Intelligent vehicle tire cleaning method and system based on fusion perception and intelligent spray gun

Abstract

The invention relates to an intelligent vehicle tire cleaning method and system based on fusion perception and an intelligent spray gun, and relates to the technical field of intelligent vehicle cleaning. The method comprises the steps that a global sensing subsystem, an intelligent spray gun execution subsystem and a central intelligent decision-making unit are built; when a vehicle is driven in, the global sensing subsystem outputs real-time dynamic parameters; the central intelligent decision-making unit allocates tasks according to a cross type tracking pre-washing mode, and drives the intelligent spray guns to cross wash key areas of tires on the opposite side and relay to pre-wash the tires on the same side; after the vehicle is stably parked, the global sensing subsystem generates a tire fine three-dimensional model and a dirt distribution diagram; the central intelligent decision-making unit controls the intelligent spray gun to execute a profiling fine cleaning mode; when the vehicle is driven out, the global sensing subsystem identifies the cleaning platform and the dirt residual area of the drainage ditch and outputs contour data; and the central intelligent decision-making unit controls all the intelligent spray guns to switch a ground curtain self-cleaning mode for synchronous spraying and sweeping. The vehicle tire cleaning device can efficiently, thoroughly and stably clean vehicle tires.

Description

Technical Field

[0001] This invention relates to the field of intelligent car wash technology, and in particular to an intelligent vehicle tire cleaning method and system based on fusion sensing and intelligent spray guns. Background Technology

[0002] During the loading, unloading, and transportation of freight vehicles, tire surfaces are highly susceptible to the accumulation of various contaminants, especially the tire tread grooves, sidewalls, and the gaps between the tires. These areas, due to their complex structure, become key and challenging areas for cleaning. Current cleaning methods in the industry still have many significant shortcomings, failing to meet the demands for efficient, thorough, and automated cleaning. Traditional cleaning methods often rely on manual handheld high-pressure water guns or fixed high-pressure nozzle arrays for rinsing. This not only results in low cleaning efficiency but also significant water waste. Furthermore, limited by fixed structures or manual operation, these methods cannot cover all areas of the tire's complex three-dimensional curved surface, leading to numerous cleaning blind spots.

[0003] While some existing technologies and related patents propose identifying dirt and adjusting cleaning parameters through vision and spectral analysis, they still fail to specifically address the complex cleaning needs of tires, a critical and difficult-to-clean area. The problem of blind spots in cleaning remains prominent, severely impacting the overall cleaning effect. Furthermore, in actual cleaning conditions, the environment of high-pressure water mist and frequent water splashes significantly interferes with the recognition performance of traditional visual aids, causing the perception system to "go blind," making it impossible to achieve stable tracking and accurate decision-making of the tires, and hindering the realization of fully automated closed-loop cleaning.

[0004] From the perspective of the execution mechanism, existing solutions either use fixed nozzles, which cannot adapt to the complex contours of freight vehicle tires, or rely on structurally complex robotic arms, which are not only costly but also prone to failure in harsh engineering environments, exhibiting poor adaptability and reliability, and incurring high subsequent maintenance costs. Furthermore, in existing cleaning processes, the stages of vehicle entry, stationary movement, and exit are isolated from each other, lacking a coherent and coordinated design, and failing to form an overall solution with optimal resource utilization. Even if some solutions possess intelligent decision-making models, the lack of cleaning execution mechanisms adaptable to the complex structure of freight vehicles prevents the effective implementation of excellent cleaning strategies, ultimately resulting in significantly reduced cleaning effects and failing to meet the needs for efficient, thorough, and stable cleaning of freight vehicle tires. Summary of the Invention

[0005] Therefore, it is necessary to provide a vehicle tire intelligent cleaning method and system based on fusion perception and intelligent spray gun that can efficiently, thoroughly and stably clean vehicle tires, addressing the aforementioned technical problems.

[0006] A vehicle tire intelligent cleaning method based on fusion sensing and intelligent spray gun, the method comprising:

[0007] Step 1: Build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit; Step 2: When the vehicle enters, the global perception subsystem senses the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-washing mode, drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-washing the same side tire. Step 3: After the vehicle comes to a complete stop, the global perception subsystem acquires multi-source data and fuses it to generate a detailed three-dimensional model of the tire and a dirt distribution map; the central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans a path based on the detailed three-dimensional model and dirt distribution map, and then controls the intelligent spray gun to perform contour-following fine cleaning mode. Step 4: When the vehicle leaves, the global perception subsystem identifies the area of ​​residual dirt on the cleaning platform and drainage ditch and outputs contour data; the central intelligent decision-making unit plans the spraying path based on the contour data and controls all intelligent spray guns to switch to the ground curtain self-cleaning mode for synchronous spraying.

