Unmanned aerial vehicle lens microfluidic cleaning method and system based on stain identification

Abstract

The invention relates to the technical field of unmanned aerial vehicles. The unmanned aerial vehicle lens microfluidic cleaning method and system based on stain recognition are provided, and the method comprises the steps that pollution area coordinate information and pollution level data of the surface of a lens are acquired; generating a cleaning solution distribution scheme based on the pollution area coordinate information and the pollution level data; micro-fluidic liquid conveying is executed according to the cleaning liquid distribution scheme, and real-time liquid distribution state data are generated; optimizing the liquid coverage parameters, and generating an optimized liquid distribution scheme; evaluating the preliminary cleaning effect and iteratively adjusting the parameters, and generating an adjusted cleaning execution scheme; and cleaning is completed based on the adjusted cleaning execution scheme, and cleaning state data is generated, so that the adaptive deviation between pollution detection and a cleaning liquid distribution scheme is improved, the real-time feedback capability of liquid coverage control is enhanced, and the verification reliability of cleaning residue evaluation is optimized.

Description

Technical Field

[0001] The present invention relates to the field of drone technology, and in particular to a microfluidic cleaning method and system for drone lenses based on stain recognition. Background Art

[0002] With the widespread use of drones in environmental monitoring, aerial photography, disaster relief, and other fields, lens cleaning has become a critical step in ensuring image quality and operational reliability. Due to the influence of the flight environment (such as wind, sand, rain, and dust), contaminants easily adhere to the surface of drone lenses, resulting in blurred images or distorted data.

[0003] However, the relevant drone lens cleaning technology has the following problems: the dynamic detection and cleaning liquid distribution scheme are mismatched, resulting in resource waste and cleaning blind spots; the microfluidic liquid coverage control lacks a real-time feedback mechanism, resulting in insufficient coverage and liquid accumulation; the cleaning residue assessment lacks a closed-loop verification mechanism, resulting in insufficient cleanliness reliability. Summary of the Invention

[0004] Based on this, it is necessary to provide a microfluidic cleaning method and system for drone lenses based on stain recognition to address the above technical problems, so as to improve the adaptation deviation between contamination detection and cleaning liquid distribution scheme, enhance the real-time feedback capability of liquid coverage control, and optimize the verification reliability of cleaning residue evaluation, thereby improving cleaning efficiency and reducing resource waste.

[0005] In a first aspect, the present application provides a microfluidic cleaning method for drone lenses based on stain recognition, the method comprising:

[0006] Obtain the coordinate information of the contaminated area and the contamination level data on the lens surface through contamination detection and quantitative processing;

[0007] Generate a cleaning fluid distribution plan based on the contaminated area coordinate information and contamination level data;

[0008] Execute microfluidic liquid delivery according to the cleaning liquid distribution plan and generate real-time liquid distribution status data;

[0009] Optimize liquid coverage parameters based on real-time liquid distribution status data and generate an optimized liquid distribution plan;

[0010] Evaluate the initial cleaning effect based on the optimized liquid distribution plan and iteratively adjust the parameters to generate an adjusted cleaning execution plan;

[0011] The cleaning is completed based on the adjusted cleaning execution plan, and cleaning status data is generated.

[0012] Furthermore, the preliminary cleaning effect is evaluated based on the optimized liquid distribution plan and parameters are iteratively adjusted to generate an adjusted cleaning execution plan, including:

[0013] Based on the optimized liquid distribution scheme, liquid coverage area data is obtained through liquid coverage area collection and processing;

[0014] The following formula is used to compare and analyze the liquid coverage area data and the initial contaminated area coordinate information, and a residual contamination distribution map is generated using the contamination residual analysis model:

[0015] D(x,y)=P(x,y)·(1-η·T(x,y))

[0016]

[0017] Among them, D(x,y) represents the residual pollution distribution function, P(x,y) represents the initial pollution distribution function, T(x,y) represents the residence time function of the liquid at point (x,y), η represents the cleaning efficiency coefficient per unit time, M(x,y) represents the pollutant distribution map after multi-sensor fusion, S i (x,y) represents the pollution distribution detected by the i-th sensor, ω i represents the weight coefficient of the i-th sensor, and m represents the total number of sensors;

[0018] Based on the residual density data of pollutants in the residual pollution distribution map, it is judged whether the preliminary cleaning effect meets the preset cleaning standard and a judgment result is obtained, which includes whether the cleaning standard is met or not;

[0019] If the cleaning standard is not met, dynamic weight distribution processing is performed based on the residual density data of pollutants and the activity parameters of the cleaning fluid to generate adjustment parameters for the spraying amount and spraying time of the cleaning fluid;

[0020] The cleaning execution parameters are recalculated through an incremental adjustment algorithm according to the adjustment parameters to generate an adjusted cleaning execution plan.

[0021] Furthermore, the liquid coverage area data and the initial contaminated area coordinate information are compared and analyzed, and a residual contamination distribution map is generated through the contamination residual analysis model, including:

[0022] Based on the liquid coverage area data and the initial contaminated area coordinate information, the contamination clearance difference value mapping data is generated through overlapping area analysis and processing;

[0023] The following formula is used to generate residual contamination density level distribution data based on the contamination clearance difference value mapping data through density gradient calculation:

[0024] L(x,y)=Classify(|ρ(x,y)|,{T1,T2,...,T k})

[0025] Among them, L(x,y) represents the residual pollution density level distribution data, |ρ(x,y)| represents the density gradient modulus, T1,T2,...,T k Indicates the predefined density level threshold;

[0026] The residual pollution distribution map is generated by fusing the residual pollution density level distribution data with the chemical activity decay parameters in the pollution residue analysis model.

[0027] Furthermore, when the judgment result is that the cleaning standard is not met, a dynamic weight distribution process is performed based on the pollutant residual density data and the cleaning liquid activity parameter to generate adjustment parameters for the cleaning liquid spraying amount and spraying time, including:

[0028] Based on the regional residual level distribution in the pollutant residual density data, the dynamic attenuation coefficient of the chemical activity of the cleaning fluid is calculated using the activity attenuation compensation algorithm;

[0029] The following formula is used to perform a weighted fusion process based on the dynamic attenuation coefficient and the basic activity value in the cleaning fluid activity parameter to generate the cleaning fluid activity compensation weight:

[0030] W c =α·D d +(1-α)·A b

[0031] Among them, W c represents the cleaning fluid activity compensation weight, D d Represents the dynamic attenuation coefficient, A b represents the basic activity value in the cleaning fluid activity parameter, α represents the weighted fusion coefficient, and its value range is [0,1];

[0032] Multi-factor normalization processing is performed based on the cleaning fluid activity compensation weight and pollutant residual density data to generate adjustment parameters for the cleaning fluid spraying amount and spraying time.

