Inspection unmanned aerial vehicle with self-cleaning lens mechanism

Abstract

The invention belongs to the technical field of unmanned aerial vehicles, and particularly relates to an inspection unmanned aerial vehicle with a self-cleaning lens mechanism. The invention discloses an inspection unmanned aerial vehicle with a self-cleaning lens mechanism, and aims to solve the problems of imaging quality reduction and frequent task interruption caused by lens pollution. A self-cleaning mechanism is integrated on the periphery of an optical window of the unmanned aerial vehicle, and the unmanned aerial vehicle comprises a pollution sensing subsystem, a central cooperative controller, a cleaning execution subsystem and a power management unit. A pollution comprehensive index is constructed through multi-source sensing data, and a cleaning action is intelligently triggered; the cleaning execution subsystem adopts a micro linear motor to drive a flexible cleaning head to be matched with trace cleaning liquid, low-disturbance one-way scraping cleaning is achieved, the time consumed in the whole process does not exceed 1.2 seconds, and the cleaning execution subsystem cooperates with a flight control system to guarantee flight stability. According to the scheme, the task continuity and the imaging reliability of the unmanned aerial vehicle in severe environments such as high humidity and high dust are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of inspection drone technology, and in particular to an inspection drone with a self-cleaning lens mechanism. Background Technology

[0002] Driven by the technological wave, drones have rapidly penetrated various industries from their initial military and consumer applications. Among them, "inspection drones," as a mainstay of industrial applications, are reshaping traditional operation and inspection models in unprecedented ways. Leveraging their unique aerial advantages and intelligent core, they have become key technological tools for ensuring public safety, improving operational efficiency, and driving industrial upgrading. Inspection drones are unmanned aerial vehicle systems specifically designed to perform equipment, facility, or environmental inspection tasks. Their core value lies in replacing or assisting manual labor, entering high-risk, high-altitude, or high-efficiency scenarios that are "inaccessible or unsuitable for human access." Equipped with diverse payloads such as high-definition visible light cameras, thermal imaging cameras, lidar, and multispectral sensors, they transform into "aerial eyes" integrating vision, thermal sensing, and three-dimensional perception, capturing details from a completely new perspective. This transforms traditional inspection work, which relies on experience and physical strength, into a digital, precise, and intelligent modern operation.

[0003] In current technological practices, to ensure image acquisition quality, mainstream inspection drones generally employ high-performance optical lenses, supplemented by protective covers or foldable lens caps. When not in operation, these lenses physically shield the optical windows to reduce the adhesion of environmental contaminants such as dust, moisture, and oil. Some high-end models further introduce hydrophobic and oleophobic coating technology, attempting to slow down the contaminant deposition rate through surface energy regulation. Such designs did alleviate lens contamination problems to some extent in early applications, especially under short-duration, low-frequency, and relatively clean operating conditions, maintaining basic image clarity. However, as inspection tasks expand towards higher frequencies, longer flight times, and more complex and harsh environments (such as high humidity, high salt spray, strong dust, and corrosive chemical atmospheres), the inherent limitations of these passive protection strategies are becoming increasingly apparent.

[0004] Fundamentally, existing protection mechanisms are essentially static defense systems, their core logic based on "avoiding contamination" rather than "active removal." Once the protective shield is activated for mission execution, the lens is completely exposed to the external environment. Hydrophobic coatings can only slow the spread of liquid contaminants and cannot effectively address the electrostatic adsorption or dry deposition of solid particles (such as dust, coal ash, and pollen). More importantly, during continuous multi-point inspections or long-term hovering observations, the cumulative effect of tiny contaminants inevitably leads to optical path attenuation, manifesting as decreased image contrast, blurred edges, thermal imaging temperature difference distortion, and even missing laser point cloud data. To address this problem, existing solutions often rely on manual intervention—that is, operators wiping and cleaning after mission interruptions. This not only severely disrupts operational continuity but also introduces new safety risks in high-altitude, high-pressure, or toxic environments. Even when some systems attempt to incorporate compressed gas purging or miniature wiper structures, the added structural complexity, energy consumption burden, and potential mechanical failure points make reliable integration on compact UAV platforms difficult. Of particular note is that such auxiliary mechanisms typically lack the ability to perceive and adaptively respond to the level of contamination in real time, which can easily lead to insufficient cleaning or excessive movement, thereby affecting flight stability and mission timeliness.

