Ant colony algorithm optimization and system for drainage channel of waterproof cabinet

Abstract

The invention provides an ant colony algorithm optimization method and system for a drainage channel of a waterproof cabinet. According to the method, the double-source heterogeneous data in the drainage channel of the waterproof cabinet is collected through visual correction. And extracting a viscosity coefficient in the humidity parameters, fusing the viscosity coefficient with the light image to generate an adhesion characteristic spectrum, and constructing a three-dimensional temperature field covering the curved surface according to the adhesion characteristic spectrum. And generating a cleaning track set by applying an ant colony collaborative optimization mechanism in the temperature field, and performing fitting degree verification on the channel structure curved surface. And if the verification result is lower than the threshold value, starting nonlinear self-calibration to correct the motion vector deviation. And channel deformation intervals are distinguished based on the deviation, the dirt deposition density and the deformation intervals are associated to predict the dirt expansion trend, a preventive maintenance operation chain of a pressure self-adaptive fused ant colony algorithm is triggered, and global optimization of a cleaning path is achieved. The cleaning efficiency of the drainage channel of the waterproof cabinet is effectively improved, dirt accumulation is prevented through global path optimization, and long-acting operation of the system is ensured.

Description

Technical Field

[0001] This application relates to the fields of intelligent cleaning system technology, pipeline maintenance optimization and artificial intelligence algorithm application technology, and in particular to an ant colony algorithm optimization method and system for the drainage channel of a waterproof cabinet. Background Technology

[0002] In high-reliability environments such as power equipment and communication base stations, the drainage channels of waterproof cabinets are constantly exposed to harsh conditions characterized by dampness and easy accumulation of dirt. Their internal structures are often complex, curved surfaces, making it difficult to completely remove adhesive dirt (such as sludge and biofilm) using conventional cleaning methods. The technological requirements lie in achieving precise sensing of dirt within the curved channels, dynamic optimization of cleaning paths to adapt to structural deformation, and predictive maintenance based on dynamic dirt deposition trends, in order to ensure drainage efficiency and extend equipment lifespan.

[0003] The most advanced solutions currently available employ laser cleaning robot systems based on structured light 3D scanning. This system first uses a laser scanner to create a high-precision geometric model of the interior of the drainage channel of the waterproof cabinet, generating an initial 3D structural topology. Then, it plans the movement path of the robotic arm according to preset coverage rules (such as fixed-interval spiral trajectories or gridded paths). During execution, onboard pressure sensors monitor the contact pressure between the cleaning tool and the channel wall in real time, preventing mechanical damage through closed-loop control.

[0004] This solution aims to automate cleaning and reduce the frequency of human intervention. Despite its automation capabilities, the system still has significant limitations: First, its path planning relies solely on a static geometric model, failing to incorporate dynamic environmental parameters such as dirt viscosity, humidity gradient, and temperature distribution. This results in the cleaning intensity being unable to adapt to the actual adhesion state of highly viscous dirt, leading to incomplete removal. Second, the robotic arm's motion trajectory is generated based on a rigid model from the initial scan, making it unable to respond to micron-level deformations in the channel caused by material aging or external stress, resulting in failure of the cleaning tool to adhere to the deformed surface. Third, it lacks the ability to model the relationship between dirt deposition density and structural deformation, and can only passively perform periodic cleaning tasks. It cannot predict the expansion trend of local dirt, nor can it proactively trigger preventative maintenance strategies, posing a risk of maintenance lag. Summary of the Invention

[0005] This application provides an ant colony algorithm optimization method and system for the drainage channel of a waterproof cabinet, which solves the problem in the prior art that the cleaning intensity cannot adapt to the actual adhesion state of highly viscous dirt, resulting in incomplete removal.

[0006] Firstly, this application provides an ant colony algorithm optimization method for the drainage channel of a waterproof cabinet, including: The dual-source heterogeneous data inside the drainage channel of the waterproof cabinet is acquired through visual correction. The dual-source heterogeneous data includes multi-angle light images and synchronously changing humidity parameter groups. The viscosity coefficient in the humidity parameter set is extracted, and the viscosity coefficient is fused with the multi-angle light image to generate an adhesion feature map. Based on the adhesion feature map, a three-dimensional temperature field covering the curved surface of the drainage channel of the waterproof cabinet is constructed. An ant colony cooperative optimization mechanism is applied in the three-dimensional temperature field to generate a set of clean trajectories under physical constraints. The fit of the set of clean trajectories with the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field is verified to obtain the corresponding verification results. When the verification result is lower than a preset threshold, a nonlinear self-calibration process is initiated to correct the motion vector deviation of the cleaning trajectory set, thereby obtaining the corrected motion vector deviation. The deformation range of the channel structure is distinguished based on the motion vector deviation, and the dirt deposition density in the three-dimensional temperature field is correlated with the deformation range to predict the dirt expansion trend. At the same time, a preventive maintenance operation chain with the swarm intelligence optimization characteristics of the pressure-adaptive ant colony algorithm is triggered to achieve global optimization of the cleaning path of the drainage channel of the waterproof cabinet.

[0007] Optionally, an ant colony cooperative optimization mechanism is applied in the three-dimensional temperature field to generate a set of clean trajectories under physical constraints, including: Multiple virtual agent points are defined in a three-dimensional temperature field, each virtual agent point representing a cleaning unit, and the starting position of the virtual agent points is initialized. A movement rule is set for each virtual agent point, and the movement probability is calculated based on the movement rule and the temperature value of the three-dimensional temperature field. Based on physical constraints including channel curvature limitations and cleaning tool size limitations, the physical constraints are added to the movement rules to adjust the movement probability; Through an iterative process, the virtual agent point moves in the three-dimensional temperature field according to the movement probability, and the movement probability is updated after each movement; Record the movement paths of all virtual agent points. When the number of iterations reaches a preset value, output the set of movement paths as the set of cleaning trajectories under the physical constraints.