[0033] Furthermore, a cleaning liquid distribution plan is generated based on the contaminated area coordinate information and the contamination level data, including:

[0034] Based on the coordinate information of the polluted area, a pollution density distribution map is generated through pollution density distribution modeling;

[0035] Based on the pollution level data and pollution density distribution map, the cleaning priority sub-areas are divided through the spatial weight allocation algorithm to generate the sub-area distribution map;

[0036] Based on the pollution density gradient data in the sub-area distribution map, the dynamic distribution parameters are generated by matching the cleaning liquid spray pressure and flow parameters through the fluid dynamics model;

[0037] The dynamic allocation parameters and pollution level data are normalized and fused to generate a cleaning fluid allocation plan.

[0038] Furthermore, based on the pollution density gradient data in the sub-area distribution map, the cleaning liquid spray pressure and flow parameters are matched through the fluid dynamics model to generate dynamic distribution parameters, including:

[0039] Based on the pollution density gradient data, the fluid dynamic partitioning process is used to divide the area where the pollution density is higher than the preset threshold and the area where the pollution density is lower than the preset threshold, thereby generating fluid partition mapping data;

[0040] According to the fluid partition mapping data, the local spraying pressure parameters of the area where the pollution density is higher than the preset threshold are matched by the adaptive pressure compensation algorithm in the fluid dynamics model;

[0041] Based on the local spraying pressure parameters and the fluid characteristic data of the microfluidic channel, flow parameter matching processing is performed to generate initial flow distribution parameters;

[0042] The initial flow distribution parameters are decoupled and fused with the pressure fluctuation suppression parameters in the adaptive pressure compensation algorithm to generate dynamic distribution parameters.

[0043] Furthermore, microfluidic liquid delivery is performed according to the cleaning liquid distribution plan to generate real-time liquid distribution status data, including:

[0044] According to the dynamic distribution parameters in the cleaning liquid distribution scheme, the spray angle parameters adapted to the contaminated area are generated through microfluidic nozzle angle adjustment processing;

[0045] Based on the spray angle parameters and the fluid characteristic data of the microfluidic channel, the spray pressure parameters of the cleaning fluid are matched through a dynamic pressure compensation algorithm;

[0046] According to the spraying pressure parameters and the pollution density gradient distribution data, incremental liquid delivery is performed on the areas where the pollution density is higher than the preset threshold through partitioned injection control processing to generate real-time liquid distribution status data.

[0047] Furthermore, according to the dynamic distribution parameters in the cleaning liquid distribution scheme, the spray angle parameters adapted to the contaminated area are generated through microfluidic nozzle angle adjustment processing, including:

[0048] Based on the pollution density gradient distribution data in the dynamic allocation parameters, the geometric feature data of the pollution area is extracted through the pollution area geometric analysis processing;

[0049] According to the geometric feature data of the contaminated area and the fluid diffusion characteristic data of the microfluidic channel, the nozzle's spray angle range parameters are matched through a dynamic angle optimization algorithm;

[0050] Based on the injection angle range parameters and the spatial distribution data of the area where the pollution density is higher than the preset threshold, the injection angle parameters adapted to the pollution area are generated through adaptive adjustment processing.

[0051] Furthermore, the cleaning is completed based on the adjusted cleaning execution plan, and cleaning status data is generated, including:

[0052] Based on the incremental cleaning parameters in the adjusted cleaning execution plan, collecting intermediate image data of the lens surface after cleaning through multispectral analysis and processing;

[0053] Perform residual pollution quantification processing on the intermediate image data and the coordinate information of the initial contaminated area to generate residual pollution coverage area data;

[0054] Perform dynamic parameter matching processing based on the residual contamination coverage area data and the preset cleanliness threshold to generate cleanliness assessment results;

[0055] Based on the cleanliness assessment results and the dynamic adjustment factors in the incremental cleaning parameters, cleaning status data is generated through a feedback optimization algorithm.

[0056] In a second aspect, the present application also provides a microfluidic cleaning system for drone lenses based on stain recognition, the system comprising:

[0057] The stain detection and pre-quantification module is used to obtain the coordinate information of the contaminated area and the contamination level data on the lens surface through contamination detection and quantification processing;

[0058] A dynamic cleaning fluid allocation module, used to generate a cleaning fluid allocation plan based on the contaminated area coordinate information and contamination level data;

[0059] A microfluidic injection module is used to perform microfluidic liquid delivery according to the cleaning liquid distribution plan and generate real-time liquid distribution status data;

[0060] Liquid coverage dynamic optimization module, used to optimize liquid coverage parameters based on real-time liquid distribution status data and generate an optimized liquid distribution plan;

[0061] The cleaning effect evaluation and iteration module is used to evaluate the initial cleaning effect and iteratively adjust the parameters based on the optimized liquid distribution plan to generate an adjusted cleaning execution plan;

[0062] The cleaning closed-loop execution module is used to complete cleaning based on the adjusted cleaning execution plan and generate cleaning status data.

[0063] The technical solution provided by this application includes the following technical effects: by providing a microfluidic cleaning method and system for drone lenses based on stain recognition, the method includes: obtaining the coordinate information of the contaminated area and the pollution level data on the lens surface through pollution detection and quantitative processing; generating a cleaning liquid distribution plan based on the contaminated area coordinate information and the pollution level data; executing microfluidic liquid delivery according to the cleaning liquid distribution plan to generate real-time liquid distribution status data; optimizing the liquid coverage parameters based on the real-time liquid distribution status data to generate an optimized liquid distribution plan; evaluating the preliminary cleaning effect according to the optimized liquid distribution plan and iteratively adjusting the parameters to generate an adjusted cleaning execution plan; completing cleaning based on the adjusted cleaning execution plan and generating cleaning status data to improve the adaptation deviation between the contamination detection and the cleaning liquid distribution plan, enhance the real-time feedback capability of the liquid coverage control, and optimize the verification reliability of the cleaning residue evaluation, thereby improving the cleaning efficiency and reducing resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 This is a flow chart of a microfluidic cleaning method for drone lenses based on stain recognition in one embodiment of the present invention;

[0066] Figure 2 A flowchart of generating a cleaning liquid distribution plan based on contaminated area coordinate information and contamination level data in one embodiment of the present invention;

[0067] Figure 3 This is a structural diagram of a microfluidic cleaning system for drone lenses based on stain recognition in one embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the above-mentioned purposes, features and advantages of the present application more clearly understood, the specific implementation methods of the present application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to fully understand the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the connotation of the application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0069] like Figure 1 As shown, the present application provides a microfluidic cleaning method for drone lenses based on stain recognition, the method comprising:

[0070] S101: Obtaining the contamination area coordinate information and contamination level data on the lens surface through contamination detection and quantification processing.