[0005] In-depth analysis reveals that the aforementioned predicament stems from a deep-seated technological contradiction: the stringent cleanliness requirements of the inspection drone's optical sensing system conflict fundamentally with its inevitable long-term exposure in dynamic, open, and uncontrollable operating environments. Traditional approaches attempt to bridge this gap by strengthening front-end protection or manual back-end maintenance, neglecting the necessity of building a closed-loop self-cleaning capability at the system architecture level—a "perception-judgment-execution" system. In other words, existing technological approaches fail to integrate lens cleanliness into the drone's autonomous decision-making and task execution, resulting in a high dependence on external conditions and human intervention for sensing reliability. This, in turn, limits the robustness and intelligence of the inspection system in truly unmanned, all-weather, and all-terrain scenarios.

[0006] Therefore, how to construct an integrated self-cleaning lens mechanism that can sense lens contamination in real time and autonomously trigger precise, efficient, and low-disturbance cleaning actions based on task requirements without significantly increasing system size, weight, and power consumption has become a key technical challenge to overcome the current environmental adaptability bottleneck of inspection drones and achieve highly reliable continuous operation capabilities. Summary of the Invention This invention provides an inspection drone with a self-cleaning lens mechanism, aiming to solve the technical problems of decreased image quality, frequent mission interruptions, and limited environmental adaptability caused by lens contamination in existing technologies. To achieve the above-mentioned objectives, this invention proposes a closed-loop self-cleaning lens mechanism that integrates contamination perception, status assessment, and precise execution. This mechanism achieves real-time monitoring, intelligent judgment, and low-disturbance active cleaning of optical window contamination without significantly increasing the overall size, weight, and power consumption. This ensures the continuity, reliability, and imaging accuracy of inspection missions in complex and harsh environments such as high humidity, high salt spray, strong dust, or corrosive atmospheres.

[0007] The inspection drone includes an airframe structure, flight control system, power system, communication module, and multi-sensing payload system. The multi-sensing payload system includes at least one optical imaging unit, which has a lens assembly with an optical window at its front end. The core of this invention lies in integrating a self-cleaning lens mechanism around the optical window. This mechanism consists of four parts: a pollution sensing subsystem, a central coordinating controller, a cleaning execution subsystem, and a power management unit. Each part is connected via hardwired connections to a dedicated communication bus for data interaction and command synchronization.

[0008] The pollution sensing subsystem includes an optical pollution detection array and an environmental parameter sensing module. The optical pollution detection array consists of four miniature photodiodes, positioned at the top, bottom, left, and right of the outer edge of the optical window. Each photodiode faces a reference light source at a fixed angle. This reference light source is an 850nm infrared LED with a radiation intensity of 850mW / sr and a half-intensity angle of ±25°. It is installed inside the UAV body, and the optical path is led to the back of the optical window through a non-imaging area. When pollutants are present on the surface of the optical window, the transmitted light intensity attenuates, and the output current value of each photodiode changes accordingly. This current signal is converted into a voltage signal by a transimpedance amplifier, sampled by a 12-bit analog-to-digital converter, and sent to the central coordinating controller. The environmental parameter sensing module includes a temperature and humidity sensor, a particulate matter concentration sensor, and a pressure sensor, used to collect the relative humidity, PM2.5 / PM10 concentration, and atmospheric pressure of the operating environment, respectively. Its output signal is connected to the central coordinating controller via an I²C bus.

[0009] The central coordinating controller employs an embedded microcontroller based on an ARM Cortex-M7 core, with a main frequency of 480MHz, a built-in 64KB instruction cache and 32KB data cache, and runs the FreeRTOS real-time operating system. This controller receives multi-source sensor data from the pollution sensing subsystem and executes a pollution state assessment algorithm. The pollution state assessment algorithm first performs differential processing on the photodiode signals from the four directions to calculate the lateral pollution gradient ΔH = |V_left - V_right| and the longitudinal pollution gradient ΔV = |V_top - V_bottom|. Simultaneously, it combines the particulate matter concentration C_p and relative humidity R_h from the environmental parameters to construct a comprehensive pollution index P_index = α·(1 - V_avg / V_0) + β·C_p + γ·R_h, where V_avg is the average value of the four voltages, V_0 is the reference voltage under clean conditions, and α, β, and γ are preset weighting coefficients with values ​​of 0.6, 0.25, and 0.15, respectively. When P_index exceeds the first threshold T1 (set to 0.35), it is determined to be lightly contaminated, and the warning mode is activated. When P_index exceeds the second threshold T2 (set to 0.65), it is determined to be heavily contaminated, and the cleaning execution command is triggered.