[0008] Optionally, the fit between the set of cleaning trajectories and the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field is verified to obtain the corresponding verification results, including: Spatial coordinates are extracted from the three-dimensional temperature field, and a three-dimensional point cloud representation of the channel structure surface of the spatial coordinates is constructed. The coordinates of each cleaning trajectory point are obtained from the cleaning trajectory set, and the cleaning trajectory points are mapped onto the three-dimensional point cloud; Calculate the distance between each cleaning trajectory point and the corresponding point on the curved surface of the channel structure, and statistically analyze the distance values ​​of all cleaning trajectory points to calculate the average distance value and the maximum distance value; Based on the average distance value and the maximum distance value, the fit is verified by a comparison function and the corresponding verification result is output. The verification result indicates the degree of fit between the cleaning trajectory set and the channel structure surface.

[0009] Optionally, when the verification result is lower than a preset threshold, a nonlinear self-calibration process is initiated to correct the motion vector deviation of the cleaning trajectory set, resulting in a corrected motion vector deviation, including: When the verification result is lower than the preset threshold, identify the motion vector representing the direction and speed of cleaning movement for each cleaning trajectory point in the cleaning trajectory set; Based on the verification results, the corresponding correction coefficients are calculated using a nonlinear function. The motion vector is then adjusted according to the correction coefficients, and the correction coefficients are applied to the motion vector to update the direction and velocity components of the motion vector. Through iterative optimization, the updated motion vectors satisfy the geometric constraints of the channel structure surface, and the updated motion vectors of all clean trajectory points are output as the corrected motion vector deviations.

[0010] Optionally, the deformation range of the channel structure is distinguished based on the motion vector deviation, and the fouling deposition density in the three-dimensional temperature field is correlated with the deformation range to predict the fouling expansion trend, including: The deviation value of each cleaning trajectory point is extracted from the motion vector deviation, and the channel structure is divided into multiple continuous intervals according to the magnitude of the deviation value, with each interval defined as a deformation interval; The dirt deposition density value at each point is obtained by converting the temperature value in the three-dimensional temperature field. Map the dirt deposition density value onto the deformation range, and calculate the average dirt deposition density for each deformation range; Based on the average dirt deposition density, the dirt growth rate for each deformation interval is calculated using a trend prediction model that considers time factors and interval geometric characteristics, and the dirt growth rate is output as the dirt expansion trend.

[0011] Optionally, the viscosity coefficient is extracted from the humidity parameter set, and the viscosity coefficient is fused with the multi-angle light image to generate an adhesion feature map, including: Multiple humidity measurements are obtained from the humidity parameter set, and the viscosity coefficient is derived based on the humidity measurements using a predefined calculation rule based on the physical relationship between humidity and the adhesion properties of substances. For multi-angle light images, the brightness change information of each pixel in the multi-angle light image is identified, and the brightness change information is converted into surface texture values; The viscosity coefficient is combined with the surface texture value, and the viscosity coefficient is assigned to the corresponding pixel position of the multi-angle light image through a position matching operation to generate a fused data layer. On the fused data layer, the weighted sum of the viscosity coefficient and the surface texture value is calculated for each pixel location to obtain the adhesion strength value at each location; Based on the adhesion strength values ​​at all pixel locations, a two-dimensional mesh structure is constructed, which covers the observation surface of the drainage channel of the waterproof cabinet, and outputs it as an adhesion feature map.

[0012] Optionally, a three-dimensional temperature field covering the curved surface of the drainage channel of the waterproof cabinet is constructed based on the adhesion feature map, including: The adhesion strength value at each location is obtained from the adhesion feature map, and the adhesion strength value is input into the heat conduction model that simulates the heat distribution on the curved surface of the drainage channel of the waterproof cabinet; In the heat conduction model, a heat source influence factor representing the degree of influence of dirt deposition on local temperature is calculated at each location based on the adhesion strength value; Based on the heat source influence factor, a spatial mapping operation is performed according to the curved geometry of the drainage channel of the waterproof cabinet to convert the two-dimensional mesh structure of the attachment feature map into three-dimensional spatial coordinates. On the three-dimensional spatial coordinates, each point is assigned a temperature value calculated using the heat source influence factor of adjacent points; By combining the temperature values ​​assigned on the three-dimensional spatial coordinates, a three-dimensional grid containing the spatial coordinates and corresponding temperature values ​​of all points is constructed to form a three-dimensional temperature field.

[0013] Secondly, this application provides an ant colony algorithm optimization system for the drainage channel of a waterproof cabinet, including: The acquisition module is used to acquire dual-source heterogeneous data inside the drainage channel of the waterproof cabinet through visual correction. The dual-source heterogeneous data includes multi-angle light images and synchronously changing humidity parameter groups. The generation module is used to extract the viscosity coefficient from the humidity parameter group, fuse the viscosity coefficient with the multi-angle light image to generate an adhesion feature map, and construct a three-dimensional temperature field covering the curved surface of the drainage channel of the waterproof cabinet based on the adhesion feature map. The verification module is used to apply an ant colony cooperative optimization mechanism in the three-dimensional temperature field to generate a set of clean trajectories under physical constraints, and to verify the fit between the set of clean trajectories and the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field to obtain the corresponding verification results. The correction module is used to initiate a nonlinear self-calibration process to correct the motion vector deviation of the cleaning trajectory set when the verification result is lower than a preset threshold, so as to obtain the corrected motion vector deviation. The prediction module is used to distinguish the deformation range of the channel structure based on the motion vector deviation, and to correlate the dirt deposition density in the three-dimensional temperature field with the deformation range to predict the dirt expansion trend. At the same time, it triggers a preventive maintenance operation chain with pressure-adaptive ant colony algorithm swarm intelligence optimization characteristics to achieve global optimization of the cleaning path of the drainage channel of the waterproof cabinet.

[0014] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the ant colony algorithm optimization method for the drainage channel of a waterproof cabinet as described in the first aspect above.

[0015] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements an ant colony algorithm optimization method for the drainage channel of a waterproof cabinet as described in the first aspect.