[0008] On the other hand, a vehicle tire intelligent cleaning system based on fusion perception and intelligent spray guns is also provided, including: The cleaning system construction module is used to build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit; The pre-cleaning module is used when a vehicle enters. The global perception subsystem perceives the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-cleaning mode, and drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-cleaning the same side tire. The contour-following fine cleaning module is used after the vehicle has come to a complete stop. The global perception subsystem acquires multi-source data and fuses it to generate a fine three-dimensional model of the tire and a dirt distribution map. The central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans a path based on the fine three-dimensional model and dirt distribution map, and then controls the intelligent spray gun to execute the contour-following fine cleaning mode. The self-cleaning module is used when the vehicle leaves. The global perception subsystem identifies the area of ​​residual dirt on the cleaning platform and drainage ditch and outputs contour data. The central intelligent decision-making unit plans the spraying path based on the contour data and controls all intelligent spray guns to switch to the self-cleaning mode for synchronous spraying.

[0009] Compared with existing technologies, the intelligent vehicle tire cleaning method and system based on fusion perception and intelligent spray gun provided by this invention has the following beneficial effects: 1. Relying on the global perception subsystem for precise tire perception and data fusion processing, combined with the flexible adjustment capability of the intelligent spray gun, it can adapt to the complex three-dimensional contours of the tire, cover traditional cleaning blind spots such as tread grooves and double wheel gaps, and achieve thorough cleaning.

[0010] 2. By employing a phased and targeted approach involving cross-tracking pre-washing, contour-following fine washing, and floor curtain self-cleaning, coupled with intelligent decision-making for task allocation and path planning, the system ensures cleaning effectiveness while reducing water waste and improving resource utilization efficiency.

[0011] 3. The global perception subsystem adopts an anti-interference fusion architecture, which effectively resists the interference of high-pressure water mist and water splash during the cleaning process, ensuring stable identification and accurate decision-making of tire status and dirt distribution, and guaranteeing the stability and reliability of the cleaning process.

[0012] 4. The entire cleaning process is automated and closed-loop, with each stage of vehicle entry, parking, and exit coordinated and seamless, requiring no human intervention. This significantly improves cleaning efficiency and solves the problems of isolated stages and high reliance on manual labor in traditional cleaning processes. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention, and those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating the intelligent vehicle tire cleaning method based on fused perception and intelligent spray gun in Example 1. Figure 2 This is a schematic diagram showing the positions of the global perception subsystem and the intelligent spray gun in Example 1; Figure 3 This is a schematic diagram of the intelligent spray gun structure in Example 1; Figure 4 This is a structural block diagram of the intelligent vehicle tire cleaning system based on fusion perception and intelligent spray gun in Example 2.

[0015] Explanation of reference numerals in the attached figures: Car wash lane 1, LiDAR 2, Industrial camera 3, Intelligent spray gun 4, Lifting component 41, High degree of freedom gimbal 42.

[0016] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly.

[0019] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0020] In this invention, unless otherwise explicitly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; 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 the internal communication of two elements or the interaction between two elements, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Furthermore, the detachable connection described in this invention includes, but is not limited to, snap-fit ​​connections, threaded connections, pin connections, magnetic connections, plug-in connections, etc., which can be selected flexibly according to the situation. The specific detachable connection methods shown in the following embodiments are just one feasible method and are not intended to be the only limitation.

[0022] It is understood that the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0023] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] Example 1 like Figure 1 As shown, this embodiment provides a vehicle tire intelligent cleaning method based on fused perception and intelligent spray gun, including the following steps: Step 1: Build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit.

[0025] Step 2: When the vehicle enters, the global perception subsystem senses the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-washing mode, drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-washing the same side tire.

[0026] Step 3: After the vehicle comes to a complete stop, the global perception subsystem acquires multi-source data and integrates it to generate a detailed 3D model of the tire and a dirt distribution map. The central intelligent decision-making unit discretizes the 3D contour of the tire and plans the path based on the detailed 3D model and dirt distribution map, and then controls the intelligent spray gun to perform contour washing mode.

[0027] Step 4: When the vehicle leaves, the global perception subsystem identifies the area of ​​residual dirt on the cleaning platform and drainage ditch and outputs contour data; the central intelligent decision-making unit plans the spraying path based on the contour data and controls all intelligent spray guns to switch to the ground curtain self-cleaning mode for synchronous spraying.

[0028] The intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun provided by this invention builds an integrated cleaning system that combines fusion perception, intelligent execution and decision control. It executes targeted cleaning modes in stages, achieving efficient, thorough and stable automated cleaning of vehicle tires. At the same time, it solves the problems of many blind spots, low efficiency and waste of water resources in traditional cleaning methods.