[0010] The cleaning execution subsystem includes a miniature linear motor, a cleaning arm assembly, a cleaning media supply unit, and a return spring mechanism. The miniature linear motor is a voice coil motor with a rated thrust of 0.8N, a stroke of 5mm, and a response time ≤8ms. Its output shaft is fixed to the base of the cleaning arm assembly via a threaded connection. The cleaning arm assembly is made of 6061-T6 aluminum alloy and has an L-shaped structure. One end is hinged to a rotating shaft on the side wall of the optical window, and the other end is equipped with a flexible cleaning head. This cleaning head consists of a medical-grade silicone scraper encased in polyurethane foam (Shore hardness 35A), with a scraper thickness of 0.3mm and a blade... The radius of curvature of the nozzle is 0.1 mm. The cleaning medium supply unit includes a micro reservoir, a piezoelectric micropump, and a flow guide capillary. The reservoir has a volume of 0.8 mL and is filled with a cleaning solution of isopropanol and deionized water in a volume ratio of 7:3. The piezoelectric micropump operates at a frequency of 120 Hz and has a single spray volume of 5 μL. Its outlet is connected to the micropore array on the back of the cleaning head through a polytetrafluoroethylene flow guide capillary with an inner diameter of 0.2 mm. The reset spring mechanism is a torsion spring structure, installed at the rotating shaft of the cleaning arm, with a preload torque of 0.015 N·m, used to automatically reset the cleaning arm to the storage position after the cleaning action is completed.

[0011] The operation flow of the cleaning execution subsystem is as follows: When the central coordinating controller issues a cleaning command, the piezoelectric micropump is first activated to spray cleaning fluid onto the back of the cleaning head once, lasting for 100ms. Subsequently, the micro linear motor is powered on and driven to rotate the cleaning arm around the axis, causing the cleaning head to slide unidirectionally from top to bottom along the surface of the optical window at a constant speed, covering the entire effective imaging area at a sliding speed of 25mm / s. During the sliding process, the silicone scraper adheres closely to the surface of the optical window to scrape off the attached contaminants, while the polyurethane foam absorbs the residual liquid. After the sliding is completed, the micro linear motor is de-energized, and the reset spring mechanism drives the cleaning arm to spring back to the initial storage position, completing one cleaning cycle. The entire cleaning process takes no more than 1.2 seconds. During this period, the flight control system synchronously adjusts the attitude control parameters to compensate for the center of gravity shift caused by the movement of the cleaning arm, ensuring flight stability.

[0012] The power management unit is integrated into the main power distribution board of the UAV and adopts a two-stage power supply architecture. The first stage is a 5V DC bus, supplied by the UAV's main battery via a DC-DC step-down module. The second stage consists of two regulated outputs: 3.3V and 12V, used for the pollution sensing subsystem and the cleaning execution subsystem, respectively. The 12V output is only turned on by the MOSFET switch during the cleaning execution phase and is turned off at other times to reduce standby power consumption. The power management unit also has a current monitoring circuit that collects the operating current of the cleaning execution subsystem in real time. If the current abnormally rises above 150mA, it is determined to be a mechanical jam, and the 12V power supply is immediately cut off and a fault alarm is sent to the flight control system.

[0013] Furthermore, the optical window is made of sapphire glass with a thickness of 1.2mm and a double-layer composite film coating. The bottom layer is a silica hydrophobic layer with a contact angle ≥110°, and the top layer is a magnesium fluoride anti-reflective layer with an average transmittance of ≥98.5% in the visible light band. The optical window is sealed to the lens barrel by an O-ring fluororubber seal (hardness 70 Shore A). The seal has a cross-sectional diameter of 1.5mm and a compression rate of 25%, ensuring that no liquid seeps into the lens during the cleaning fluid spraying process.

[0014] In a preferred embodiment of the present invention, the reference light source in the pollution sensing subsystem operates in pulse modulation mode with a modulation frequency of 1kHz and a duty cycle of 50% to eliminate ambient light interference; the photodiode signal acquisition and analog-to-digital conversion operations are completed within the high-level phase of each modulation cycle to ensure that the signal-to-noise ratio is not less than 45dB.

[0015] In another preferred embodiment of the present invention, the central collaborative controller is connected to the UAV flight control system via a CAN bus with a baud rate of 1Mbps. When a heavy pollution cleaning command is triggered, the central collaborative controller sends a "cleaning in progress" status flag to the flight control system. The flight control system then suspends the image acquisition task and switches the current flight mode to the hovering stable mode. At the same time, it records the timestamp and pollution index of the cleaning event for subsequent task log analysis.

[0016] In another preferred embodiment of the present invention, the liquid storage chamber of the cleaning medium supply unit is equipped with a liquid level detection electrode. When the remaining amount of cleaning liquid is less than 0.2 mL, a low liquid level signal is triggered, the central coordinating controller will prohibit subsequent cleaning actions, and send a liquid replenishment prompt to the ground station through the communication module.