[0016] This application achieves fusion perception of optical features and humidity dynamics by simultaneously acquiring dual-source heterogeneous data consisting of multi-angle light images and humidity parameter sets within the channel through visual correction. Furthermore, it extracts the viscosity coefficient from the humidity parameters and fuses it with the light images to generate an adhesion feature map, quantifying the dirt adhesion characteristics. Based on this map, a three-dimensional temperature field covering the curved surface is constructed to dynamically map the microscopic thermodynamic environment of dirt deposition. An ant colony collaborative optimization mechanism is applied in the temperature field to generate a set of physically constrained cleaning trajectories, overcoming the bottleneck of path optimization under the constraints of complex geometric structures and thermodynamic coupling. By verifying the fit between the trajectory and the curved surface of the channel structure, path deviations caused by deformation are detected in real time. When the fit is insufficient, a nonlinear self-calibration process is triggered to correct the motion vector, significantly improving the system's adaptability to structural deformation. Finally, by correlating dirt deposition density with the deformation range, the dirt expansion trend is predicted, triggering a pressure-adaptive preventive maintenance operation chain that integrates ant colony algorithm swarm intelligence. This achieves global collaborative optimization of the cleaning path while proactively avoiding blockage risks, systematically solving the problems of incomplete cleaning and delayed maintenance caused by the one-sided environmental perception, static and rigid models, and lack of prediction mechanisms in traditional methods.

[0017] Furthermore, an ant colony cooperative optimization method is applied in a three-dimensional temperature field to generate a set of cleaning trajectories. Specifically, this includes: initializing virtual agent points representing cleaning units in the three-dimensional temperature field and setting movement probability rules that incorporate the influence of the temperature field; converting physical constraints into probability weights to adjust the movement strategy based on channel curvature limitations and cleaning tool size limitations; updating the probability through multiple iterations and recording the agent point paths; and outputting a set of cleaning trajectories that satisfy complex geometric constraints after iteration convergence. Through the dynamic probability optimization mechanism of virtual agent points, physical constraints such as channel curvature and tool size are encoded into probability weight functions of the ant colony algorithm. Combined with temperature field gradient iterative optimization of the movement path, this effectively solves the problems of insufficient coverage of local dead zones and path conflicts with physical conditions in curved structures, as seen in traditional algorithms, and outputs cleaning trajectories with high conformity to actual engineering constraints.

[0018] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart of an ant colony algorithm optimization method for a waterproof cabinet drainage channel provided in this application is shown; Figure 2 This paper presents a schematic diagram of the structure of an ant colony algorithm optimization system for a waterproof cabinet drainage channel provided in this application; Figure 3 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0022] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0023] Current mainstream cleaning technologies for waterproof tank drainage channels, such as systems utilizing laser scanning and robotic arms, have several key shortcomings. First, they focus solely on the shape of the pipes themselves, failing to effectively detect significant changes in the internal environment, especially the actual viscosity of dirt in different locations (e.g., whether oil is more difficult to remove). Second, these systems have fixed cleaning routes; if the pipes deform slightly due to temperature changes or aging, the cleaning tools may not adhere properly to the pipe surface, leaving cleaning blind spots. Finally, they operate solely according to a preset schedule, lacking the ability to assess the rate of dirt accumulation and predict which locations are prone to blockage. Problems are only discovered when water flow is obstructed, making maintenance highly reactive. The root cause lies in the fact that existing systems are "invisible" to the characteristics of dirt, "unable to sense" subtle changes in the pipes, and unable to "calculate" where the risk of blockage lies.

[0024] To address the aforementioned challenges, this invention provides a drainage channel cleaning optimization method that intelligently adapts to the characteristics of dirt and changes in pipes. The core of this method lies in the collaborative operation of machine vision and humidity sensors, essentially adding "eyes" and "skin" to the system. This allows the system to not only clearly see the internal structure of the pipe but also perceive the viscosity and humidity of dirt in various locations in real time. Then, an intelligent algorithm simulating "ant colony collaboration to find the optimal path" (ant colony algorithm) is introduced. This algorithm integrates information such as the current curvature of the pipe, the detected distribution and viscosity of dirt, and whether there is any deformation, to dynamically calculate the optimal cleaning route and intensity in real time. Simultaneously, the system can analyze the trend of dirt accumulation, issuing early warnings for potential blockages and focusing on cleaning those areas. The benefits are significant: it automatically adjusts the cleaning intensity based on the viscosity of the dirt, ensuring that stubborn dirt is thoroughly removed; when the pipe undergoes slight deformation, the system immediately adjusts the trajectory of the cleaning tool to ensure a tight fit without leaving any blind spots; and it proactively identifies potential blockage risks, addressing them before the problem worsens, transforming maintenance from "remedial action" to "prevention." This approach effectively overcomes the shortcomings of existing technologies in terms of cleanliness adaptability, deformation response capability, and maintenance predictability through dynamic sensing and intelligent computing.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Figure 1 This application provides a flowchart of an ant colony algorithm optimization method for the drainage channel of a waterproof cabinet, as shown in the embodiments of this application. Figure 1 As shown, the method includes: 101. Obtain dual-source heterogeneous data inside the drainage channel of the waterproof cabinet through visual correction, wherein the dual-source heterogeneous data includes multi-angle light images and synchronously changing humidity parameter groups; In the above scheme, visual correction acquisition refers to the process of automatically adjusting the position and angle of the camera equipment through image analysis technology to ensure that the captured image is clear and distortion-free. The drainage channel of a waterproof cabinet specifically refers to the pipe structure at the bottom of industrial equipment used to drain accumulated water. Dual-source heterogeneous data represents two different types of data acquired simultaneously, including light images and humidity parameters, with different data structures and acquisition methods. Multi-angle light images refer to visible light photographs taken from multiple directions, such as the top of the front side of the drainage channel. The humidity parameter set refers to a sequence of humidity values ​​synchronously recorded by multiple sensors, with each data point including a timestamp and location marker.

