An adaptive ear canal cleaning system and method based on visual recognition

Abstract

The present application relates to ear canal cleaning instrument technical field, disclose a kind of self-adapting ear canal cleaning system and method based on visual identification.The present application is directed to the problems of single identification dimension, lack of effect evaluation and safety protection relying on host software of existing visual feedback type ear canal flushing device, the system includes: image acquisition device, including camera and fill light source, for obtaining real-time image inside ear canal;Image processing and control unit, according to real-time image recognition earwax type, degree of blockage and ear canal wall state, generate at least including pulse frequency and water flow pressure parameter flushing strategy, and reacquire image after flushing to evaluate flushing effect, adjust flushing strategy based on evaluation result;Pulse water flow execution unit, output pulse frequency and water flow pressure adjustable pulse water flow;Safety monitoring unit, independent of image processing and control unit, directly shut off pulse water flow when physical parameter is over limit.The present application is suitable for clinical ear canal cleaning and daily ear canal care.

Description

Technical Field

[0001] This invention relates to the field of ear canal cleaning device technology, and more specifically to an adaptive ear canal cleaning system and method based on visual recognition. Background Technology

[0002] Ear canal cleaning is a common procedure in clinical otolaryngology diagnosis and treatment as well as in daily personal ear canal care. It is mainly used to remove earwax, secretions, and foreign objects from the external auditory canal to maintain ear canal patency and hygiene. With the improvement of people's health awareness, ear canal cleaning equipment has evolved from traditional manual tools to electric devices with a certain degree of automation.

[0003] Currently, ear canal cleaning devices mainly include the following categories: (1) Manual tools, such as ear picks, cotton swabs, and squeeze ear cleaners, rely entirely on manual control by the operator during the cleaning process. The force and angle are difficult to control precisely, which may result in incomplete cleaning or damage to the ear canal mucosa.

[0004] (2) Visual aids, such as the external auditory canal visual constant temperature irrigator disclosed in Chinese invention patent application CN120094020A, which integrates a high-definition camera and a supplementary light at the front end of the irrigator, allowing the operator to observe the condition inside the ear canal. The irrigating fluid is provided by a manually squeezed elastic fluid supply mechanism. This type of device achieves visual assistance in the irrigating process, but the irrigating parameters are still manually adjusted by the operator, and the degree of automation is limited.

[0005] (3) Pressure-sensing adaptive types, such as the visual ear suction device disclosed in Chinese invention patent CN120114682B, collect pressure distribution data of the inner wall of the ear canal through a flexible contact array, match it with a preset ear canal feature model to calculate the target water flow pressure value and automatically adjust the water pump output. This type of solution achieves adaptive adjustment of water flow pressure based on contact pressure sensing, but does not use image recognition technology to evaluate earwax type and ear canal wall condition, and the sensor sensing data is affected by contact pressure, resulting in a relatively simple evaluation dimension.

[0006] (4) Pulse flushing based on visual feedback, such as the adaptive variable frequency pulse ear canal flushing device based on visual feedback disclosed in Chinese invention patent application CN121971232A, acquires ear canal images through a visual acquisition unit at the end of the flushing gun, identifies the target object attributes by an image processing unit, and selects the corresponding flushing mode based on the identification results to achieve automatic adjustment of pulse frequency and water pressure. This type of solution has initially achieved adaptive control capability driven by image recognition, but it has the following shortcomings: First, image recognition only distinguishes the type of target object, does not quantitatively assess the degree of blockage, and does not monitor the health status of the ear canal wall itself, so the generation of flushing strategy lacks multi-dimensional quantitative basis; Second, safety protection mainly relies on the software level control of the main controller, and does not set up a hardware-level safety monitoring mechanism independent of the main control system. When the main control system malfunctions, there is a lack of a safety baseline that can act independently; Third, although integrating the camera into the front end of the flushing gun achieves real-time image acquisition during the flushing process, this integrated structure increases the size of the front end of the flushing gun, makes handheld operation inconvenient, and liquid splashing and equipment vibration during the flushing process will interfere with the image acquisition quality and affect the accuracy of image recognition.