[0029] In the specific implementation of step 1, the global perception subsystem adopts an anti-interference fusion architecture that uses LiDAR 2 as the main sensor, industrial camera 3 as a secondary sensor, and auxiliary sensors. For example... Figure 2 As shown, the lidar 2 uses 2-4 long-wavelength anti-water-mist lidars, which are vertically arranged on both sides of the car wash channel 1. The arrangement position should be such that it can collect global information of the vehicle. This type of lidar has strong penetration and is less affected by water mist. It is used to acquire three-dimensional point cloud data of the vehicle and tires in real time.

[0030] The industrial camera 3 uses 2-4 high dynamic range industrial cameras, which are deployed on the top and sides of the car wash channel 1. The deployment position should be such that it can collect global information of the vehicle. It works with LiDAR to achieve multi-angle image acquisition. At the same time, the industrial camera 3 has automatic exposure and a splash-proof housing.

[0031] Auxiliary sensors are installed on the floor and side walls of the cleaning channel 1 to work with the industrial camera 3 to assist in identification and verification during non-direct rinsing periods.

[0032] The intelligent spray gun execution subsystem includes multiple intelligent spray guns 4 symmetrically deployed on both sides of the washing channel 1. The intelligent spray guns 4 are installed on the ground on both sides of the washing channel 1 in a completely symmetrical layout. The number and spacing of the intelligent spray guns 4 are configured according to the length of the car wash channel 1 and the required cleaning intensity, generally 4-6 units, to ensure that the tires can be effectively covered throughout the entire length of the vehicle.

[0033] like Figure 3 As shown, each intelligent spray gun 4 integrates a lifting component 41 and a high-degree-of-freedom gimbal 42. The lifting component 41 is a component with lifting function, mainly used to adjust the overall height of the intelligent spray gun 4. The lifting component 41 is equipped with a cylindrical protective sleeve on its outer periphery. The protective sleeve is made of thickened stainless steel and has an integrated guide rail inside to ensure smooth lifting. It uses a high-thrust electric push rod as the lifting drive power source, which allows the intelligent spray gun head to be raised and lowered quickly and stably. Its vertical lifting stroke can reach up to 1 meter, which can exceed the height of large engineering tires. A displacement feedback device is installed inside the electric push rod to provide real-time feedback on the height of the intelligent spray gun head. Preferably, the lifting component 41 can use mature components such as a lifting column, which will not be described in detail here.

[0034] The high-degree-of-freedom gimbal 42 is mounted on the upper end of the lifting component 41. It has horizontal rotation and pitch swing functions to adjust the horizontal rotation angle and pitch angle of the nozzle. Its horizontal rotation structure is driven by a servo motor, with a rotation range of ±100° based on the perpendicular direction of vehicle travel. The pitch swing structure is also driven by a servo motor, with the maximum upward pitch angle reaching the cleaning of the raised tire sides or the edge of the vehicle chassis, and the maximum downward pitch angle pointing towards the ground and drainage ditches. Both the horizontal and pitch axes are equipped with encoders, which can provide real-time feedback on the gimbal angle.

[0035] A smart spray gun 4 is installed at the upper end of the high-degree-of-freedom gimbal 42. The smart spray gun 4 is an intelligent water cannon with adjustable water pressure and flow rate to adapt to different cleaning conditions. For example, to combat stubborn deposits, in the contour-following fine washing mode, the water jet of the smart spray gun 4 is adjusted to a high-pressure focusing form according to the dirt distribution map, to perform targeted penetrating rinsing of highly polluted areas; to perform large-area dirt rinsing, in the cross-tracking pre-wash mode and the floor curtain self-cleaning mode, the water jet of the smart spray gun 4 is adjusted to a wide-angle fan-shaped water flow form for wide-area coverage, removing surface dust, mud, and dirt from the ground. Preferably, the overall structure of the high-degree-of-freedom gimbal 42 and the smart spray gun 4 can use mature components such as fire cannons and electronic water cannons from sprinkler trucks, which will not be elaborated here.

[0036] The central intelligent decision-making unit has a built-in adaptive cleaning model and a collaborative controller. By inputting feature data from different stages, it outputs different cleaning modes and selects the best cleaning mode for each tire at different stages. Under different cleaning modes, the collaborative controller decomposes the cleaning mode into specific, time-sequential control commands to ensure that the intelligent spray gun performs the collaborative operation of the subsystem.

[0037] The cleaning system built in this step, through the coordinated operation of the global perception subsystem, the intelligent spray gun execution subsystem, and the central intelligent decision-making unit, provides stable and precise hardware support and control foundation for subsequent cleaning stages. Compared with traditional cleaning equipment, it has stronger anti-interference capabilities and wider adaptability, and can meet the complex cleaning needs of freight vehicle tires.