[0017] The self-cleaning lens mechanism is installed in the annular bracket at the front end of the lens assembly, without exceeding the outline of the original protective cover; its maximum instantaneous power consumption is 1.8W, and its average standby power consumption is 12mW, which meets the stringent constraints of small inspection drones on payload weight and energy consumption.

[0018] In summary, this invention, by constructing a closed-loop control architecture of contamination perception—state assessment—precise execution, achieves real-time quantitative identification and on-demand cleaning of lens contamination status, completely eliminating the traditional passive protection and manual intervention mode. The self-cleaning lens mechanism is highly integrated in structure, environmentally adaptive in control, and ensures low disturbance and high reliability in execution, fundamentally solving the problem of perception failure caused by lens contamination in inspection drones in complex and harsh environments, and significantly improving the system's all-weather operation capability, task continuity, and intelligence level. Attached Figure Description

[0019] Figure 1 This is a block diagram of the overall structure of the present invention; Figure 2 This is a block diagram of the cleaning execution subsystem of the present invention; Figure 3 This is a flowchart illustrating the cleaning execution state of the present invention. Detailed Implementation To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. In addition, the terms "first," "second," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] This invention discloses an inspection drone with a self-cleaning lens mechanism, the specific implementation of which is described below. The overall structure of the inspection drone includes a fuselage, a flight control system, a power system, a communication module, and a multi-sensor payload system. The multi-sensor payload system includes at least one optical imaging unit, which has a lens assembly with an optical window at its front end. A self-cleaning lens mechanism is integrated around the optical window. This mechanism consists of four parts: a contamination sensing subsystem, a central coordinating controller, a cleaning execution subsystem, and a power management unit. Each part is connected via hardwired connections to a dedicated communication bus for data interaction and command synchronization.

[0026] The pollution sensing subsystem includes an optical pollution detection array and an environmental parameter sensing module. The optical pollution detection array consists of four miniature photodiodes, positioned at the top, bottom, left, and right of the outer edge of the optical window. Each photodiode faces a reference light source at a fixed angle. This reference light source is an 850nm infrared LED with a radiation intensity of 850mW / sr and a half-intensity angle of ±25°. It is installed inside the UAV body, and the light path is led to the back of the optical window through a non-imaging area. When pollutants are present on the surface of the optical window, the transmitted light intensity attenuates, and the output current value of each photodiode changes accordingly. This current signal is converted into a voltage signal by a transimpedance amplifier, sampled by a 12-bit analog-to-digital converter, and sent to the central coordinating controller. The environmental parameter sensing module includes a temperature and humidity sensor, a particulate matter concentration sensor, and a pressure sensor, used to collect the relative humidity, PM2.5 / PM10 concentration, and atmospheric pressure of the operating environment, respectively. Its output signals are connected to the central coordinating controller via an I²C bus.

[0027] In a preferred embodiment of the present invention, the reference light source operates using pulse modulation with a modulation frequency of 1 kHz and a duty cycle of 50% to eliminate ambient light interference. The photodiode signal acquisition and analog-to-digital conversion operations are completed within the high-level phase of each modulation cycle, ensuring a signal-to-noise ratio of no less than 45 dB.

[0028] The central coordinating controller employs an embedded microcontroller based on an ARM Cortex-M7 core, with a main frequency of 480MHz, a built-in 64KB instruction cache and 32KB data cache, and runs the FreeRTOS real-time operating system. This controller receives multi-source sensor data from the pollution sensing subsystem and executes a pollution state assessment algorithm. The pollution state assessment algorithm first performs differential processing on the photodiode signals from the four directions to calculate the lateral pollution gradient ΔH = |V_left - V_right| and the longitudinal pollution gradient ΔV = |V_top - V_bottom|. Simultaneously, it combines the particulate matter concentration C_p and relative humidity R_h from the environmental parameters to construct a comprehensive pollution index P_index = α·(1 - V_avg / V_0) + β·C_p + γ·R_h, where V_avg is the average of the four voltages, V_0 is the reference voltage under clean conditions, and α, β, and γ are preset weighting coefficients with values ​​of 0.6, 0.25, and 0.15, respectively. When P_index exceeds the first threshold T1 (set to 0.35), it is determined to be lightly contaminated and the warning mode is activated; when P_index exceeds the second threshold T2 (set to 0.65), it is determined to be heavily contaminated and the cleaning execution command is triggered.

