational errors caused by difficult vehicle positioning and environmental interference, comprehensively optimizes the loading process, and provides a strong guarantee for the efficient and accurate loading of slag powder.

[0113] An embodiment of the present application further discloses an IoT-based precise parking control system, comprising a processor, wherein the processor executes the steps of any one of the IoT-based precise parking control methods described above.

[0114] An embodiment of the present application further discloses a storage medium, in which a program is stored. When the program is executed by a processor, the steps of any one of the above-mentioned methods for precise parking control based on the Internet of Things are implemented.

[0115] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A precise parking control method based on the Internet of Things, characterized in that: The steps include: Acquire first image data and a marking position; identifying a first target graphic corresponding to a first target from the first image data based on a preset first algorithm; If the first target graphic is recognized, calculating the graphic position of the first target graphic as the target position; Acquire height data of the first target, and adjust the height parameter of the marking position in a positive correlation manner according to the height data; Calculating a position distance value between the target position and the mark position; If the position distance value is greater than or equal to the distance reference value, calculating the distance difference between the position distance value and the distance reference value, calculating the position direction between the target position and the mark position, and outputting the distance difference and the position direction; If the position distance value is less than a preset distance reference value, issuing a command to drive the second target; identifying a second target graphic from the first image data based on a preset second algorithm; If the second target graphic is identified, a moving distance value of the second target graphic is calculated; if the moving distance value is greater than a preset moving reference value, contour coincidence data of the second target graphic and the first target graphic is calculated; If the contour overlap data is within the preset contour overlap range, a prompt is given indicating that parking is successful; otherwise, a prompt is given indicating that parking is unsuccessful, wherein the movement reference value is inversely correlated with the height parameter; The method for obtaining the first image data includes the following steps: Capturing a first image at a frequency of N images per unit time using a camera device tilted toward the first target; Calculating the coefficient of dispersion of the height data within a set time period; If the dispersion coefficient is greater than the set threshold, the size of N is adjusted according to the dispersion coefficient. The larger the dispersion coefficient is, the larger N is; the smaller the dispersion coefficient is, the smaller N is. Based on two pairs of laser ranging devices symmetrically arranged on both sides of the second target in different directions and tilted toward the first target, each pair of laser ranging devices includes two sets of mutually symmetrical laser ranging devices, and the method includes the following steps: Obtaining the time difference between the emission and reception of each laser pulse in each group of laser ranging devices, and calculating the height data of each laser ranging device from the first target based on the time difference based on a preset algorithm, so as to obtain two sets of height data for each pair; The matching degree of the two sets of height data in each pair of laser ranging devices is calculated. If the matching degree is greater than a preset reference value, an instruction to drive the second target is issued.

2. The precise parking control method based on the Internet of Things according to claim 1 is characterized in that: After acquiring the first image data, the method further includes the following steps: Based on a preset dust and fog recognition algorithm, identifying target features from the first image; If the target feature is identified, an early warning prompt is issued and an emergency stop command is triggered, and the positioning information of the target feature is sent to the background terminal.

3. The precise parking control method based on the Internet of Things according to claim 1 is characterized in that: Based on a plurality of distance sensors facing horizontally toward the first target side wall, the method further comprises the following steps: If the first target pattern is recognized, obtaining a set of lateral distance parameters measured by a plurality of distance sensors; Fitting a lateral fluctuation curve based on a set of lateral distance parameters, matching the lateral fluctuation curve with a preset lateral distance curve by fluctuation amplitude, and calculating the lateral curve matching degree; If the lateral curve matching degree is lower than a preset threshold, the target position is corrected according to a set of lateral distance parameters.

4. The precise parking control method based on the Internet of Things according to claim 1 is characterized in that: The method for obtaining the height data of the first target comprises the following steps: Based on a plurality of distance sensors located on top of the first target and pointing vertically downward; If the first target figure is recognized, obtaining a set of vertical distance parameters measured by multiple distance sensors; Calculating the height data of the first target according to a set of vertical distance parameters; A vertical fluctuation curve is fitted based on a set of vertical distance parameters, and the vertical fluctuation curve is matched with a preset vertical distance curve in terms of fluctuation amplitude, and the vertical curve matching degree is calculated; If the vertical curve matching degree is lower than a preset threshold, the target position is corrected according to a set of vertical distance parameters.

5. The precise parking control method based on the Internet of Things according to claim 1 is characterized in that: The method further comprises the steps of: Acquiring noise parameters, and performing frequency analysis on the noise parameters; If a characteristic frequency is extracted from the noise parameter, the energy value of the characteristic frequency is calculated; The size of N is adjusted according to the positive correlation of the energy value. The larger the energy value, the larger the N, and the smaller the energy value, the smaller the N; And the cleaning time is adjusted in positive correlation with the energy value. The greater the energy value is, the longer the cleaning time is, and the smaller the energy value is, the shorter the cleaning time is.

6. A precise parking control system based on the Internet of Things, characterized in that: The method comprises a processor, wherein the processor executes the steps of the precise parking control method based on the Internet of Things according to any one of claims 1 to 5.

7. A storage medium, characterized in that: The medium stores a program, and when the program is executed by the processor, the steps of the precise parking control method based on the Internet of Things according to any one of claims 1 to 5 are implemented.

Images