[0038] In the specific implementation of step 2, real-time dynamic parameters refer to the core data set that can reflect the real-time motion state of the vehicle tires, including the three-dimensional coordinates of each row of wheels in the global coordinate system and the vehicle's speed. and tire rotational angular velocity .

[0039] Specifically, in the global perception subsystem, vehicle point cloud data is acquired via LiDAR 2, and image data is acquired via industrial camera 3. After fusing the vehicle point cloud data and image data, the Kalman filter algorithm is used to track and estimate the displacement between consecutive frames. Then, the time series difference method is used to calculate the tire rotation speed and vehicle speed. The calculation formula is as follows: ; ; ; In the formula, This represents the average displacement change of the wheel axle between two adjacent sampling times; express The position of the wheel axle at any given moment; express The position of the wheel axle at any given moment; Indicates the sampling period; Indicates the vehicle's speed; Indicates the angular velocity of the tire rotation; This represents the change in tire angle between consecutive frames; Simultaneously, by fitting the wheel profile center coordinates using Hough transform and combining them with the spatiotemporal calibration results of auxiliary sensors, the three-dimensional coordinates of the wheel in the global coordinate system are calculated. Finally, these three-dimensional coordinates and the vehicle's speed are... and tire rotational angular velocity The integrated parameters are output as real-time dynamic parameters.

[0040] The central intelligent decision-making unit predicts tire trajectories based on real-time dynamic parameters and allocates tasks according to a cross-tracking pre-washing mode. This drives four intelligent spray guns to cross-clean key areas of the opposite tires and then relay pre-wash the tires on the same side. The cross-tracking pre-washing mode specifically employs a spatiotemporal coupling scheduling algorithm, combining real-time wheel dynamic parameters to calculate the spatiotemporal windows for each intelligent spray gun's intervention and exit from cleaning, assigning a dynamic cleaning time window and spatial path point sequence to each intelligent spray gun. The specific calculation formula is as follows: ; ; In the formula, Indicates when the intelligent spray gun intervenes in the cleaning process; Indicates the preset cleaning start distance; Indicates when the intelligent spray gun exits the cleaning process; This indicates the cleaning time is dynamically adjusted based on the degree of dirt on the tires. It can be understood that this algorithm can accurately allocate the working time and coverage area of ​​each intelligent spray gun, driving the intelligent spray guns to cross-clean key areas of the opposite tire, such as tire tread grooves and sidewalls, areas prone to dirt accumulation, thus improving the targeted nature of the pre-wash.

[0041] When pre-washing the tires on the same side in relay, the start-up time of the downstream intelligent spray gun must be precisely matched with the cleaning progress of the upstream spray gun to avoid missed cleaning or repeated cleaning. The calculation formula is as follows: ; In the formula, Indicates the start-up time of the downstream intelligent spray gun; Indicates the start-up time of the upstream intelligent spray gun; This indicates the tire radius. This calculation method enables seamless connection of the tire cleaning areas on the same side. After the upstream spray gun completes the first half of the cleaning, the downstream spray gun immediately continues the cleaning of the second half. Phase difference control ensures cleaning coverage while reducing ineffective water spraying, improving pre-washing efficiency and water resource utilization.

[0042] In addition, the central intelligent decision-making unit will plan the movement path of the intelligent spray gun 4 based on the predicted tire trajectory, so that the nozzle focus point is incident along the normal direction of the outer edge of the tire as much as possible, realizing the spatiotemporal matching of the cleaning force field and the geometric features of the tire surface, and further improving the pre-washing effect.

[0043] The spatial path sequence of the tire trajectory is generated into a smooth trajectory through B-spline curve interpolation, ensuring that the nozzle accurately follows the wheel motion envelope in three-dimensional space. This achieves spatiotemporal matching between the cleaning force field and the geometric features of the tire surface, ensuring cleaning coverage while avoiding water flow interference, and improving resource utilization efficiency and system response robustness.

[0044] This step combines fusion sensing technology with intelligent scheduling algorithms to achieve dynamic and precise pre-washing during vehicle operation. Compared with the traditional fixed nozzle pre-washing mode, it not only has a more comprehensive coverage, but also can dynamically adjust the cleaning strategy according to the real-time movement of the vehicle, quickly removing floating dust and mud from the tire surface, laying a good foundation for the subsequent fine washing process, while effectively reducing water waste.