[0007] In summary, existing technologies have made some progress in applying image recognition technology to adaptive control of ear canal irrigation. However, further improvements are needed in the comprehensiveness of recognition dimensions, the refinement of irrigation strategies, the post-irreversible evaluation mechanism for irrigation effects, and the reliability of the safety protection system. There is an urgent need for an intelligent ear canal cleaning system and method that can comprehensively assess earwax type, degree of blockage, and ear canal wall condition, adaptively generate multi-parameter pulsed water flow irrigation strategies, achieve iterative optimization of strategies by re-evaluating the ear canal condition after irrigation, and possess independent hardware-level safety monitoring capabilities. Summary of the Invention

[0008] The present invention aims to solve the following technical problems of existing ear canal cleaning devices: First, the safety protection mechanism relies entirely on the software judgment of the main controller and does not have independent hardware-level bottom-line protection, thus resulting in insufficient flushing safety and single-point failure risk; Second, it fails to comprehensively assess the type of earwax, degree of blockage and health status of the ear canal wall before flushing to generate an optimized flushing strategy, and lacks a mechanism for evaluating the effect after flushing and iterative optimization of the strategy, making it difficult to guarantee flushing efficiency and effect.

[0009] This invention provides an adaptive ear canal cleaning system based on visual recognition, comprising: an image acquisition device including a camera and a supplementary light source, configured to acquire real-time images of the inside of the ear canal; an image processing and control unit connected to the image acquisition device, configured to: identify earwax type, degree of blockage, and ear canal wall state based on the real-time images, generate a flushing strategy based on the identification results, the flushing strategy including at least pulse frequency parameters and water pressure parameters, and after the pulse water flow execution unit performs flushing according to the flushing strategy, re-acquire real-time images of the inside of the ear canal through the image acquisition device to evaluate the flushing effect, the evaluation of the flushing effect including at least evaluating the earwax removal rate, and adjusting the flushing strategy based on the evaluation results; a pulse water flow execution unit connected to the image processing and control unit, configured to output a pulse water flow with adjustable pulse frequency and water pressure according to the flushing strategy; and a safety monitoring unit including at least one sensor independent of the image processing and control unit, configured to detect physical parameters, and when the physical parameters exceed a preset safety threshold, directly shut off the water flow output of the pulse water flow execution unit through hardware circuitry independent of the image processing and control unit.

[0010] Furthermore, the pulsed water flow execution unit includes a rinsing handle, and the image acquisition device and the rinsing handle are independent and separate mechanisms. The image acquisition device and the rinsing handle are used alternately in different operation stages. The image acquisition device is used to insert into the ear canal to acquire images before and after rinsing, and the rinsing handle is used to insert into the ear canal to perform pulsed water flow rinsing operation during rinsing.

[0011] Furthermore, the evaluation of the flushing effect includes evaluating the earwax removal rate and changes in the ear canal wall condition; the adjustment of the flushing strategy based on the evaluation results includes adjusting at least one of the pulse frequency parameter and the water flow pressure parameter according to the earwax removal rate, and reducing at least one of the water flow pressure and the pulse frequency or terminating the flushing when the ear canal wall condition is identified as congested or damaged.

[0012] Furthermore, the sensor of the safety monitoring unit is connected to the power supply circuit of the pulse water flow execution unit through an independent comparison circuit. The comparison circuit is configured to directly cut off the power supply of the power supply circuit when the physical parameter exceeds the preset safety threshold, without needing to be processed by the image processing and control unit. The physical parameter includes at least one of water flow pressure and temperature.

[0013] Furthermore, the safety monitoring unit includes a temperature sensor and a pressure sensor independent of the image processing and control unit. The temperature sensor is used to monitor the temperature of the flushing fluid, and the pressure sensor is used to monitor the water flow pressure.

[0014] Furthermore, in the flushing strategy, the pulse frequency parameter is determined according to the earwax type, wherein dry earwax corresponds to a first frequency range, wet earwax corresponds to a second frequency range, and the lower limit of the first frequency range is greater than the upper limit of the second frequency range; the water flow pressure parameter is determined according to the degree of blockage and the ear canal wall condition, wherein the water flow pressure increases with the degree of blockage, but does not exceed a preset safe pressure upper limit; when the ear canal wall condition is identified as congested or damaged, the water flow pressure is limited to below the preset safe pressure value regardless of the degree of blockage; the earwax type includes dry earwax, wet earwax, and mixed earwax; the degree of blockage is quantified as mild blockage, moderate blockage, or severe blockage based on the proportion of earwax in the ear canal cross-sectional area.