[0027] In this embodiment, visual correction and acquisition are first performed using a robotic arm equipped with a high-definition camera. A feature matching algorithm is used to identify the contour of the drainage channel. When the initial position deviates, the system automatically calculates the movement and drives the robotic arm to adjust to align with the channel opening. Secondly, dual-source heterogeneous data is acquired simultaneously. After correction, three light images are captured from 0 degrees frontal view, 45 degrees side view, and 90 degrees top view. Simultaneously, ten humidity sensors installed in the channel are triggered to record data five times per second. A time synchronization signal is generated at the moment the images are captured to ensure accurate correspondence between the humidity readings and the image. Finally, a heterogeneous data package is generated, integrating the three images and the humidity data within the corresponding time window into a structured data file. For example, the top view and the humidity sequence from 0.1 seconds before to 0.1 seconds after the capture moment are merged and marked as the same data source.

[0028] In a practical application, during the inspection of a waterproof electrical distribution cabinet at Factory A, after the operator initiated the inspection program, the system automatically controlled a robotic arm to move the camera to the inlet of the drainage channel, eliminating any previous positional deviations through visual correction. Subsequently, the camera captured images of the channel's interior from three fixed angles; for example, a top-down view clearly captured traces of deposited silt. Simultaneously, the acquired humidity data showed an abnormal increase in the P5 value of the position sensor, reaching 70%. All data was packaged, labeled with inspection point number Q3, and uploaded to a cloud-based analysis platform.

[0029] This solution ensures stable and usable image quality through visual correction, eliminating the risk of misjudgment caused by angular deviation; precise synchronization of dual-source data establishes the correlation between physical features and humidity changes, for example, the cracks shown in the image and the corresponding humidity abrupt change form a corresponding evidence chain; structured data packages provide a cross-modal fusion analysis basis for subsequent intelligent diagnosis.

[0030] 102. Extract the viscosity coefficient from the humidity parameter group, fuse the viscosity coefficient with the multi-angle light image to generate an adhesion feature map, and construct a three-dimensional temperature field covering the curved surface of the drainage channel of the waterproof cabinet based on the adhesion feature map. Optionally, step 102 may specifically include the following steps: 1021. Obtain multiple humidity measurement values ​​from the humidity parameter group, and derive the viscosity coefficient based on the humidity measurement values ​​through a predefined calculation rule based on the physical relationship between humidity and material adhesion properties; 1022. For a multi-angle light image, identify the brightness change information of each pixel in the multi-angle light image, and convert the brightness change information into a surface texture value; 1023 The viscosity coefficient is combined with the surface texture value, and the viscosity coefficient is assigned to the corresponding pixel position of the multi-angle light image through a position matching operation to generate a fused data layer; 1024. On the fused data layer, calculate the weighted sum of the viscosity coefficient and the surface texture value for each pixel location to obtain the adhesion strength value for each location; 1025. Based on the adhesion strength values ​​at all pixel locations, a two-dimensional mesh structure is constructed, which covers the observation surface of the drainage channel of the waterproof cabinet, and outputs it as an adhesion feature map.

[0031] In the above scheme, the viscosity coefficient is a value calculated from humidity parameters that reflects the liquid's adhesion characteristics; a higher value indicates that the liquid adheres more easily to the surface. The surface texture value is a quantitative index of surface roughness extracted from changes in image brightness; a greater change in brightness results in a higher value. The adhesion feature map is a two-dimensional color image formed by fusing the viscosity coefficient and surface texture, used to visualize the adhesion intensity at various locations. The three-dimensional temperature field is a three-dimensional temperature distribution model constructed based on the adhesion features, covering the curved inner wall of the drainage channel.

[0032] In this embodiment of the application, firstly, through step 1021, the humidity measurement values ​​of ten measurement positions are extracted from the humidity parameter group, such as position P5 with a humidity of 65%. The viscosity coefficient is calculated based on a physical formula based on the adhesion properties of humidity and substances: viscosity coefficient = humidity value × α + β, where α = 0.05 and β = 0.2. For example, 65 × 0.05 + 0.2 = 3.45.

[0033] Secondly, in step 1022, pixel-level analysis is performed on the three multi-angle light images to identify the brightness change information of each pixel in the multi-angle light images, and the brightness gradient algorithm is used to convert the brightness change into a surface texture value of 0-1. For example, if a sudden change in brightness is detected in the rust spot area, the texture value is calculated to be 0.85.

[0034] Next, in step 1023, the viscosity coefficient is combined with the surface texture value, the sensor position is matched with the image coordinates, such as P5 being mapped to image coordinates X150, Y300, and the viscosity coefficient 3.45 is assigned to the corresponding pixel to generate the fusion data layer.

[0035] Subsequently, in step 1024, a weighted sum of adhesion strength values ​​is calculated for each pixel: viscosity coefficient × 0.6 + surface texture value × 0.4, for example, 3.45 × 0.6 + 0.85 × 0.4 = 2.07 + 0.34 = 2.41.

[0036] Finally, in step 1025, the adhesion intensity of all pixels is constructed into a two-dimensional grid matrix, such as the bending intensity of 2.41 marked in red, and the adhesion feature map covering the surface of the channel is output.

[0037] In a practical application, in the factory A inspection case: the system first obtains humidity values ​​ranging from 60% to 70% from five sensors in the middle section of the drainage channel, and calculates the viscosity coefficient to be 3.1 to 3.7 using a formula. Simultaneously, analysis of the three-angle images reveals a texture value of 0.8 at the channel corner. After matching, the adhesion strength of 3.32 is calculated at the corner coordinates, generating a two-dimensional mesh map showing the corner area as a red high-adhesion zone. This map is used as input to construct a three-dimensional temperature field, showing that the corner area is three degrees warmer than the surrounding area.

[0038] This solution reveals adhesion characteristics through humidity parameters, captures surface conditions through optical images, and fuses them to generate a visual adhesion feature map that intuitively displays the problem area. Finally, the constructed three-dimensional temperature field provides a three-dimensional thermodynamic model for predicting blockage and leakage risks.

[0039] 103. An ant colony cooperative optimization mechanism is applied in the three-dimensional temperature field to generate a set of clean trajectories under physical constraints, and the fit of the set of clean trajectories with the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field is verified to obtain the corresponding verification results.