[0045] In the specific implementation of step 3, after the vehicle comes to a complete stop, the global perception subsystem continues to operate. The LiDAR 2, industrial camera 3, and auxiliary sensors collaboratively acquire multi-source data, providing data support for the generation of a detailed 3D tire model and a contaminant distribution map. The detailed 3D tire model refers to a model that accurately reflects the three-dimensional geometry of the tire, while the contaminant distribution map is a distribution map containing information such as the location and density of contaminants on the tire surface. The generation of both requires multi-source data fusion processing, specifically including the following steps: Step 301: Obtain multi-view point cloud data of the tire surface by scanning with LiDAR 2, and collect texture information of the tire surface by industrial camera 3; take the multi-view point cloud data and texture information as input, perform registration and fusion by ICP algorithm to eliminate spatial deviation of multi-source data, and output fused point cloud data in a unified coordinate system.

[0046] Step 302: The fused point cloud data is used as input. Voxel filtering is used to remove point cloud noise and improve data quality. Then, Poisson reconstruction is used to optimize the surface geometry of the point cloud so that the model fits the actual tire contour better and outputs a detailed 3D tire model.

[0047] Step 303 involves using the fused point cloud data as input, identifying the attachment regions in the fused point cloud data based on a deep learning semantic segmentation network, accurately distinguishing between dirt and the tire body, quantifying the pollution density of each region, and generating a pixel-by-pixel dirt heatmap to visually present the pollution distribution. The deep learning semantic segmentation network can employ lightweight architectures such as U-Net and PointPillars, adapted to the fusion features of point clouds and images.

[0048] Step 304: Align the pixel-by-pixel dirt heatmap with the detailed 3D model of the tire in terms of spatial coordinates, so that the dirt information and the geometric position of the tire are accurately matched, and output a dirt distribution map containing the geometric features of the tire and the dirt distribution information.

[0049] Based on the aforementioned detailed 3D model and dirt distribution map, the central intelligent decision-making unit discretizes the 3D contour of the tire and plans its path. The specific process is as follows: Step 311: Extract the pollution density of each area based on the pollution distribution map. By comparing the preset pollution density threshold with the pollution density of each area, high-pollution areas and conventionally polluted areas can be identified, providing a basis for differentiated cleaning.

[0050] Step 312: Based on the detailed 3D model of the tire, for conventionally polluted areas, at sampling intervals... Extract the coordinates of contour points along the tire circumference to form an ordered point set. For highly polluted areas, the sampling interval is shortened to perform encrypted extraction, outputting an encrypted ordered point set, thereby improving the cleaning accuracy of highly polluted areas. Preferably, for highly polluted areas, the sampling interval is set to... This ensures that the sampling density in key areas is doubled.

[0051] Step 313: After discretizing the ordered point set and the encrypted ordered point set, the Bézier curve is used for fitting and optimization to generate a continuous and differentiable spray gun trajectory, ensuring smooth spray gun movement and avoiding jamming, and outputting the cleaning path planning result.

[0052] When the intelligent spray gun 4 executes the contour-following fine cleaning mode, the central intelligent decision-making unit converts the above path planning results into control commands, driving the lifting component 41 and the high-degree-of-freedom gimbal 42 of the intelligent spray gun to work together. Specifically, the contour-following fine cleaning mode is as follows: the lifting component 41 dynamically adjusts the overall height of the spray gun according to the tire height to ensure that the nozzle is always at the optimal cleaning height; the high-degree-of-freedom gimbal 42 adjusts the nozzle angle through horizontal rotation and pitch swing, so that the nozzle of the intelligent spray gun 4 always conforms to the tire contour, and can accurately cover even complex areas such as the gaps between the two tires; at the same time, according to the distribution of dirt, the water pressure and flow rate of the intelligent spray gun 4 are adjusted to cross-wash difficult-to-clean areas such as the gaps between the two tires, ensuring that dirt is completely removed.

[0053] Furthermore, after cleaning, the global perception subsystem also provides feedback on cleanliness characteristics, determining whether to perform incremental or targeted intensive cleaning based on preset cleanliness thresholds. Specifically, the global perception subsystem performs a rapid local scan of the cleaned tire, collecting tire surface image data and extracting features such as texture complexity and color uniformity as cleanliness evaluation indicators, which are then compared with preset cleanliness thresholds. If the cleanliness does not meet the threshold requirements, a supplementary cleaning path is generated for the residual dirt areas, and the intelligent spray gun 4 is controlled to perform incremental cleaning, cleaning only the areas that do not meet the standards, avoiding the waste of resources caused by full rework. If there are stubborn dirt causing the cleanliness to be significantly lower than the threshold, such as below 95% of the threshold, it is determined that there is stubborn dirt in the area, and the water jet of the intelligent spray gun 4 is switched to a high-pressure focused water flow for targeted intensive rinsing. The focused high-pressure water flow specifically targets stubborn dirt until a re-inspection shows that the standards are met, forming a cleaning closed loop to ensure the cleaning effect.