[0029] In another preferred embodiment of the present invention, the central collaborative controller is connected to the UAV flight control system via a CAN bus with a baud rate of 1Mbps. When a heavy pollution cleaning command is triggered, the central collaborative controller sends a "cleaning in progress" status flag to the flight control system. The flight control system then suspends the image acquisition task and switches the current flight mode to a hovering stable mode. At the same time, it records the timestamp and pollution index of the cleaning event for subsequent task log analysis.

[0030] The cleaning execution subsystem includes a miniature linear motor, a cleaning arm assembly, a cleaning media supply unit, and a return spring mechanism. The miniature linear motor is a voice coil motor with a rated thrust of 0.8 N, a stroke of 5 mm, and a response time ≤8 ms. Its output shaft is fixed to the base of the cleaning arm assembly via a threaded connection. The cleaning arm assembly is made of 6061-T6 aluminum alloy and has an L-shaped structure. One end is hinged to a rotating shaft on the side wall of the optical window, and the other end is fitted with a flexible cleaning head. This cleaning head consists of a medical-grade silicone scraper encased in polyurethane foam (Shore hardness 35A), with a scraper thickness of 0.3 mm and a blade curvature radius of 0.1 mm. The cleaning media supply unit includes a miniature reservoir, a piezoelectric micropump, and a flow-guiding capillary. The reservoir has a volume of 0.8 mL and is filled with a cleaning solution of isopropanol and deionized water at a volume ratio of 7:3. The piezoelectric micropump operates at a frequency of 120Hz and has a single spray volume of 5μL. Its outlet is connected to the micropore array on the back of the cleaning head via a 0.2mm inner diameter PTFE capillary tube. The reset spring mechanism is a torsion spring structure, installed at the rotating shaft of the cleaning arm, with a preload torque of 0.015N·m, used to automatically reset the cleaning arm to the storage position after the cleaning action is completed.

[0031] The operation flow of the cleaning execution subsystem is as follows: When the central coordinating controller issues a cleaning command, the piezoelectric micropump is first activated to spray cleaning fluid onto the back of the cleaning head once, lasting for 100ms. Subsequently, the micro linear motor is powered on and driven to rotate the cleaning arm around the axis, causing the cleaning head to slide unidirectionally from top to bottom along the surface of the optical window at a constant speed, covering the entire effective imaging area at a sliding speed of 25mm / s. During the sliding process, the silicone scraper adheres closely to the surface of the optical window to scrape off the attached contaminants, while the polyurethane foam absorbs the residual liquid. After the sliding is completed, the micro linear motor is de-energized, and the reset spring mechanism drives the cleaning arm to spring back to the initial storage position, completing one cleaning cycle. The entire cleaning process takes no more than 1.2 seconds. During this period, the flight control system synchronously adjusts the attitude control parameters to compensate for the center of gravity shift caused by the movement of the cleaning arm, ensuring flight stability.

[0032] In another preferred embodiment of the present invention, the liquid storage chamber of the cleaning medium supply unit is equipped with a liquid level detection electrode. When the remaining amount of cleaning liquid is less than 0.2 mL, a low liquid level signal is triggered, the central coordinating controller will prohibit subsequent cleaning actions, and send a liquid replenishment prompt to the ground station through the communication module.

[0033] The power management unit is integrated into the UAV's main power distribution board and adopts a two-stage power supply architecture. The first stage is a 5V DC bus, supplied by the UAV's main battery via a DC-DC step-down module; the second stage consists of two regulated outputs, 3.3V and 12V, used for the pollution sensing subsystem and the cleaning execution subsystem, respectively. The 12V output is only turned on by a MOSFET switch during the cleaning execution phase, remaining off at other times to reduce standby power consumption. The power management unit also includes a current monitoring circuit that collects the operating current of the cleaning execution subsystem in real time. If the current abnormally rises above 150mA, it is determined to be a mechanical jam, and the 12V power supply is immediately cut off, sending a fault alarm to the flight control system.

[0034] The optical window is made of sapphire glass with a thickness of 1.2 mm and a double-layer composite coating. The bottom layer is a hydrophobic silica layer with a contact angle ≥110°, and the top layer is a magnesium fluoride anti-reflective layer with an average transmittance of ≥98.5% in the visible light band. The optical window is sealed to the lens barrel by an O-ring fluororubber seal (hardness 70 Shore A). The seal has a cross-sectional diameter of 1.5 mm and a compression rate of 25%, ensuring that no liquid seeps into the lens during the cleaning fluid spraying process.

[0035] The self-cleaning lens mechanism has overall dimensions of 32mm×28mm×15mm and weighs 8.7g. It is installed in the ring bracket at the front of the lens assembly without exceeding the outline of the original protective cover. Its maximum instantaneous power consumption is 1.8W and the average standby power consumption is 12mW, which meets the stringent constraints of payload weight and energy consumption for small inspection drones.