[0040] Optionally, step 103 may specifically include the following steps: 1031. Define multiple virtual agent points in a three-dimensional temperature field, each virtual agent point representing a cleaning unit, and initialize the starting position of the virtual agent points; 1032. Set movement rules for each virtual agent point, and calculate the movement probability based on the movement rules and the temperature value of the three-dimensional temperature field; 1033. Based on physical constraints including channel curvature limitations and cleaning tool size limitations, the physical constraints are added to the movement rules to adjust the movement probability; 1034. Through an iterative process, the virtual agent point moves in the three-dimensional temperature field according to the movement probability, and the movement probability is updated after each movement; 1035. Record the movement paths of all virtual agent points. When the number of iterations reaches a preset value, output the set of movement paths as the set of clean trajectories under the physical constraints.

[0041] In the above scheme, virtual agent points refer to the calculated points that simulate the movement path of cleaning tools, with each point representing a cleaning unit. Movement rules are probabilistic calculation mechanisms that define the movement direction of agent points. Physical constraints include channel curvature limits such as maximum turning angles and cleaning tool size limits such as minimum passing diameter. The cleaning trajectory set is the integrated result of all agent point movement paths. Fit verification involves comparing the generated trajectories with the actual channel structure to check the feasibility of the path.

[0042] In this embodiment of the application, ten virtual agent points are first defined and their positions are initialized in the three-dimensional temperature field model through step 1031. For example, all points are evenly distributed at the entrance of the channel with coordinates X0, Y0, and Z0 starting from zero.

[0043] Subsequently, in step 1032, movement rules are set for each point, and the movement probability is calculated based on the temperature. The formula is: movement probability = 0.3 + (target area temperature - average temperature) × 0.01. For example, the probability of high temperature area is increased to twice the base value. For example, when a corner area with a temperature of 32 degrees Celsius is detected, the probability increases from 0.3 to 0.6.

[0044] Next, physical constraints are added in step 1033. When the direction of movement exceeds the maximum turning angle of 60 degrees allowed by the channel, the probability is forced to zero. For example, if a point attempts to move in the 85-degree direction, it will be intercepted by the system.

[0045] Next, iterative movement is performed through step 1034. The virtual agent points gradually cover the high-temperature area according to the updated probability. For example, after a point stops at a corner, the probability of that area drops to 1.3 times the base value, guiding other points to enter. Finally, all trajectories are recorded through step 1035. After 20 iterations, ten paths are output, such as trajectory T1 forming a spiral coverage at the bend, which are integrated into a clean trajectory set.

[0046] In practical applications, during the inspection of drainage channels in Factory A, the system initialized ten proxy points at the channel entrance. Based on the three-dimensional temperature field, the corner area was identified as a high-temperature zone and marked as a risk area. During the movement, the proxy points preferentially moved towards the high-temperature zone. However, due to a 55-degree sharp bend in the channel (with physical constraints limiting the maximum bend to 60 degrees), one trajectory was automatically corrected into a segmented turning path. The final output trajectory set was verified to fit the channel surface perfectly. Eight paths completely fit the structure, such as trajectory T2, which perfectly matches the bend. Two paths were marked as not fitting due to the tool size being too large in narrow locations.

[0047] This solution simulates the cleaning process using virtual agent points, prioritizes high-risk areas based on temperature field characteristics, ensures path feasibility through physical constraints, and verifies the fit between the path and the physical structure to provide an executable solution for automated cleaning.

[0048] 104. When the verification result is lower than a preset threshold, a nonlinear self-calibration process is initiated to correct the motion vector deviation of the cleaning trajectory set, thereby obtaining the corrected motion vector deviation; Optionally, step 104 may specifically include the following steps: 1041. When the verification result is lower than the preset threshold, identify the motion vector representing the direction and speed of cleaning movement for each cleaning trajectory point in the cleaning trajectory set; 1042. Based on the verification results, the corresponding correction coefficient is calculated using a nonlinear function. The motion vector is adjusted according to the correction coefficient, and the correction coefficient is applied to the motion vector to update the direction and velocity components of the motion vector. 1043. Through iterative optimization, the updated motion vectors satisfy the geometric constraints of the channel structure surface, and the updated motion vectors of all clean trajectory points are output as the corrected motion vector deviations.

[0049] In the above scheme, the verification result refers to the score of the degree of matching between the cleaning trajectory and the channel structure; the lower the score, the greater the deviation. The motion vector is a mathematical vector describing the direction and velocity of movement of the cleaning trajectory points. Nonlinear self-calibration is the process of dynamically correcting the motion vector through a curve function. The correction coefficient is an adjustment factor calculated based on the verification result, used to update the direction and velocity components of the motion vector.

[0050] In this embodiment of the application, firstly, through step 1041, when the fit verification result is lower than a preset standard, such as 70 points (e.g., trajectory T3 scores 60), the system extracts the motion vector of each point in the trajectory set. For example, the motion vector of trajectory T2 at a certain point in the bend is θ=50°, v=8cm / s.

[0051] Secondly, in step 1042, a correction coefficient is calculated based on the difference in the verification results using a nonlinear function k = 0.3 + ΔS × 0.02. For example, when the fit difference is 10, the function outputs a correction coefficient k = 0.3 + 10 × 0.02 = 0.5. Multiplying this coefficient by the original motion vector, the new direction is adjusted to θ_new = θ × k = 50° × 0.5 = 25°, and the new velocity becomes v_new = v × k = 8 cm / s × 0.5 = 4 cm / s.

[0052] Finally, iterative optimization is performed through step 1043 to check whether the new motion vector meets the channel geometric constraints, such as a minimum turning radius of 10 centimeters. If it does not meet the constraints, the correction coefficient is recalculated. For example, if the constraint is still exceeded after the first adjustment, the direction is corrected to 10 degrees and the speed to 2 centimeters per second after the second iteration, and finally, a set of motion vectors that meet the requirements is output.