[0071] Specifically, high-resolution imaging technology is combined with multi-spectral analysis technology to scan and image the surface of the drone lens, capture changes in the optical properties of the lens surface, and identify existing contaminated areas. At the same time, a pollution feature database is used for comparative analysis to determine the type of pollution (such as dust, oil, water stains, etc.). Through image processing technology, edge detection and feature extraction are performed on the acquired lens surface image to locate the coordinate position of the contaminated area on the lens surface, providing a spatial positioning basis for the subsequent spraying of cleaning fluid. Using light intensity analysis technology, the light reflection and refraction characteristics of the contaminated area are quantitatively analyzed. Combined with the type of pollution, the severity of the pollution is assessed and pollution level data is generated to provide a reference basis for the amount of cleaning fluid used and the spraying intensity.

[0072] S102: Generate a cleaning liquid distribution plan based on the contaminated area coordinate information and the contamination level data.

[0073] Specifically, the acquired coordinate information of the contaminated area is analyzed and processed to clarify the distribution of the contaminated area on the lens surface. Combined with the pollution level data, a quantitative assessment of the severity of the contamination is performed. Based on the contaminated area coordinate information and pollution level data, the spray pressure, flow rate, and spray time parameters of the cleaning fluid are preliminarily determined. A mapping relationship is established between the contaminated area coordinate information and the cleaning fluid spray parameters to achieve more accurate distribution of the cleaning fluid. Based on the spatial distribution characteristics of the contaminated area, the contamination density gradient is analyzed to optimize the distribution plan of the cleaning fluid. Based on the above analysis results, a cleaning fluid distribution plan is generated to ensure that the cleaning fluid can act on the contaminated area more accurately and efficiently.

[0074] S103: Execute microfluidic liquid delivery according to the cleaning liquid distribution plan to generate real-time liquid distribution status data.

[0075] Specifically, based on the cleaning liquid distribution plan, the initial position and spray angle of the microfluidic nozzle are adjusted to align it with the contaminated area on the lens surface. The microfluidic liquid delivery system is activated to deliver the cleaning liquid according to the predetermined flow and pressure parameters. During the liquid delivery process, optical sensors and image processing technology are used to monitor the distribution of the cleaning liquid on the lens surface in real time. Feedback data from the sensors is collected, including information such as liquid coverage area, distribution uniformity, and liquid thickness. This collected data is processed and analyzed to generate real-time liquid distribution status data, which is used to evaluate the liquid delivery effect and provide a basis for subsequent adjustments.

[0076] S104: Optimizing liquid coverage parameters based on real-time liquid distribution status data to generate an optimized liquid distribution plan.

[0077] Specifically, the system collects real-time liquid distribution data from sensors, including information such as liquid coverage area, distribution uniformity, and liquid thickness. Based on this analysis, the system evaluates the effectiveness of current liquid coverage parameters and identifies areas of insufficient or excessive liquid coverage. Based on this evaluation, an optimization strategy is developed to adjust liquid coverage parameters, such as the nozzle's spray angle, flow rate, and pressure. Based on this optimization strategy, an optimized liquid distribution plan is generated, ensuring that the cleaning liquid covers the contaminated area more evenly and effectively.

[0078] S105: Evaluate the preliminary cleaning effect according to the optimized liquid distribution plan and iteratively adjust the parameters to generate an adjusted cleaning execution plan.

[0079] Specifically, based on the optimized liquid distribution plan, the cleaning effect evaluation program is started to collect images of the lens surface and liquid distribution data after preliminary cleaning. Using high-resolution imaging technology and multispectral analysis technology, the remaining contaminated areas on the lens surface are identified, and the distribution and density information of the residual contamination is obtained. By comparing and analyzing the initial contamination data with the current residual contamination data, combined with the preset cleanliness standards, it is determined whether the initial cleaning effect meets the standards. For areas that do not meet the standards, a dynamic adjustment algorithm is used to optimize parameters such as the cleaning liquid spray volume, spray time, and the injection angle of the microfluidic nozzle based on the residual contamination distribution and density information. The optimized parameters are integrated to generate a new cleaning execution plan to provide guidance for subsequent cleaning and ensure that the cleanliness of the lens surface meets the preset requirements.

[0080] S106: Complete cleaning based on the adjusted cleaning execution plan and generate cleaning status data.

[0081] Specifically, according to the adjusted cleaning execution plan, the microfluidic system's parameters, such as the nozzle spray angle, cleaning fluid flow rate, and spray pressure, are configured to ensure the equipment is in optimal operating condition for cleaning tasks. The microfluidic injection system is activated, and the cleaning fluid is sprayed onto the contaminated areas on the lens surface according to the set parameters for efficient cleaning. During the cleaning process, optical sensors and image processing technology are used to monitor the cleanliness of the lens surface in real time, collecting data such as liquid distribution and residual contamination. By analyzing the collected data, the cleaning effect is evaluated to determine whether the contamination has been effectively removed. The uniformity of the liquid distribution is also checked to ensure that the cleaning process meets preset standards. Based on the cleaning effect evaluation results, cleaning status data is generated to record the condition of the lens surface after cleaning, providing a basis for subsequent cleaning effect verification and equipment operation records.

[0082] An embodiment of the present application provides a microfluidic cleaning method for drone lenses based on stain recognition, including: obtaining contamination area coordinate information and contamination level data on the lens surface through contamination detection and quantification processing; generating a cleaning liquid distribution plan based on the contamination area coordinate information and the contamination level data; executing microfluidic liquid delivery according to the cleaning liquid distribution plan to generate real-time liquid distribution status data; optimizing liquid coverage parameters based on the real-time liquid distribution status data to generate an optimized liquid distribution plan; evaluating the preliminary cleaning effect according to the optimized liquid distribution plan and iteratively adjusting the parameters to generate an adjusted cleaning execution plan; completing cleaning based on the adjusted cleaning execution plan and generating cleaning status data to improve the adaptation deviation between contamination detection and the cleaning liquid distribution plan, enhance the real-time feedback capability of liquid coverage control, and optimize the verification reliability of cleaning residue evaluation, thereby improving cleaning efficiency and reducing resource waste.