[0036] In one specific embodiment, an inspection drone equipped with the self-cleaning lens mechanism described in this invention performs routine inspection tasks at a coastal wind farm. The operating environment is characterized by complex meteorological conditions including high salt spray, high humidity (85% relative humidity), and wind speed of 6 m / s. After takeoff, the pollution sensing subsystem continuously collects the transmitted light intensity and environmental parameters through the optical window. Initially, the output voltages of the four photodiodes are V_top=2.48V, V_bottom=2.47V, V_left=2.49V, V_right=2.48V, V_avg=2.48V, and V_0=2.50V; the particulate matter concentration is C_p=120μg / m³ (PM10), and the relative humidity is R_h=85%. Substituting these values ​​into the comprehensive pollution index formula: P_index = 0.6 × (1 - 2.48 / 2.50) + 0.25 × (120 / 500) + 0.15 × (85 / 100) = 0.6 × 0.008 + 0.25 × 0.24 + 0.15 × 0.85 = 0.0048 + 0.06 + 0.1275 = 0.1923 When the value is below the first threshold T1 (0.35), the system maintains normal imaging mode.

[0037] Approximately 25 minutes into the flight, salt spray accumulated on the surface of the optical window, causing a decrease in light transmittance. At this point, V_top = 2.10V, V_bottom = 2.05V, V_left = 2.30V, V_right = 2.32V, and V_avg = 2.1925V were measured; C_p increased to 180 μg / m³, and R_h = 88%. Calculations yielded: ΔH = |2.30 - 2.32| = 0.02V ΔV = |2.10 - 2.05| = 0.05V P_index = 0.6 × (1 - 2.1925 / 2.50) + 0.25 × (180 / 500) + 0.15 × (88 / 100) = 0.6 × 0.123 + 0.25 × 0.36 + 0.15 × 0.88 = 0.0738 + 0.09 + 0.132 = 0.2958 The value was still below T1, but the vertical gradient ΔV increased significantly, indicating that contaminants were mainly deposited in the lower part of the window. The system recorded this trend but did not trigger cleaning.

[0038] Continuing flight to the 40th minute, V_top=1.85V, V_bottom=1.70V, V_left=2.10V, V_right=2.12V, V_avg=1.9425V; C_p=210μg / m³, R_h=90%. Calculations show: P_index = 0.6 × (1 - 1.9425 / 2.50) + 0.25 × (210 / 500) + 0.15 × 0.90 = 0.6 × 0.223 + 0.25 × 0.42 + 0.135 = 0.1338 + 0.105 + 0.135 = 0.3738 When the value exceeds T1 (0.35), the system enters early warning mode, sends a "light lens pollution" alert to the ground station, and increases the pollution sensing sampling frequency to once every 5 seconds.

[0039] After 10 minutes, V_bottom further decreased to 1.50V, V_avg=1.76V, C_p=230μg / m³, R_h=92%, then: P_index = 0.6 × (1 - 1.76 / 2.50) + 0.25 × (230 / 500) + 0.15 × 0.92 = 0.6 × 0.296 + 0.25 × 0.46 + 0.138 = 0.1776 + 0.115 + 0.138 = 0.4306 Although T2 was not reached, ΔV = |1.85 - 1.50| = 0.35V, which is much higher than the historical average. The system judged that there was a risk of localized severe pollution and triggered cleaning preparations in advance.

[0040] After another 8 minutes, V_bottom = 1.30V, V_avg = 1.60V, C_p = 250μg / m³, R_h = 93%, and the calculations are as follows: P_index = 0.6 × (1 - 1.60 / 2.50) + 0.25 × 0.5 + 0.15 × 0.93 = 0.6 × 0.36 + 0.125 + 0.1395 = 0.216 + 0.125 + 0.1395 = 0.4805 It still did not exceed T2. However, due to the continuous deterioration of the vertical gradient and the obvious blurring of the image edges, the system actively triggered a cleaning command when P_index=0.48 according to the preset strategy (this is the dynamic threshold mechanism after strategy optimization, which is not the core of this embodiment and is only used for illustration).

[0041] Cleaning process: The piezoelectric micropump operates for 100ms, spraying 5μL of cleaning fluid; the micro linear motor drives the cleaning arm to slide down at a speed of 25mm / s, with a stroke of 22mm (covering the window height), taking 0.88 seconds; the return spring takes 0.25 seconds to rebound; the total time is 1.13 seconds. Immediately after cleaning, the following measurements were taken: V_top=2.45V, V_bottom=2.42V, V_left=2.47V, V_right=2.46V, V_avg=2.45V, P_index dropped back to 0.21, and image clarity was restored.