[0053] In a practical application, in the factory A inspection case, trajectory T3 achieved a fit score of only 60 points at a narrow bend, below the 70-point threshold. The system identified the trajectory point's motion vector as a direction of 50 degrees and a speed of 8 cm / s. A fit difference of 20 was calculated, corresponding to a correction factor of 0.6, resulting in an adjusted direction of 30 degrees and a speed of 4.8 cm / s. However, the new direction still exceeded the maximum permissible turning angle of 55 degrees. After a second iteration, the direction was corrected to 48 degrees and the speed to 4.3 cm / s, ultimately passing the channel structure verification.

[0054] This solution addresses the mismatch between the path and the physical structure by dynamically adjusting the direction and velocity of the motion vector. Iterative optimization ensures that the final path meets physical constraints, thereby improving the feasibility and safety of the cleaning operation.

[0055] 105. Based on the motion vector deviation, the deformation range of the channel structure is distinguished, and the dirt deposition density in the three-dimensional temperature field is correlated with the deformation range to predict the dirt expansion trend. At the same time, a preventive maintenance operation chain with pressure-adaptive ant colony algorithm swarm intelligence optimization characteristics is triggered to achieve global optimization of the cleaning path of the drainage channel of the waterproof cabinet.

[0056] Optionally, step 105 may specifically include the following steps: 1051. Extract the deviation value of each cleaning trajectory point from the motion vector deviation, and divide the channel structure into multiple continuous intervals according to the magnitude of the deviation value, with each interval defined as a deformation interval; 1052. Based on the temperature values ​​obtained in the three-dimensional temperature field, the dirt deposition density value of each point is obtained by conversion; 1053. Map the dirt deposition density value onto the deformation range, and calculate the average dirt deposition density for each deformation range; 1054. Based on the average dirt deposition density, calculate the dirt growth rate for each deformation interval using a trend prediction model that considers time factors and interval geometric characteristics, and output the dirt growth rate as the dirt expansion trend.

[0057] In the above scheme, the deformation range refers to the channel deformation area divided according to the motion vector deviation value; the larger the deviation, the more severe the deformation. Fouling deposition density is an indicator of the degree of contaminant accumulation converted from temperature values; the higher the temperature, the greater the density. Fouling expansion trend predicts the future rate and scale of fouling growth. The pressure-adaptive fusion ant colony algorithm is a computational model that combines physical pressure parameters with population path optimization for planning preventative maintenance paths.

[0058] In this embodiment, firstly, step 1051 extracts the deviation value of each cleaning trajectory point from the motion vector deviation, and divides the 10-meter channel into three deformation intervals based on the magnitude of the deviation value. For example, the 0.5-meter to 3-meter interval is classified as a high-risk deformation zone because the deviation value exceeds the threshold. Secondly, step 1052 converts the three-dimensional temperature field data and calculates the dirt deposition density using the temperature difference, where the symbol Di represents the dirt deposition density value at location i, the temperature detection value Ti represents the measured temperature data at that location in degrees Celsius, and the temperature conversion coefficient α is a fixed parameter with a default value of 5. ; For example, the density value calculated using the formula is: at a temperature of 32 degrees Celsius. .

[0059] Next, in step 1053, the density values ​​are mapped to the deformation range, and the average dirt deposition density for each segment is calculated, where the symbol is... This represents the average density of all monitoring points within the interval, and is also a dimensionless index. The parameter is the number of monitoring points included in the interval, expressed in units of 3, with a typical value of 3. A point set refers to the set of all points that are divided into interval k.

[0060] ; For example, the average density in high-risk areas: ; Then, in step 1054, a prediction model is established based on the average fouling deposition density, time factor, and bending angle to calculate the fouling growth rate. This represents the monthly increase in dirt within interval k. The density growth factor β is a fixed coefficient set to zero, determining the linear effect of density on the growth rate. Geometric characteristic factor. This depends on the characteristics of the channel structure, especially the bending angle.

[0061] ; For example, the 90-degree bend in high-risk areas leads to a decrease in the monthly growth rate. Ultimately, preventative maintenance is triggered: a three-month cleaning plan is generated using a swarm optimization model, taking into account dirt growth rate and channel pressure limits, such as planning a cleaning path for Wednesday morning to avoid the high water pressure period on Monday.

[0062] In practical application, during the maintenance of the drainage system at Factory A: a motion vector deviation exceeding 40% was detected in the 5-7 meter section of channel 3, classifying it as a deformation zone. The average temperature of this section was 34 degrees Celsius, corresponding to a dirt density of 3.2, with a predicted monthly increase of 0.35 density units based on the bend structure. The system automatically generated a maintenance plan: activating 20 virtual cleaning units, avoiding the peak water pressure period on Thursdays, planning an S-shaped cleaning path around the deformation zone, and simultaneously optimizing the robotic arm's movement speed.

[0063] This solution precisely locates structural risks within deformation ranges, predicts dirt trends to provide early warnings of blockages, and optimizes group paths to balance physical constraints and cleaning efficiency, forming a closed-loop maintenance system to extend equipment lifespan.

[0064] A complete embodiment of steps 101-105 includes: In the task of inspecting the drainage channel of the waterproof distribution cabinet in Factory A, the first step is to control a robotic arm equipped with a high-definition camera to automatically align with the channel entrance through a vision correction acquisition system. Simultaneously, it acquires three-angle light images (0° front view, 45° side view, and 90° top view) and humidity parameter sets from 10 sensors, such as 68% humidity at location P5. Then, the humidity parameters are extracted to calculate the viscosity coefficient 68×0.05+0.2=3.6, and the surface texture values ​​generated from the light images are fused, such as a rust area texture value of 0.9. The adhesion strength value 3.6×0.6+0.9×0.4=2.52 is then calculated through weighted averages to construct a three-dimensional temperature field model covering the curved surface of the channel. Finally, 10 virtual agent points are initialized in the temperature field, and an ant colony optimization mechanism is applied. A set of cleaning trajectories is generated. For example, trajectory T1 forms a spiral path at a 90° bend. When trajectory T3 scores only 60 points in the fit verification in a narrow area, which is below the 70-point threshold, nonlinear self-calibration is triggered. The motion vector of this point is extracted with a 50° direction and a speed of 8 cm / s, and updated to a 30° direction and a speed of 4.8 cm / s with a correction coefficient of 0.6. Finally, the deformation range is divided according to the deviation of the motion vector. For example, the deviation of 0.45 in the 7-8 meter section is classified as a high-risk area. Combined with the average temperature of 34℃ in this area, the dirt deposition density is converted to 34-20 / 5=2.8. The monthly growth rate is predicted to be 2.8×0.1+0.1=0.38, and the pressure adaptive optimization model is triggered to generate an 8-shaped double spiral preventive cleaning plan for 1:00 AM on Wednesday.