[0015] This invention also provides an adaptive ear canal cleaning method based on visual recognition, comprising the following steps: S1. Acquiring real-time images of the inside of the ear canal through an image acquisition device, the image acquisition device including a camera and a supplementary light source; S2. Performing image analysis on the real-time images to identify earwax type, degree of blockage, and ear canal wall condition; S3. Automatically generating a flushing strategy based on the recognition results, the flushing strategy including at least pulse frequency parameters and water flow pressure parameters; S4. Controlling a pulse water flow execution unit according to the flushing strategy to output pulse water flow with adjustable pulse frequency and water flow pressure to flush the ear canal; S5. After flushing, re-acquiring real-time images of the inside of the ear canal through the image acquisition device to evaluate the flushing effect, and adjusting the flushing strategy based on the evaluation results. If the flushing effect does not meet a preset standard, returning to step S4 to perform the next round of flushing with the adjusted flushing strategy; and S6. During the flushing process, independently of steps S2 to S4, monitoring physical parameters, and when the physical parameters exceed a preset safety threshold, forcibly interrupting the output of the pulse water flow through an independent hardware circuit.

[0016] Furthermore, the evaluation of the flushing effect includes evaluating the earwax removal rate and changes in the ear canal wall condition; the adjustment of the flushing strategy based on the evaluation results in step S5 includes adjusting at least one of the pulse frequency parameter and the water flow pressure parameter according to the earwax removal rate, and reducing at least one of the water flow pressure and the pulse frequency or terminating the flushing when the ear canal wall condition is identified as congested or damaged.

[0017] Furthermore, the forced interruption of the pulsed water flow output in step S6 includes directly cutting off the power supply to the pulsed water flow execution unit through an independent comparison circuit that is not connected to the signals in steps S2 to S4.

[0018] Further, the automatic flushing strategy generated in step S3 includes: determining the pulse frequency parameter based on the earwax type, wherein the pulse frequency corresponding to dry earwax is higher than that corresponding to wet earwax; determining the water flow pressure parameter based on the degree of blockage and the ear canal wall condition, wherein the water flow pressure increases with the degree of blockage but does not exceed a preset safe pressure limit; when the ear canal wall condition is identified as congested or damaged, the water flow pressure is limited to below the preset safe pressure value regardless of the degree of blockage; the earwax type includes dry earwax, wet earwax, and mixed earwax; the degree of blockage is quantified as mild blockage, moderate blockage, or severe blockage based on the proportion of earwax in the ear canal cross-sectional area.

[0019] Compared with the prior art, the present invention has the following beneficial effects: First, this invention achieves proactive perception and risk avoidance of the ear canal wall's physiological state. Through image processing and a control unit, it comprehensively assesses the type of earwax, the degree of blockage, and the condition of the ear canal wall before flushing, identifying whether the ear canal wall is congested or damaged, and using this as one of the bases for generating a flushing strategy. When congestion or damage is detected, the system reduces at least one of the water flow pressure and pulse frequency during the flushing strategy generation stage, avoiding secondary stimulation of the damaged tissue at the source, rather than passively responding during the flushing process. Compared to existing technologies that only select the flushing mode based on earwax type without considering the ear canal wall's own condition, this invention makes substantial improvements in both perception and decision-making logic, enhancing the safety and intelligence of the flushing process.

[0020] Secondly, the safety monitoring unit provides independent and necessary hardware-level safety assurance during the rinsing process. In a preferred embodiment of the invention, the image acquisition device and the rinsing handle are separate mechanisms. Since real-time monitoring of the ear canal status is impossible during rinsing via image recognition, the safety monitoring unit serves as the sole safety assurance mechanism during the rinsing phase. The safety monitoring unit of this invention includes at least one sensor independent of the image processing and control unit. When the physical parameters of the rinsing fluid exceed a preset safety threshold, the water flow output of the pulse water flow execution unit is directly shut off via an independent hardware circuit. This hardware-level independent safety protection channel does not rely on the software processing of the image processing and control unit. Even if the image processing and control unit malfunctions due to faults, misjudgments, or crashes, the safety monitoring unit can still independently complete parameter detection and forced shutdown. Compared to the single-layer protection architecture in the prior art, which aggregates all sensor signals to the main controller and has the main controller software uniformly judge and execute protective actions, this invention provides an indispensable bottom-line safety guarantee for invasive ear canal rinsing operations from the system architecture level. In a variant embodiment of the present invention, the image acquisition device and the rinsing handle can also be integrated into a single structure. During the rinsing process, ear canal images can be acquired simultaneously to assist in monitoring, but the independent hardware shutdown function of the safety monitoring unit is still retained as the last line of defense for safety.

[0021] Third, the parameters of the rinsing strategy are more precisely matched. This invention determines the pulse frequency differently according to the type of earwax. The pulse frequency range for dry earwax is 2.0 to 3.0 Hz, and the pulse frequency range for wet earwax is 0.8 to 1.5 Hz. The two do not intersect on the frequency axis, avoiding the ambiguous area of ​​frequency selection. At the same time, corresponding water flow pressure levels are set for different degrees of blockage and ear canal congestion or damage, so that the rinsing parameters form a more accurate correspondence with the actual working conditions of the ear canal, taking into account both cleaning efficiency and user comfort.