[0042] To verify the technical effectiveness of this invention, a comparative test was conducted using a comparative example.

[0043] Comparative Example: Using the same model of inspection drone, but with the lens assembly only equipped with a passive protective cover (no self-cleaning mechanism), an inspection mission of the same duration (60 minutes) was performed in the same coastal wind farm environment. Before the mission began, the lens was clean, and V_avg = 2.50V. After 30 minutes, V_avg dropped to 2.20V, and P_index ≈ 0.32; after 45 minutes, V_avg = 1.90V, and P_index ≈ 0.45; at the end of 60 minutes, V_avg = 1.65V, and P_index = 0.51, the image was severely blurred, and the edge detail loss rate exceeded 60%, forcing the mission to be aborted and the drone to return to base.

[0044] The performance comparison data between the examples and the comparative examples are shown in the table below:

[0045] The above data demonstrates that the self-cleaning lens mechanism described in this invention can effectively maintain the cleanliness of the optical window in complex and harsh environments, ensuring the continuity and stability of imaging quality. The cleaning process is fast and low-disturbance, without affecting flight safety; the contamination perception and assessment algorithm has good environmental adaptability and discrimination accuracy; the overall mechanism is lightweight and low-power, making it fully compatible with small inspection drone platforms.

[0046] Furthermore, the components of the self-cleaning lens mechanism must meet the following process requirements during manufacturing and assembly: the sealed assembly of the optical window and lens barrel adopts a laser welding-assisted positioning fixture to ensure that the O-ring compression rate is strictly controlled within 25%±2%; the cleaning arm shaft adopts a precision-ground stainless steel shaft with a surface roughness Ra≤0.2μm and a fitting clearance of 0.01mm to ensure smooth rotation without shaking; the threaded connection between the micro linear motor and the cleaning arm base is cured with anaerobic adhesive to prevent vibration-induced loosening; the flow guide capillary and the micropore array on the back of the cleaning head are sealed together through a hot-melt embedding process, with a leakage rate of less than 1×10⁻⁶. -6Pa·m³ / s; all electronic components are coated with conformal coating (acrylate-based) with a thickness of 25–35 μm, conforming to IPC-CC-830B Class B standard.

[0047] At the software level, the firmware of the central collaborative controller includes the following functional modules: a sensor driver module, a data filtering module (using a combination of moving average and Kalman filtering), a pollution index calculation module, a threshold judgment and state machine module, a cleaning execution timing control module, a fault diagnosis module, and a communication interface module. Specifically, the pollution index calculation module performs a complete evaluation every 200ms; the cleaning execution timing control module precisely controls the piezoelectric micropump start-up time, the motor drive current ramp-up / down curve (to avoid impact), and the release timing of the reset spring; the fault diagnosis module continuously monitors the motor current, pump operating voltage, and liquid level signal, and immediately initiates a safe shutdown procedure upon detecting an anomaly.

[0048] Furthermore, to adapt to different application scenarios, the weighting coefficients α, β, and γ in the pollution state assessment algorithm can be adjusted online based on environmental characteristics. For example, in a desert dust environment, β can be increased to 0.35 and α decreased to 0.5; in a humid tropical rainforest, γ can be increased to 0.25 and β decreased to 0.15. This adjustment can be configured remotely via a ground station or automatically loaded with a preset parameter set by the flight control system based on GPS positioning.

[0049] In summary, this invention, through its highly integrated electromechanical structure design, multi-source fusion pollution perception method, closed-loop feedback intelligent decision-making mechanism, and rapid, low-interference execution strategy, achieves autonomous, reliable, and efficient cleaning of inspection drone lenses in complex and harsh environments. This provides solid technical support for all-weather intelligent inspection of unmanned systems in critical infrastructure sectors such as power, oil and gas, transportation, and environmental protection. Those skilled in the art can make various modifications and equivalent substitutions to the above embodiments without departing from the spirit and scope of this invention, but all such modifications and substitutions should fall within the protection scope defined by the claims of this invention.