[0065] This solution constructs a three-dimensional temperature field model through high-precision data fusion to accurately identify structural risks and dirt deposition hotspots; intelligent path planning generates cleaning trajectories that adapt to physical constraints, effectively fitting complex curved structures; a self-calibration mechanism corrects trajectory deviations in real time to ensure operational feasibility; deformation range and dirt prediction models drive a shift from passive maintenance to proactive defense; and finally, it integrates a pressure-adaptive preventive maintenance strategy to achieve synergistic optimization of cleaning efficiency, safety, and equipment lifespan.

[0066] This plan Figure 2 This application provides a schematic diagram of the structure of an ant colony algorithm optimization system for the drainage channel of a waterproof cabinet, as shown in the embodiment of this application. Figure 2 As shown, the system includes: The acquisition module 21 is used to acquire dual-source heterogeneous data inside the drainage channel of the waterproof cabinet through visual correction. The dual-source heterogeneous data includes multi-angle light images and synchronously changing humidity parameter groups. The generation module 22 is used to extract the viscosity coefficient in the humidity parameter group, fuse the viscosity coefficient with the multi-angle light image to generate an adhesion feature map, and construct a three-dimensional temperature field covering the curved surface of the drainage channel of the waterproof cabinet based on the adhesion feature map. Verification module 23 is used to apply an ant colony cooperative optimization mechanism in the three-dimensional temperature field to generate a set of clean trajectories under physical constraints, and to verify the fit between the set of clean trajectories and the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field to obtain the corresponding verification results. Correction module 24 is used to initiate a nonlinear self-calibration process to correct the motion vector deviation of the cleaning trajectory set when the verification result is lower than a preset threshold, so as to obtain the corrected motion vector deviation; The prediction module 25 is used to distinguish the deformation range of the channel structure based on the motion vector deviation, and to correlate the dirt deposition density in the three-dimensional temperature field with the deformation range to predict the dirt expansion trend. At the same time, it triggers a preventive maintenance operation chain with pressure-adaptive ant colony algorithm swarm intelligence optimization characteristics to achieve global optimization of the cleaning path of the waterproof cabinet drainage channel.

[0067] Figure 2 The aforementioned ant colony algorithm optimization system for the drainage channel of a waterproof cabinet can execute... Figure 1 The ant colony optimization method for the drainage channel of a waterproof cabinet, as described in the illustrated embodiment, will not be elaborated further on its implementation principle and technical effects. The specific methods by which each module and unit performs operations in the ant colony optimization system for the drainage channel of a waterproof cabinet as described in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0068] In one possible design, Figure 2 The ant colony algorithm optimization system for the drainage channel of a waterproof cabinet, as shown in the embodiment, can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0069] The processing component 32 is used for the above Figure 1 The embodiment describes an ant colony algorithm optimization method for the drainage channel of a waterproof cabinet.

[0070] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0071] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0072] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0073] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0074] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0075] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0076] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 An ant colony algorithm optimization method for the drainage channel of a waterproof cabinet is shown in the embodiment.

[0077] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0078] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0079] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An ant colony algorithm optimization method for waterproof cabinet drainage channels, characterized in that, Comprising; acquiring double-source heterogeneous data inside the drainage passage of the waterproof cabinet by visual correction acquisition, the double-source heterogeneous data including multi-angle light images and a group of synchronous changing humidity parameters; extracting a viscosity coefficient in the group of humidity parameters, fusing the viscosity coefficient with the multi-angle light images to generate an attached feature map, and constructing a three-dimensional temperature field covering a curved surface of the drainage passage of the waterproof cabinet according to the attached feature map; applying an ant colony cooperation optimization mechanism in the three-dimensional temperature field to generate a cleaning trajectory set under physical constraints, and verifying the cleaning trajectory set with a channel structure surface derived from spatial coordinates of the three-dimensional temperature field to obtain a corresponding verification result; when the verification result is lower than a preset threshold, starting a nonlinear self-calibration process to correct a motion vector deviation of the cleaning trajectory set to obtain a corrected motion vector deviation; distinguishing a deformation interval of the channel structure according to the motion vector deviation, and correlating a dirt deposition density in the three-dimensional temperature field with the deformation interval to predict a dirt expansion trend, while triggering a preventive maintenance operation chain of the group intelligence optimization feature of the pressure-adaptive fusion ant colony algorithm to realize global optimization of the cleaning path of the waterproof cabinet drainage passage.

2. The method of claim 1, wherein, applying an ant colony cooperation optimization mechanism in the three-dimensional temperature field to generate a cleaning trajectory set under physical constraints, comprising: defining a plurality of virtual agent points in the three-dimensional temperature field, each virtual agent point representing a cleaning unit, and initializing the starting position of the virtual agent point; setting a movement rule for each virtual agent point, calculating a movement probability based on the temperature value of the three-dimensional temperature field according to the movement rule; based on the physical constraint conditions including channel bending curvature limit and cleaning tool size limit, adding the physical constraint conditions to the movement rule to adjust the movement probability; through an iterative process, the virtual agent points move in the three-dimensional temperature field according to the movement probability, and the movement probability is updated after each movement; record the movement path of all virtual agent points, when the iteration times reach the preset value, output the set of movement paths as the cleaning trajectory set under the physical constraint conditions.