[0083] Furthermore, the preliminary cleaning effect is evaluated based on the optimized liquid distribution plan and parameters are iteratively adjusted to generate an adjusted cleaning execution plan, including:

[0084] Based on the optimized liquid distribution scheme, liquid coverage area data is obtained through liquid coverage area collection and processing;

[0085] The following formula is used to compare and analyze the liquid coverage area data and the initial contaminated area coordinate information, and a residual contamination distribution map is generated using the contamination residual analysis model:

[0086] D(x,y)=P(x,y)·(1-η·T(x,y))

[0087]

[0088] Among them, D(x,y) represents the residual pollution distribution function, P(x,y) represents the initial pollution distribution function, T(x,y) represents the residence time function of the liquid at point (x,y), η represents the cleaning efficiency coefficient per unit time, M(x,y) represents the pollutant distribution map after multi-sensor fusion, S i (x,y) represents the pollution distribution detected by the i-th sensor, ω i represents the weight coefficient of the i-th sensor, and m represents the total number of sensors;

[0089] Based on the residual density data of pollutants in the residual pollution distribution map, it is judged whether the preliminary cleaning effect meets the preset cleaning standard and a judgment result is obtained, which includes whether the cleaning standard is met or not;

[0090] If the cleaning standard is not met, dynamic weight distribution processing is performed based on the residual density data of pollutants and the activity parameters of the cleaning fluid to generate adjustment parameters for the spraying amount and spraying time of the cleaning fluid;

[0091] The cleaning execution parameters are recalculated through an incremental adjustment algorithm according to the adjustment parameters to generate an adjusted cleaning execution plan.

[0092] Specifically, based on the optimized liquid distribution scheme, high-precision sensors are used to monitor and collect data on the liquid coverage area in real time, obtaining detailed data such as the actual coverage range, distribution density, and coverage uniformity of the liquid on the lens surface. The residence time of the liquid in different areas is also recorded, providing time-dimensional information for subsequent analysis. The collected liquid coverage area data is compared and analyzed with the coordinate information of the initial contaminated area, and a residual contamination distribution map is generated using a contamination residual analysis model. This model comprehensively considers factors such as liquid coverage, residence time, and cleaning efficiency coefficient, and combines the contaminant distribution information obtained by multi-sensor fusion technology to locate and quantify the residual contamination area, providing a scientific basis for judging the cleaning effect.

[0093] The residual contaminant density data from the residual contamination distribution map is compared with the preset cleaning standard to determine whether the initial cleaning effect meets the standard, resulting in a clear judgment result: whether the cleaning standard has been met or not. This process ensures an accurate assessment of the cleaning effect and provides a key basis for subsequent decision-making. If the cleaning standard is not met, a dynamic weight allocation algorithm is used to adjust the spray volume and spray time of the cleaning solution based on the residual contaminant density data and the activity parameters of the cleaning solution, generating corresponding adjustment parameters. This process fully considers the residual contamination situation in different areas and the chemical properties of the cleaning solution, achieving optimal configuration of cleaning parameters.

[0094] Based on the adjusted parameters, an incremental adjustment algorithm is used to recalculate the cleaning execution parameters and generate a new cleaning execution plan. This plan is optimized and improved on the original basis, ensuring the accuracy and efficiency of the cleaning process and providing a strong guarantee for achieving a more thorough cleaning of the lens surface.

[0095] Furthermore, the liquid coverage area data and the initial contaminated area coordinate information are compared and analyzed, and a residual contamination distribution map is generated through the contamination residual analysis model, including:

[0096] Based on the liquid coverage area data and the initial contaminated area coordinate information, the contamination clearance difference value mapping data is generated through overlapping area analysis and processing;

[0097] The following formula is used to generate residual contamination density level distribution data based on the contamination clearance difference value mapping data through density gradient calculation:

[0098] L(x,y)=Classify(|ρ(x,y)|,{T1,T2,...,T k})

[0099] Among them, L(x,y) represents the residual pollution density level distribution data, |ρ(x,y)| represents the density gradient modulus, T1,T2,...,T k Indicates the predefined density level threshold;

[0100] The residual pollution distribution map is generated by fusing the residual pollution density level distribution data with the chemical activity decay parameters in the pollution residue analysis model.

[0101] Specifically, the liquid-covered area data and the coordinate information of the initial contaminated area are analyzed for overlap. By comparing the coordinates of the liquid-covered area with the initial contaminated area, the areas that have been covered by liquid and those that still remain contaminated are determined. This analysis identifies the differences between the liquid-covered and contaminated areas, generating contamination clearance difference mapping data to clearly define the effectiveness of contamination clearance and the location of remaining contamination.

[0102] Next, the pollution clearance difference mapping data is used to perform density gradient calculations. By calculating the spatial gradient of the pollution clearance difference values, the residual pollution density distribution data is obtained. This step quantifies the concentration of residual pollution and reveals the distribution patterns of pollution in different areas, providing more detailed information for subsequent pollution assessments.

[0103] The residual contamination density distribution data and the chemical activity decay parameters from the residual contamination analysis model are then fused. The chemical activity decay parameters reflect how the cleaning effectiveness of the cleaning fluid in different areas changes over time. This fusion process comprehensively considers the effects of contamination density and cleaning fluid activity decay to generate a residual contamination distribution map. This map intuitively displays the spatial distribution of residual contamination on the lens surface, providing critical data support for evaluating cleaning effectiveness and formulating further cleaning strategies.

[0104] Furthermore, when the judgment result is that the cleaning standard is not met, a dynamic weight distribution process is performed based on the pollutant residual density data and the cleaning liquid activity parameter to generate adjustment parameters for the cleaning liquid spraying amount and spraying time, including:

[0105] Based on the regional residual level distribution in the pollutant residual density data, the dynamic attenuation coefficient of the chemical activity of the cleaning fluid is calculated using the activity attenuation compensation algorithm;

[0106] The following formula is used to perform a weighted fusion process based on the dynamic attenuation coefficient and the basic activity value in the cleaning fluid activity parameter to generate the cleaning fluid activity compensation weight:

[0107] W c =α·D d +(1-α)·Ab

[0108] Among them, W c represents the cleaning fluid activity compensation weight, D d Represents the dynamic attenuation coefficient, A b represents the basic activity value in the cleaning fluid activity parameter, α represents the weighted fusion coefficient, and its value range is [0,1];

[0109] Multi-factor normalization processing is performed based on the cleaning fluid activity compensation weight and pollutant residual density data to generate adjustment parameters for the cleaning fluid spraying amount and spraying time.

[0110] Specifically, based on the regional residual level distribution in the pollutant residual density data, an activity decay compensation algorithm is used to calculate the dynamic decay coefficient of the chemical activity of the cleaning fluid. This step takes into account the severity of the pollution residue in different areas and the changes in the activity of the cleaning fluid over time or usage, thereby obtaining an attenuation index that reflects the actual cleaning ability of the cleaning fluid in each area. Through weighted fusion processing, the dynamic attenuation coefficient is combined with the basic activity value in the cleaning fluid activity parameter to generate the cleaning fluid activity compensation weight. In this step, taking into account the basic activity of the cleaning fluid and its decay in different polluted areas, a specific weighted fusion mechanism is used to obtain a compensation weight that can reflect the actual activity level of the cleaning fluid in each area, providing a basis for subsequent adjustments to the spraying amount and time.