[0054] This step achieves precise cleaning of the complex contours of the tire and key contaminated areas through refined perception modeling, path planning, and closed-loop control, effectively solving the blind spot problem of traditional cleaning methods. At the same time, the cleanliness judgment mechanism ensures the stability and consistency of the cleaning effect.

[0055] In the specific implementation of step 4, after the contour washing in step 3 is completed, the global perception subsystem will first scan the dirt on the cleaning platform. The central intelligent decision-making unit will drive the intelligent spray guns to simultaneously perform tire rinsing and platform preliminary self-cleaning. During tire rinsing, a fan-shaped water curtain will be used to thoroughly rinse the tires with low pressure to remove residual cleaning agent and tiny dirt. During the platform preliminary self-cleaning, all intelligent spray guns will be uniformly pointed at the road surface driven by the wheels, and the dirt on the ground will be cleaned in a planned manner according to the vehicle's driving route to avoid secondary pollution caused by the vehicle driving out and running over the dirt on the ground. The cleaning method is the same as the ground curtain self-cleaning mode after the vehicle has completely driven out.

[0056] Then, the global perception subsystem identifies the areas of residual dirt on the cleaning platform and in the drainage ditch and outputs contour data, specifically: Step 401: Point cloud data of the cleaning platform and drainage ditch is collected using LiDAR 2. This point cloud data includes information such as the location and thickness of dirt on the ground. The collected point cloud data is filtered to remove noise points and reflections from irrelevant objects, such as reflections from surrounding debris, and the effective point cloud data of the ground area is output. Image data is collected using industrial camera 3, and the height difference between the high-humidity mud and water residue area and the ground is calculated based on the image data.

[0057] Step 402: Based on the effective point cloud data of the ground area, the high humidity mud and water residue area is segmented according to the preset reflection intensity threshold. The reflection intensity of the high humidity area is significantly different from that of the dry ground, which can be quickly distinguished. Then, the boundary of the high humidity mud and water residue area is identified by combining the difference in height from the ground, and the distribution range of the dirt is clarified.

[0058] Step 403: The high-humidity muddy water residue area is clustered using a Euclidean clustering algorithm. The discrete mud point cloud is merged into several continuous regions. The outer contour coordinates of each mud region are extracted, and the contour data of the mud residue area is output, providing a precise basis for path planning. The formula for calculating the contour data of the mud residue area is: ; In the formula, Indicates the first Point cloud clustering of individual contaminated areas; Represents the original point cloud set; Indicates the cluster center; Indicates the neighborhood radius; Point cloud The intensity of reflection; Indicates the humidity discrimination threshold; Point cloud The vertical coordinates; Indicates the ground reference elevation; This indicates the height tolerance threshold.

[0059] After the vehicle has completely driven away, the central intelligent decision-making unit plans the spraying path based on the aforementioned contour data. The path planning must ensure that all areas with residual dirt are covered, and that the spraying ranges of adjacent intelligent spray guns have reasonable overlap to avoid cleaning dead spots. Subsequently, the central intelligent decision-making unit controls all intelligent spray guns 4 to switch to the ground curtain self-cleaning mode. At this time, the high-degree-of-freedom gimbal 42 of the intelligent spray gun 4 is adjusted to a downward angle, pointing towards the ground and drainage ditch. The intelligent spray gun 4 switches to a wide-angle fan-shaped water flow pattern, and the proportional adjustment valve is adjusted to the maximum flow level. All intelligent spray guns 4 synchronously perform reciprocating swing or fan-shaped scanning spraying according to the planned path. The combined force of the water flow pushes the sewage and residual dirt into the drainage ditch, completing the comprehensive cleaning of the cleaning platform and drainage ditch. This ensures that the car wash platform is clean and free of residue, and that the drainage is unobstructed and unblocked, providing a clean and tidy working environment for the next vehicle to be washed. This achieves the automation of the cleaning station's self-maintenance and ensures the continuous and efficient operation of the equipment.

[0060] This step achieves a self-cleaning closed loop in the cleaning process, avoiding secondary pollution caused by dirt residue, while ensuring the stability of subsequent equipment operation and extending the equipment's service life.

[0061] It should be understood that, although this embodiment Figure 1 The steps are shown sequentially as indicated by the arrows, but they are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are performed; they can be executed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0062] Example 2 Based on the intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun in Embodiment 1, this embodiment discloses an intelligent vehicle tire cleaning system based on fusion perception and intelligent spray gun, such as... Figure 4 As shown, the intelligent vehicle tire cleaning system based on fusion perception and intelligent spray guns includes: a cleaning system construction module 401, a pre-cleaning module 402, a contour-following fine cleaning module 403, and a floor curtain self-cleaning module 404, wherein: The cleaning system building module 401 is used to build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit.