3. The method of claim 1, wherein, verifying the cleaning trajectory set with the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field to obtain a corresponding verification result, comprising: extracting spatial coordinates from the three-dimensional temperature field to construct a three-dimensional point cloud representation of the channel structure surface; obtaining the coordinates of each cleaning trajectory point from the cleaning trajectory set, and mapping the cleaning trajectory point to the three-dimensional point cloud; calculating the distance value of each cleaning trajectory point and the corresponding point of the channel structure surface, and statistically calculating the distance values of all cleaning trajectory points to calculate the average distance value and the maximum distance value; according to the average distance value and the maximum distance value, the fitting degree is verified by a comparison function and the corresponding verification result is output, the verification result represents the fitting degree of the cleaning trajectory set and the channel structure surface.

4. The method of claim 1, wherein, When the verification result is lower than a preset threshold, a nonlinear self-calibration process is started to correct the motion vector deviation of the set of cleaning trajectories, to obtain a corrected motion vector deviation, including: When the verification result is lower than a preset threshold, identify the motion vector of each cleaning trajectory point in the set of cleaning trajectories, which represents the direction and speed of cleaning movement; Based on the verification result and through a nonlinear function, a corresponding correction coefficient is calculated, and the motion vector is adjusted according to the correction coefficient, and the correction coefficient is applied to the motion vector to update the direction and speed components of the motion vector; Through iterative optimization, the updated motion vector satisfies the geometric constraint of the channel structure surface, and the updated motion vector of all cleaning trajectory points is output as the corrected motion vector deviation.

5. The method of claim 1, wherein, According to the motion vector deviation, the deformation interval of the channel structure is distinguished, and the dirt deposition density in the three-dimensional temperature field is associated with the deformation interval to predict the dirt expansion trend, including: Extract the deviation value of each cleaning trajectory point from the motion vector deviation, and divide the channel structure into multiple continuous intervals according to the size of the deviation value, and each interval is defined as a deformation interval; Based on the temperature value in the three-dimensional temperature field, the dirt deposition density value of each point is obtained by conversion; Map the dirt deposition density value to the deformation interval, and calculate the average dirt deposition density of each deformation interval; Based on the average dirt deposition density, the dirt growth rate of each deformation interval is calculated by considering the trend prediction model of the time factor and the interval geometric characteristics, and the dirt growth rate is output as the dirt expansion trend.

6. The method of claim 1, wherein, Extract the viscous coefficient in the humidity parameter group, and fuse the viscous coefficient with the multi-angle light image to generate an adhesion feature map, including: Obtain multiple humidity measurement values from the humidity parameter group, and derive the viscous coefficient according to the humidity measurement values through a calculation rule predefined based on the physical relationship between humidity and material adhesion characteristics; For the multi-angle light image, identify the brightness change information of each pixel point in the multi-angle light image, and convert the brightness change information into a surface texture value; Combine the viscous coefficient with the surface texture value, and through a position matching operation, distribute the viscous coefficient to the corresponding pixel point position of the multi-angle light image to generate a fusion data layer; On the fusion data layer, calculate the weighted sum of the viscous coefficient and the surface texture value for each pixel point position to obtain the adhesion intensity value of each position; Based on the adhesion intensity value of all pixel point positions, a two-dimensional grid structure is constructed, which covers the observation surface of the waterproof cabinet drainage channel, and is output as an adhesion feature map.

7. The method of claim 1, wherein, According to the adhesion feature map, a three-dimensional temperature field covering the curved surface of the waterproof cabinet drainage channel is constructed, including: Obtain the adhesion intensity value of each position from the adhesion feature map, and input the adhesion intensity value into a heat conduction model simulating the heat distribution of the curved surface of the waterproof cabinet drainage channel; In the heat conduction model, a heat source influence factor representing the degree of influence of dirt deposition on local temperature is calculated for each position according to the adhesion strength value; Based on the heat source influence factor, a spatial mapping operation is performed according to the curved geometry of the waterproof cabinet drain channel to convert the two-dimensional grid structure of the adhesion feature map into three-dimensional spatial coordinates; On the three-dimensional spatial coordinates, a temperature value calculated by the heat source influence factor of adjacent points is assigned to each point; Combined with the assigned temperature values on the three-dimensional spatial coordinates, a three-dimensional grid containing the spatial coordinates and corresponding temperature values of all points is constructed to form a three-dimensional temperature field.

8. An ant colony optimization system for optimizing a drain channel of a waterproof cabinet, characterized by, It includes: Acquire double-source heterogeneous data inside the waterproof cabinet drain channel through visual correction acquisition, which includes multi-angle light images and synchronously changing humidity parameter groups; Extract the viscosity coefficient in the humidity parameter group, and fuse the viscosity coefficient with the multi-angle light images to generate an adhesion feature map, and construct a three-dimensional temperature field covering the curved surface of the waterproof cabinet drain channel according to the adhesion feature map; In the three-dimensional temperature field, an ant colony cooperation optimization mechanism is applied to generate a cleaning trajectory set under physical constraints, and the cleaning trajectory set is verified with the channel structure surface derived from the spatial coordinates of the three-dimensional temperature field to obtain a corresponding verification result; When the verification result is lower than a preset threshold, start a nonlinear self-calibration process to correct the motion vector deviation of the cleaning trajectory set to obtain a corrected motion vector deviation; According to the motion vector deviation, the deformation interval of the channel structure is distinguished, and the dirt deposition density in the three-dimensional temperature field is associated with the deformation interval to predict the dirt expansion trend, and the preventive maintenance operation chain of the group intelligence optimization feature of the pressure adaptive fusion ant colony algorithm is triggered, realizing the global optimization of the cleaning path of the waterproof cabinet drain channel.

9. A computing device, comprising: It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to realize the ant colony algorithm optimization method of the waterproof cabinet drain channel according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that, A computer program is stored, and when the computer program is executed by a computer, the ant colony algorithm optimization method of the waterproof cabinet drain channel according to any one of claims 1 to 7 is realized.

Images