[0050] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An inspection drone with a self-cleaning lens mechanism, comprising a fuselage structure, a flight control system, a power system, a communication module, and a multi-sensing payload system, wherein the multi-sensing payload system includes at least one optical imaging unit, the optical imaging unit having a lens assembly, and the front end of the lens assembly having an optical window; characterized in that, A self-cleaning lens mechanism is integrated around the optical window. This mechanism includes a contamination sensing subsystem, a central coordinating controller, a cleaning execution subsystem, and a power management unit. The contamination sensing subsystem includes an optical contamination detection array and an environmental parameter sensing module. The optical pollution detection array consists of four miniature photodiodes, which are arranged at the top, bottom, left, and right of the outer edge of the optical window. Each photodiode faces a reference light source at a fixed angle. The reference light source is an infrared light-emitting diode with a wavelength of 850nm, which is installed inside the UAV body and the light path is led to the back of the optical window through the non-imaging area. The environmental parameter sensing module includes a temperature and humidity sensor, a particulate matter concentration sensor, and an air pressure sensor. The central collaborative controller receives multi-source sensor data from the pollution sensing subsystem and executes a pollution status assessment algorithm. When the comprehensive pollution index exceeds a preset threshold, it triggers a cleaning execution command. The cleaning execution subsystem includes a miniature linear motor, a cleaning arm assembly, a cleaning medium supply unit, and a reset spring mechanism. The output shaft of the miniature linear motor is fixedly connected to the base of the cleaning arm assembly. One end of the cleaning arm assembly is hinged to the rotating shaft of the optical window sidewall, and the other end is equipped with a flexible cleaning head. The flexible cleaning head is composed of a medical-grade silicone scraper wrapped with polyurethane foam. The cleaning medium supply unit includes a miniature liquid storage chamber, a piezoelectric micropump, and a flow guide capillary. The piezoelectric micropump sprays cleaning liquid onto the back of the cleaning head through the flow guide capillary. The reset spring mechanism is a torsion spring structure, installed at the rotating shaft of the cleaning arm, used to automatically reset the cleaning arm to the storage position after the cleaning action is completed; the power management unit adopts a two-stage power supply architecture, in which the 12V output is only turned on during the cleaning execution phase and turned off at other times.

2. The inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The pollution status assessment algorithm calculates the lateral pollution gradient ΔH = |V_left - V_right| and the longitudinal pollution gradient ΔV = |V_top - V_bottom|, and combines the particulate matter concentration C_p and relative humidity R_h to construct a comprehensive pollution index P_index = α·(1 - V_avg / V_0) + β·C_p + γ·R_h, where V_avg is the average value of the four voltages, V_0 is the reference voltage under clean conditions, and α, β, and γ are 0.6, 0.25, and 0.15, respectively. When P_index exceeds the first threshold T1 = 0.35, it is judged as light pollution; when P_index exceeds the second threshold T2 = 0.65, it is judged as heavy pollution and a cleaning execution command is triggered.

3. The inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The reference light source operates using pulse modulation with a modulation frequency of 1kHz and a duty cycle of 50%. The photodiode signal acquisition and analog-to-digital conversion operations are completed within the high-level phase of each modulation cycle.

4. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The central coordinating controller is connected to the UAV flight control system via a CAN bus with a baud rate of 1Mbps. When a heavy pollution cleaning command is triggered, the central coordinating controller sends a "cleaning in progress" status flag to the flight control system, which then suspends the image acquisition task and switches the flight mode to hovering stable mode.

5. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The liquid storage chamber of the cleaning medium supply unit is equipped with a liquid level detection electrode. When the remaining amount of cleaning liquid is less than 0.2 mL, a low liquid level signal is triggered, the central coordinating controller prohibits subsequent cleaning actions, and sends a liquid replenishment prompt to the ground station through the communication module.

6. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The power management unit is equipped with a current monitoring circuit to collect the operating current of the cleaning execution subsystem in real time. If the current rises abnormally to more than 150mA, it is determined to be mechanical jamming, and the 12V power supply is immediately cut off and a fault alarm is sent to the flight control system.

7. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The optical window is made of sapphire glass with a thickness of 1.2 mm. The surface is coated with a double-layer composite film system. The bottom layer is a silica hydrophobic layer with a contact angle of not less than 110°, and the top layer is a magnesium fluoride anti-reflection layer with an average transmittance of not less than 98.5% in the visible light band.

8. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The optical window is sealed to the lens barrel by an O-ring fluororubber seal. The O-ring fluororubber seal has a hardness of 70 Shore A, a cross-sectional diameter of 1.5 mm, and a compression rate of 25%.

9. An inspection drone with a self-cleaning lens mechanism according to claim 1, characterized in that: The micro linear motor is a voice coil motor with a rated thrust of 0.8N, a stroke of 5mm, and a response time of no more than 8ms. The cleaning arm assembly is made of 6061-T6 aluminum alloy and has an overall L-shaped structure. The silicone scraper has a thickness of 0.3mm and a blade curvature radius of 0.1mm. The piezoelectric micropump operates at a frequency of 120Hz and has a single spray volume of 5μL. The guide capillary has an inner diameter of 0.2mm and is made of polytetrafluoroethylene.

Images