[0111] Based on the cleaning fluid activity compensation weight and pollutant residual density data, a multi-factor normalization process is performed to generate adjustment parameters for the cleaning fluid spray volume and spray time. This step integrates multiple factors, such as the cleaning fluid activity compensation weight and pollutant residual density, into a unified framework through normalization. The resulting adjusted cleaning fluid spray volume and spray time parameters ensure that the cleaning fluid is effectively distributed and used based on the residual contaminant situation and the cleaning fluid activity.

[0112] like Figure 2 As shown, a cleaning liquid distribution plan is generated based on the contaminated area coordinate information and contamination level data, including:

[0113] S201: Generate a pollution density distribution map through pollution density distribution modeling based on the pollution area coordinate information;

[0114] S202: Based on the pollution level data and the pollution density distribution map, the cleaning priority sub-areas are divided by a spatial weight allocation algorithm to generate a sub-area distribution map;

[0115] S203: Based on the pollution density gradient data in the sub-region distribution map, the cleaning liquid spraying pressure and flow parameters are matched through a fluid dynamics model to generate dynamic distribution parameters;

[0116] S204: Perform normalization and fusion processing on the dynamic allocation parameters and the pollution level data to generate a cleaning liquid allocation plan.

[0117] Specifically, the acquired coordinates of the contaminated areas are analyzed in detail to determine the exact distribution and extent of contamination on the lens surface. Using specialized modeling techniques, a contamination density distribution model is constructed based on the coordinates of the contaminated areas. This model reflects the concentration of contamination in different areas. Through model calculation and visualization, a contamination density distribution map is generated. This map intuitively displays the contamination density of each area on the lens surface, providing basic data support for subsequent cleaning fluid distribution.

[0118] Combined with pollution level data, the severity of pollution is quantified and integrated to obtain pollution level information for each contaminated area. A spatial weight allocation algorithm, combining the pollution density distribution map and pollution level data, determines the priority of different areas during the cleaning process. Based on the algorithm's calculation results, the lens surface is divided into multiple cleaning priority sub-areas and a sub-area distribution map is generated, providing more detailed spatial guidance for the precise distribution of cleaning fluid.

[0119] Contamination density gradient data is extracted from the sub-area distribution map to analyze the changing trends of contamination density across different sub-areas. Using a fluid dynamics model, the corresponding cleaning fluid spray pressure and flow rate parameters are matched to the contamination density gradient data to ensure that the cleaning fluid effectively covers the contaminated area and achieves the desired cleaning effect. Combined with the model calculation results, dynamic distribution parameters are generated. These parameters include information such as spray pressure, flow rate, and spray time for the cleaning fluid in different sub-areas, providing more precise parameter support for dynamic distribution of the cleaning fluid.

[0120] The generated dynamic distribution parameters are combined with the contamination level data for a comprehensive analysis. Through normalized fusion processing, the cleaning fluid distribution is more accurately aligned with the contamination situation. Based on the fusion results, the cleaning fluid distribution plan is optimized to ensure that the distribution of cleaning fluid in each sub-area meets the required contamination density and level while achieving optimal cleaning results. A complete cleaning fluid distribution plan is then generated, detailing parameters such as spray pressure, flow rate, time, and sequence for each contaminated area on the lens surface, providing accurate guidance and a basis for subsequent cleaning operations.

[0121] Furthermore, based on the pollution density gradient data in the sub-area distribution map, the cleaning liquid spray pressure and flow parameters are matched through the fluid dynamics model to generate dynamic distribution parameters, including:

[0122] Based on the pollution density gradient data, the fluid dynamic partitioning process is used to divide the area where the pollution density is higher than the preset threshold and the area where the pollution density is lower than the preset threshold, thereby generating fluid partition mapping data;

[0123] According to the fluid partition mapping data, the local spraying pressure parameters of the area where the pollution density is higher than the preset threshold are matched by the adaptive pressure compensation algorithm in the fluid dynamics model;

[0124] Based on the local spraying pressure parameters and the fluid characteristic data of the microfluidic channel, flow parameter matching processing is performed to generate initial flow distribution parameters;

[0125] The initial flow distribution parameters are decoupled and fused with the pressure fluctuation suppression parameters in the adaptive pressure compensation algorithm to generate dynamic distribution parameters.

[0126] Specifically, based on the contamination density gradient data in the sub-region distribution map, dynamic fluid zoning is used to identify and demarcate high-contamination areas (where contamination density exceeds a preset threshold) and low-contamination areas (where contamination density falls below a preset threshold). This generates fluid zoning mapping data, providing a zoning basis for subsequent cleaning fluid spray parameter matching. Based on this fluid zoning mapping data, an adaptive pressure compensation algorithm within the fluid dynamics model is employed to match local spray pressure parameters to highly contaminated areas. This algorithm dynamically adjusts the spray pressure based on the contamination density and fluid dynamics characteristics of the highly contaminated area, ensuring that the cleaning fluid effectively covers and cleans the highly contaminated area.

[0127] Flow parameter matching is performed by combining local spray pressure parameters with the fluid characteristics of the microfluidic channel. By analyzing the fluid dynamics of the microfluidic channel, initial flow distribution parameters are determined to match the local spray pressure and meet the cleaning requirements of different contaminated areas. Decoupling and fusion are performed on the initial flow distribution parameters and the pressure fluctuation suppression parameters in the adaptive pressure compensation algorithm. This decoupling and fusion process comprehensively considers the dynamic changes in flow and pressure to generate dynamic distribution parameters, ensuring more stable and uniform distribution of the cleaning fluid during the spraying process, thereby improving cleaning effectiveness.

[0128] Furthermore, microfluidic liquid delivery is performed according to the cleaning liquid distribution plan to generate real-time liquid distribution status data, including:

[0129] According to the dynamic distribution parameters in the cleaning liquid distribution scheme, the spray angle parameters adapted to the contaminated area are generated through microfluidic nozzle angle adjustment processing;

[0130] Based on the spray angle parameters and the fluid characteristic data of the microfluidic channel, the spray pressure parameters of the cleaning fluid are matched through a dynamic pressure compensation algorithm;

[0131] According to the spraying pressure parameters and the pollution density gradient distribution data, incremental liquid delivery is performed on the areas where the pollution density is higher than the preset threshold through partitioned injection control processing to generate real-time liquid distribution status data.