[0063] The pre-cleaning module 402 is used when a vehicle enters. The global perception subsystem perceives the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-cleaning mode, drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-cleaning the same side tire.

[0064] The contour-following fine cleaning module 403 is used after the vehicle has come to a complete stop. The global perception subsystem acquires multi-source data and integrates it to generate a fine three-dimensional model of the tire and a dirt distribution map. The central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans the path based on the fine three-dimensional model and dirt distribution map, and then controls the intelligent spray gun to execute the contour-following fine cleaning mode.

[0065] The ground curtain self-cleaning module 404 is used to identify the residual dirt areas on the cleaning platform and drainage ditch and output contour data when the vehicle leaves; the central intelligent decision unit plans the spraying path according to the contour data and controls all intelligent spray guns to switch to the ground curtain self-cleaning mode for synchronous spraying.

[0066] In this embodiment, the specific working process and working principle of the cleaning system construction module 401, pre-cleaning module 402, contour-following fine cleaning module 403, and floor curtain self-cleaning module 404 are the same as those in Embodiment 1, and therefore will not be described again in this embodiment. Each unit module can be implemented entirely or partially through software, hardware, or a combination thereof. Each unit module can be embedded in or independent of the processor in the computer device in hardware form, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above unit modules.

[0067] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0069] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for intelligent vehicle tire cleaning based on fusion sensing and intelligent spray guns, characterized in that, The method includes: Step 1: Build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit; Step 2: When the vehicle enters, the global perception subsystem senses the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-washing mode, drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-washing the same side tire. Step 3: After the vehicle comes to a complete stop, the global perception subsystem acquires multi-source data and fuses it to generate a detailed three-dimensional model of the tire and a dirt distribution map; the central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans a path based on the detailed three-dimensional model and dirt distribution map, and then controls the intelligent spray gun to perform contour-following fine cleaning mode. Step 4: When the vehicle leaves, the global perception subsystem identifies the area of ​​residual dirt on the cleaning platform and drainage ditch and outputs contour data; the central intelligent decision-making unit plans the spraying path based on the contour data and controls all intelligent spray guns to switch to the ground curtain self-cleaning mode for synchronous spraying.

2. The intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun according to claim 1, characterized in that, In step 1, the global perception subsystem is an anti-interference fusion architecture built with lidar, industrial cameras, and auxiliary sensors; The intelligent spray gun execution subsystem includes intelligent spray guns symmetrically deployed on both sides of the cleaning channel, and each intelligent spray gun is equipped with a lifting component and a high degree of freedom gimbal. The central intelligent decision-making unit incorporates an adaptive cleaning model and a collaborative controller.

3. The intelligent vehicle tire cleaning method based on fused perception and intelligent spray gun according to claim 1, characterized in that, In step 2, the collected vehicle point cloud data and image data are fused to extract wheel contour and position feature information, and combined with the vehicle speed, real-time dynamic parameters are output, including: After fusing the collected vehicle point cloud data and image data, the Kalman filter algorithm is used to track and estimate the displacement between consecutive frames. The tire rotation speed and vehicle speed are calculated using the time series difference method. The calculation formula is as follows: ； ； ； In the formula, This represents the average displacement change of the wheel axle between two adjacent sampling times; express The position of the wheel axle at any given moment; express The position of the wheel axle at any given moment; Indicates the sampling period; Indicates the vehicle's speed; Indicates the angular velocity of the tire rotation; This represents the change in tire angle between consecutive frames; By fitting the center coordinates of the wheel profile using Hough transform and combining them with the spatiotemporal calibration results of the auxiliary sensor, the three-dimensional coordinates of the wheel in the global coordinate system are calculated. The three-dimensional coordinates and vehicle speed and tire rotational angular velocity As a real-time dynamic parameter output.

4. The intelligent vehicle tire cleaning method based on fused perception and intelligent spray gun according to claim 3, characterized in that, In step 2, the central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters. When allocating tasks according to the cross-tracking pre-washing mode, it uses a spatiotemporal coupling scheduling algorithm to calculate the spatiotemporal window for each intelligent spray gun to intervene and exit the cleaning process, combining the real-time dynamic parameters of the wheels. The calculation formula is as follows: ； ； In the formula, Indicates when the intelligent spray gun intervenes in the cleaning process; Indicates the preset cleaning start distance; Indicates when the intelligent spray gun exits the cleaning process; This indicates the cleaning time, which is dynamically adjusted based on the degree of dirt on the tires.