[0132] Specifically, dynamic distribution parameters are obtained from the cleaning fluid distribution plan. These parameters include the spraying requirements for different contaminated areas. Based on the dynamic distribution parameters, the angle of the microfluidic nozzle is adjusted through an angle adjustment mechanism to more accurately target each contaminated area, thereby generating spray angle parameters adapted to the contaminated area. This step takes into account factors such as the location, shape, and size of the contaminated area to ensure that the cleaning fluid is sprayed more accurately to the target area.

[0133] Combined with the generated spray angle parameters, their impact on the spray pressure of the cleaning liquid is analyzed, because changes in the spray angle will affect the spray range and force of the liquid. The fluid characteristic data of the microfluidic channel is obtained, including the viscosity and surface tension of the fluid, which will affect the flow and spraying effect of the cleaning liquid in the channel. Based on the above analysis results, a dynamic pressure compensation algorithm is used to match the cleaning liquid spray pressure parameters suitable for the current spray angle and fluid characteristics. This algorithm can dynamically adjust the spray pressure according to changes in the spray angle and fluid characteristics, ensuring that the cleaning liquid can be sprayed at a better pressure and speed to achieve the ideal cleaning effect.

[0134] Based on the matched spray pressure parameters, the microfluidic injection system's spray pressure is set to power the delivery of the cleaning fluid. Combined with the contamination density gradient distribution data, key areas with contamination densities above a preset threshold are identified. These areas require more cleaning fluid to meet cleaning standards. Zoned injection control technology is used to implement incremental liquid delivery to key contaminated areas. This control method can more accurately increase the liquid delivery volume based on the actual conditions of the contaminated area, improving cleaning efficiency. During the delivery process, the distribution of the cleaning fluid on the lens surface is monitored in real time, including the liquid coverage, thickness, and uniformity, generating real-time liquid distribution status data. This data can reflect the spraying effectiveness of the cleaning fluid and provide a basis for subsequent cleaning effect evaluation and parameter adjustment.

[0135] Furthermore, according to the dynamic distribution parameters in the cleaning liquid distribution scheme, the spray angle parameters adapted to the contaminated area are generated through microfluidic nozzle angle adjustment processing, including:

[0136] Based on the pollution density gradient distribution data in the dynamic allocation parameters, the geometric feature data of the pollution area is extracted through the pollution area geometric analysis processing;

[0137] According to the geometric feature data of the contaminated area and the fluid diffusion characteristic data of the microfluidic channel, the nozzle's spray angle range parameters are matched through a dynamic angle optimization algorithm;

[0138] Based on the injection angle range parameters and the spatial distribution data of the area where the pollution density is higher than the preset threshold, the injection angle parameters adapted to the pollution area are generated through adaptive adjustment processing.

[0139] Specifically, the contamination density gradient distribution data is obtained from the dynamic allocation parameters. This data describes the variation in contamination density across different areas of the lens surface. By performing geometric analysis on the contaminated areas, the geometric features of the contaminated areas, including their shape, size, location, and boundary information, are extracted. This provides a more precise geometric basis for subsequent nozzle angle adjustments.

[0140] The extracted geometric feature data of the contaminated area is then combined with the fluid diffusion characteristics of the microfluidic channel. This fluid diffusion characteristic data covers the flow characteristics of the cleaning fluid in the microfluidic channel, such as diffusion range, flow velocity, and direction. A dynamic angle optimization algorithm, taking into account the geometric characteristics of the contaminated area and the fluid diffusion characteristics, calculates the appropriate nozzle spray angle range parameters to ensure that the cleaning fluid effectively covers the contaminated area.

[0141] Next, based on the spray angle range parameters and the spatial distribution data of areas with contamination density above a preset threshold, an adaptive adjustment process is performed to generate spray angle parameters that are tailored to the contaminated areas. This step fine-tunes the spray angle based on the specific spatial distribution of highly contaminated areas to ensure that the cleaning fluid is sprayed more accurately in these areas, achieving a more efficient cleaning effect.

[0142] Furthermore, the cleaning is completed based on the adjusted cleaning execution plan, and cleaning status data is generated, including:

[0143] Based on the incremental cleaning parameters in the adjusted cleaning execution plan, collecting intermediate image data of the lens surface after cleaning through multispectral analysis and processing;

[0144] Perform residual pollution quantification processing on the intermediate image data and the coordinate information of the initial contaminated area to generate residual pollution coverage area data;

[0145] Perform dynamic parameter matching processing based on the residual contamination coverage area data and the preset cleanliness threshold to generate cleanliness assessment results;

[0146] Based on the cleanliness assessment results and the dynamic adjustment factors in the incremental cleaning parameters, cleaning status data is generated through a feedback optimization algorithm.

[0147] Specifically, based on the incremental cleaning parameters in the adjusted cleaning plan, multispectral analysis and processing technology is used to scan the cleaned lens surface and collect intermediate image data that reflects the surface contamination. Multispectral analysis can obtain information from different spectral bands, making the characteristics of contaminated areas more distinct and providing rich data for subsequent contamination quantification.

[0148] The collected intermediate image data is compared with the coordinates of the initial contaminated areas. Image processing algorithms are used to quantify the residual contamination and generate residual contamination coverage data that more accurately represents the extent of the residual contamination. This process uses image recognition technology to determine which areas still contain contamination and calculate the area of ​​these areas.

[0149] The residual contamination coverage data is compared with a preset cleanliness threshold. Dynamic parameter matching is used to comprehensively consider factors such as contamination type, contamination level, and cleaning requirements to generate a cleanliness assessment result that reflects the cleaning effect. The cleanliness threshold is an indicator set according to actual application requirements to determine whether the cleaning has met the expected standards.

[0150] Combining cleanliness assessment results with the dynamic adjustment factors in the incremental cleaning parameters, a feedback optimization algorithm is applied to evaluate and summarize the entire cleaning process, generating cleaning status data. The feedback optimization algorithm analyzes parameter changes and cleaning effectiveness during the cleaning process, while the dynamic adjustment factors optimize subsequent cleaning operations based on real-time feedback from the cleaning process. The cleaning status data reflects the cleanliness of the lens surface and the efficiency of the cleaning process.