5. The intelligent vehicle tire cleaning method based on fused perception and intelligent spray gun according to claim 3, characterized in that, In step 2, when pre-washing the tires on the same side, the downstream intelligent spray gun starts at the following time: ； In the formula, Indicates the start-up time of the downstream intelligent spray gun; Indicates the start-up time of the upstream intelligent spray gun; Indicates the tire radius.

6. The intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun according to any one of claims 2 to 5, characterized in that, Step 3, the process by which the global perception subsystem acquires multi-source data and fuses it to generate a detailed 3D model of the tire and a dirt distribution map, includes: Step 301: Obtain multi-view point cloud data of the tire surface by scanning with LiDAR, and collect texture information of the tire surface by using an industrial camera; take the multi-view point cloud data and the texture information as input, perform registration and fusion by ICP algorithm, and output fused point cloud data in a unified coordinate system. Step 302: The fused point cloud data is used as input, and point cloud noise is removed by voxel filtering and the point cloud surface geometry is optimized by Poisson reconstruction to output a fine three-dimensional model of the tire. Step 303: Using the fused point cloud data as input, identify the attachment regions in the fused point cloud data based on a deep learning semantic segmentation network, quantify the pollution density of each region, and generate a pixel-by-pixel pollution heat map. Step 304: Align the pixel-by-pixel dirt heat map with the fine 3D model of the tire in spatial coordinates, and output a dirt distribution map containing tire geometric features and dirt distribution information.

7. The intelligent vehicle tire cleaning method based on fused perception and intelligent spray gun according to claim 6, characterized in that, In step 3, the central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans a path based on the detailed three-dimensional model and the dirt distribution map, including: Step 311: Extract the pollution density of each area based on the pollution distribution map. Based on a preset pollution density threshold, the pollution density of each area is compared to identify high-pollution areas and conventionally polluted areas. Step 312: Based on the detailed 3D model of the tire, for areas with conventional pollution, at sampling intervals... Extract the coordinates of contour points along the tire circumference to form an ordered point set. For highly polluted areas, the sampling interval is shortened to perform encrypted extraction and output an encrypted ordered point set. Step 313: After discretizing the ordered point set and the encrypted ordered point set, the Bézier curve is used for fitting and optimization to generate a continuous and differentiable spray gun motion trajectory, and the cleaning path planning result is output.

8. The intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun according to any one of claims 1 to 5, characterized in that, Step 3 also includes: after cleaning is completed, the global perception subsystem feeds back the cleanliness characteristics and determines whether to perform incremental or targeted enhanced cleaning based on the preset cleanliness threshold.

9. The intelligent vehicle tire cleaning method based on fusion perception and intelligent spray gun according to any one of claims 2 to 5, characterized in that, In step 4, the global perception subsystem identifies the areas of residual dirt on the cleaning platform and in the drainage ditch and outputs contour data, including: Step 401: Collect point cloud data of the cleaning platform and drainage ditch using lidar, perform filtering processing, and output effective point cloud data of the ground area. Step 402: Based on the effective point cloud data of the ground area, the high humidity mud and water residue area is segmented according to the preset reflection intensity threshold, and the boundary of the high humidity mud and water residue area is identified by combining the height difference from the ground. Step 403: Cluster the high-humidity mud and water residue area using Euclidean clustering algorithm, extract the outer contour coordinates of each dirt area, and output the contour data of the dirt residue area.

10. A vehicle tire intelligent cleaning system based on fusion perception and intelligent spray gun, characterized in that, The system includes: The cleaning system construction module is used to build a global perception subsystem, an intelligent spray gun execution subsystem, and a central intelligent decision-making unit; The pre-cleaning module is used when a vehicle enters. The global perception subsystem perceives the vehicle's motion status in real time, fuses the collected vehicle point cloud data and image data, extracts wheel contour and position feature information, and outputs real-time dynamic parameters in combination with the vehicle's speed. The central intelligent decision-making unit predicts the tire trajectory based on the real-time dynamic parameters, allocates tasks according to the cross-tracking pre-cleaning mode, and drives the intelligent spray gun to cross-clean the key areas of the opposite tire and take over pre-cleaning the same side tire. The contour-following fine cleaning module is used after the vehicle has come to a complete stop. The global perception subsystem acquires multi-source data and fuses it to generate a fine three-dimensional model of the tire and a dirt distribution map. The central intelligent decision-making unit discretizes the three-dimensional contour of the tire and plans a path based on the fine three-dimensional model and dirt distribution map, and then controls the intelligent spray gun to execute the contour-following fine cleaning mode. The self-cleaning module is used when the vehicle leaves. The global perception subsystem identifies the area of ​​residual dirt on the cleaning platform and drainage ditch and outputs contour data. The central intelligent decision-making unit plans the spraying path based on the contour data and controls all intelligent spray guns to switch to the self-cleaning mode for synchronous spraying.

Images