[0151] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0152] In one embodiment, Figure 3 As shown, the present application also provides a microfluidic cleaning system 300 for drone lenses based on stain recognition, and the system 300 includes:

[0153] The stain detection and pre-quantification module 301 is used to obtain the coordinate information of the contaminated area and the contamination level data on the lens surface through contamination detection and quantification processing;

[0154] A cleaning liquid dynamic allocation module 302 is used to generate a cleaning liquid allocation plan based on the contaminated area coordinate information and the contamination level data;

[0155] The microfluidic injection module 303 is used to perform microfluidic liquid delivery according to the cleaning liquid distribution plan and generate real-time liquid distribution status data;

[0156] Liquid coverage dynamic optimization module 304, for optimizing liquid coverage parameters based on real-time liquid distribution state data and generating an optimized liquid distribution plan;

[0157] The cleaning effect evaluation and iteration module 305 is used to evaluate the preliminary cleaning effect according to the optimized liquid distribution plan and iteratively adjust the parameters to generate an adjusted cleaning execution plan;

[0158] The cleaning closed-loop execution module 306 is configured to complete cleaning based on the adjusted cleaning execution plan and generate cleaning status data.

[0159] Specifically, the contamination detection and pre-quantification module 301 uses high-resolution imaging and multispectral analysis techniques to scan and image the drone lens surface through contamination detection and quantification, capturing changes in the lens' optical properties and identifying areas of possible contamination. Simultaneously, it uses a database of contamination signatures for comparative analysis to determine the type of contamination (e.g., dust, oil, water stains, etc.) and generate the coordinates of the contaminated area and contamination level data.

[0160] The cleaning fluid dynamic allocation module 302 generates a pollution density distribution map using pollution area coordinate information and pollution level data through pollution density distribution modeling. Based on the pollution level data and pollution density distribution map, a spatial weight allocation algorithm is used to divide cleaning priority sub-regions and generate a sub-region distribution map. Based on the pollution density gradient data in the sub-region distribution map, a fluid dynamics model is used to match the cleaning fluid spray pressure and flow parameters to generate dynamic allocation parameters. The dynamic allocation parameters are then normalized and integrated with the pollution level data to generate a cleaning fluid allocation plan.

[0161] Based on the dynamic distribution parameters of the cleaning liquid distribution plan, the microfluidic injection module 303 generates injection angle parameters tailored to the contaminated area through microfluidic nozzle angle adjustment. Based on the injection angle parameters and the fluid characteristics of the microfluidic channel, a dynamic pressure compensation algorithm is used to match the cleaning liquid spray pressure parameters. Based on the spray pressure parameters and the contamination density gradient distribution data, a zoned injection control process performs incremental liquid delivery to areas where the contamination density exceeds a preset threshold, generating real-time liquid distribution status data.

[0162] The liquid coverage dynamic optimization module 304 acquires liquid coverage area data through liquid coverage area acquisition and processing based on real-time liquid distribution data. A residual contamination analysis model is used to compare and analyze the liquid coverage area data with the initial contamination area coordinate information to generate a residual contamination distribution map. Based on the residual contamination density data in the residual contamination distribution map, it is determined whether the initial cleaning effect meets the preset cleaning standard. If the cleaning standard is not met, dynamic weighting is performed based on the residual contamination density data and the cleaning liquid activity parameters to generate adjustment parameters for the cleaning liquid spray volume and spray time. Based on the adjustment parameters, an incremental adjustment algorithm is used to recalculate the cleaning execution parameters to generate an optimized liquid distribution plan.

[0163] Based on the optimized liquid distribution plan, the cleaning effect evaluation and iteration module 305 acquires liquid coverage area data through liquid coverage area acquisition and processing. A residual contamination analysis model is used to compare and analyze the liquid coverage area data with the initial contamination area coordinate information to generate a residual contamination distribution map. Based on the residual contamination density data in the residual contamination distribution map, it is determined whether the initial cleaning effect meets the preset cleaning standard. If the cleaning standard is not met, a dynamic weight allocation process is performed based on the residual contamination density data and the cleaning liquid activity parameters to generate adjustment parameters for the cleaning liquid spray volume and spray time. Based on the adjustment parameters, an incremental adjustment algorithm is used to recalculate the cleaning execution parameters to generate an adjusted cleaning execution plan.

[0164] Based on the adjusted incremental cleaning parameters in the cleaning execution plan, the cleaning closed-loop execution module 306 collects intermediate image data of the cleaned lens surface through multispectral analysis. Residual contamination is quantified using the intermediate image data and the coordinates of the initial contaminated area to generate residual contamination coverage data. Dynamic parameter matching is performed between the residual contamination coverage data and a preset cleanliness threshold to generate a cleanliness assessment result. Based on the cleanliness assessment result and the dynamic adjustment factors in the incremental cleaning parameters, a feedback optimization algorithm is used to generate cleaning status data.

[0165] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0166] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A microfluidic cleaning method for drone lenses based on stain recognition, characterized in that: The method comprises: Obtain the coordinate information of the contaminated area and the contamination level data on the lens surface through contamination detection and quantitative processing; generating a cleaning fluid distribution plan based on the contaminated area coordinate information and the contamination level data; executing microfluidic liquid delivery according to the cleaning liquid distribution scheme to generate real-time liquid distribution status data; Optimizing liquid coverage parameters based on the real-time liquid distribution state data to generate an optimized liquid distribution plan; Evaluate the preliminary cleaning effect according to the optimized liquid distribution plan and iteratively adjust the parameters to generate an adjusted cleaning execution plan; Cleaning is completed based on the adjusted cleaning execution plan, and cleaning status data is generated.

2. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 1 is characterized in that: The step of evaluating the preliminary cleaning effect according to the optimized liquid distribution scheme and iteratively adjusting parameters to generate an adjusted cleaning execution scheme includes: Based on the optimized liquid distribution scheme, liquid coverage area data is acquired through liquid coverage area acquisition and processing; The liquid coverage area data and the initial contaminated area coordinate information are compared and analyzed using the following formula, and a residual contamination distribution map is generated using the contamination residual analysis model: D(x,y)=P(x,y)·(1-η·T(x,y)) Among them, D(x,y) represents the residual pollution distribution function, P(x,y) represents the initial pollution distribution function, T(x,y) represents the residence time function of the liquid at point (x,y), η represents the cleaning efficiency coefficient per unit time, M(x,y) represents the pollutant distribution map after multi-sensor fusion, S i (x,y) represents the pollution distribution detected by the i-th sensor, ω i represents the weight coefficient of the i-th sensor, and m represents the total number of sensors; Determining whether the preliminary cleaning effect meets a preset cleaning standard based on the residual contamination density data in the residual contamination distribution map, and obtaining a judgment result, wherein the judgment result includes whether the cleaning standard is met or not met; When the judgment result is that the cleaning standard is not met, a dynamic weight distribution process is performed based on the pollutant residual density data and the cleaning liquid activity parameter to generate adjustment parameters for the cleaning liquid spraying amount and spraying time; The cleaning execution parameters are recalculated according to the adjustment parameters through an incremental adjustment algorithm to generate the adjusted cleaning execution plan.

3. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 2 is characterized in that: The comparative analysis of the liquid coverage area data and the initial contaminated area coordinate information, and the generation of a residual contamination distribution map using a contamination residual analysis model, includes: Based on the liquid coverage area data and the initial contaminated area coordinate information, generating contamination removal difference value mapping data through overlapping area analysis processing; The following formula is used to generate residual contamination density level distribution data based on the pollution clearance difference value mapping data through density gradient calculation processing: L(x,y)=Classify(|ρ(x,y)|,{T1,T2,...,T k }) Among them, L(x,y) represents the residual pollution density level distribution data, |ρ(x,y)| represents the density gradient modulus, T1,T2,...,T k Indicates the predefined density level threshold; The residual pollution distribution map is generated based on the fusion processing of the residual pollution density level distribution data and the chemical activity decay parameters in the pollution residue analysis model.

4. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 2 is characterized in that: When the judgment result is that the cleaning standard is not met, dynamic weight distribution processing is performed based on the pollutant residual density data and the cleaning liquid activity parameter to generate adjustment parameters for the cleaning liquid spraying amount and spraying time, including: Calculating a dynamic attenuation coefficient of the chemical activity of the cleaning fluid using an activity attenuation compensation algorithm based on the regional residual level distribution in the pollutant residual density data; The following formula is used to perform weighted fusion processing based on the dynamic attenuation coefficient and the basic activity value in the cleaning liquid activity parameter to generate the cleaning liquid activity compensation weight: W c =α·D d +(1-a)·A b Among them, W c represents the cleaning fluid activity compensation weight, D d Represents the dynamic attenuation coefficient, A b represents the basic activity value in the cleaning fluid activity parameter, α represents the weighted fusion coefficient, and its value range is [0,1]; A multi-factor normalization process is performed based on the cleaning liquid activity compensation weight and the pollutant residual density data to generate adjustment parameters for the cleaning liquid spraying amount and spraying time.

5. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 1 is characterized in that: Generating a cleaning liquid distribution plan based on the contaminated area coordinate information and the contamination level data includes: Based on the coordinate information of the polluted area, a pollution density distribution map is generated through pollution density distribution modeling; Dividing the cleaning priority sub-areas according to the pollution level data and the pollution density distribution map by a spatial weight allocation algorithm to generate a sub-area distribution map; Based on the pollution density gradient data in the sub-area distribution map, the cleaning liquid spray pressure and flow parameters are matched through a fluid dynamics model to generate dynamic distribution parameters; The dynamic allocation parameters and the pollution level data are normalized and fused to generate the cleaning liquid allocation plan.

6. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 5 is characterized in that: The method of generating dynamic allocation parameters by matching the cleaning liquid spray pressure and flow parameters through a fluid dynamics model based on the pollution density gradient data in the sub-area distribution map includes: Based on the pollution density gradient data, the fluid dynamic partitioning process is used to divide the area where the pollution density is higher than a preset threshold value and the area where the pollution density is lower than the preset threshold value to generate fluid partition mapping data; According to the fluid partition mapping data, the local spraying pressure parameters of the area where the pollution density is higher than the preset threshold are matched by an adaptive pressure compensation algorithm in the fluid dynamics model; Performing flow parameter matching processing based on the local spraying pressure parameter and the fluid characteristic data of the microfluidic channel to generate initial flow distribution parameters; The initial flow distribution parameter and the pressure fluctuation suppression parameter in the adaptive pressure compensation algorithm are decoupled and fused to generate the dynamic distribution parameter.

7. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 1 is characterized in that: The performing of microfluidic liquid delivery according to the cleaning liquid distribution scheme to generate real-time liquid distribution state data includes: According to the dynamic distribution parameters in the cleaning liquid distribution scheme, the spray angle parameters adapted to the contaminated area are generated by adjusting the angle of the microfluidic nozzle; Based on the spray angle parameters and the fluid characteristic data of the microfluidic channel, the spray pressure parameters of the cleaning liquid are matched by a dynamic pressure compensation algorithm; According to the spraying pressure parameters and the pollution density gradient distribution data, incremental liquid delivery is performed on the area where the pollution density is higher than a preset threshold through partitioned injection control processing to generate the real-time liquid distribution state data.

8. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 7 is characterized in that: The method of generating a spray angle parameter adapted to the contaminated area by adjusting the angle of the microfluidic nozzle according to the dynamic distribution parameters in the cleaning liquid distribution scheme includes: Extracting the geometric feature data of the polluted area through geometric analysis of the polluted area based on the pollution density gradient distribution data in the dynamic allocation parameters; According to the geometric feature data of the contaminated area and the fluid diffusion characteristic data of the microfluidic channel, the spray angle range parameters of the nozzle are matched by a dynamic angle optimization algorithm; Based on the injection angle range parameter and the spatial distribution data of the area where the pollution density is higher than a preset threshold, the injection angle parameter adapted to the pollution area is generated through adaptive adjustment processing.

9. The microfluidic cleaning method for drone lenses based on stain recognition according to claim 1 is characterized in that: The step of completing cleaning based on the adjusted cleaning execution plan and generating cleaning status data includes: Based on the incremental cleaning parameters in the adjusted cleaning execution plan, collecting intermediate image data of the lens surface after cleaning through multispectral analysis processing; Performing residual pollution quantification processing on the intermediate image data and the initial contaminated area coordinate information to generate residual pollution coverage area data; Performing dynamic parameter matching processing based on the residual contamination coverage area data and a preset cleanliness threshold to generate a cleanliness assessment result; The cleaning status data is generated by a feedback optimization algorithm based on the cleanliness evaluation result and the dynamic adjustment factor in the incremental cleaning parameter.

10. The microfluidic cleaning system for drone lenses based on stain recognition is characterized by: The system comprises: The stain detection and pre-quantification module is used to obtain the coordinate information of the contaminated area and the contamination level data on the lens surface through contamination detection and quantification processing; a cleaning liquid dynamic allocation module, configured to generate a cleaning liquid allocation plan based on the contaminated area coordinate information and the contamination level data; a microfluidic injection module, configured to perform microfluidic liquid delivery according to the cleaning liquid distribution scheme and generate real-time liquid distribution status data; a liquid coverage dynamic optimization module, configured to optimize liquid coverage parameters based on the real-time liquid distribution state data and generate an optimized liquid distribution plan; A cleaning effect evaluation and iteration module, configured to evaluate the preliminary cleaning effect and iteratively adjust parameters according to the optimized liquid distribution scheme to generate an adjusted cleaning execution scheme; The cleaning closed-loop execution module is used to complete cleaning based on the adjusted cleaning execution plan and generate cleaning status data.

Images