[0022] Fourth, the flushing strategy is iteratively optimized through post-flushing effect evaluation. After each flushing cycle, the invention re-acquires ear canal images using an image acquisition device to evaluate the earwax removal rate and changes in the ear canal wall condition, and adjusts the flushing strategy parameters based on the evaluation results. If the flushing effect does not meet the preset standard, the system returns to perform the next flush, gradually removing earwax through multiple iterations. This mechanism allows the system to dynamically optimize strategy parameters based on the actual flushing effect, balancing the safety of a single flush with the cumulative cleaning efficiency of multiple flushes. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall system structure of an embodiment of the present invention, showing the spatial arrangement of the image acquisition device, the rinsing handle, the control panel, and the constant temperature water tank; Figure 2This is a schematic diagram of the overall system structure from another perspective in an embodiment of the present invention, showing the pipeline connections and internal components such as the peristaltic pump and motor; Figure 3 This is an exploded view of the overall system structure according to an embodiment of the present invention, showing the disassembled positional relationship of each major component; Figure 4 This is an exploded structural diagram of the flushing handle according to an embodiment of the present invention, showing the internal structure of the nozzle, pipeline interface, and handle housing; Figure 5 This is an exploded view of the constant temperature water tank according to an embodiment of the present invention, showing the positional relationship between the heating / cooling components and the temperature sensor; Figure 6 This is a system control architecture block diagram of an embodiment of the present invention, showing the signal connection relationships between the image acquisition device, the image processing and control unit, the pulse water flow execution unit, the safety monitoring unit, and the human-machine interaction module; Figure 7 This is a flowchart of the adaptive flushing control method according to an embodiment of the present invention, showing the execution sequence of steps S1 to S6 and the feedback loop of step S5; Figure 8 This is a circuit block diagram of the safety monitoring unit according to an embodiment of the present invention, showing the signal link of the sensor being turned off by the power supply circuit controlled by the comparison circuit and the relay.

[0024] The corresponding reference numerals in the attached figures are as follows: 1—Image acquisition device; 2—Camera; 3—Supplemental lighting source; 4—Image processing and control unit; 5—Pulse water flow execution unit; 6—Peristaltic pump; 7—Motor; 8—Safety monitoring unit; 9—Sensor; 10—Comparison circuit; 11—Power supply circuit; 12—Constant temperature water tank; 13—Heating / cooling component; 14—Temperature sensor; 15—Pressure sensor; 16—Pipeline; 17—Sprayer head; 18—Washing handle; 19—Human-machine interface module; 20—Control panel; 21—High-speed data interface; 22—Relay; 23—Normally closed contact. Detailed Implementation

[0025] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. The embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0026] Reference Figure 1 and Figure 2This embodiment provides an adaptive ear canal cleaning system based on visual recognition, including an image acquisition device 1, an image processing and control unit 4, a pulse water flow execution unit 5, and a safety monitoring unit 8.

[0027] The image acquisition device 1 includes a camera 2 and a supplementary light source 3. The supplementary light source 3 provides illumination inside the ear canal, enabling the camera 2 to acquire real-time images. In this embodiment, the image acquisition device 1 and the rinsing handle 18 are two separate mechanisms. The image acquisition device 1 is inserted into the ear canal alone to acquire images inside the ear canal during use, while the rinsing handle 18 is inserted into the ear canal during rinsing to perform pulsed water flow rinsing. The two are used alternately in different operating phases. The camera 2 is a miniature CMOS image sensor with a diameter not exceeding 3mm, and its imaging angle is directed towards the depth of the ear canal. The supplementary light source 3 is an LED cold light source, arranged in a ring around the camera 2 to provide uniform, shadow-free illumination. LED cold light sources have high luminous efficiency and low heat generation, making them suitable for long-term operation in the enclosed and confined space of the ear canal. In another specific embodiment of the invention, the supplementary light source 3 uses an optical fiber transmission light source. The light source generator is located at the rear end of the image acquisition device 1, and the light is transmitted to the front end through an optical fiber bundle. This method helps to further reduce the heat generation at the front end and is suitable for applications sensitive to temperature rise. In another specific embodiment of the present invention, the supplementary lighting source 3 uses a laser diode light source, combined with a diffuse optical element to achieve uniform illumination, suitable for scenarios requiring higher brightness to obtain clearer images. The camera 2 is connected to the image processing and control unit 4 via a high-speed data interface 21, transmitting the acquired real-time image data to the image processing and control unit 4 for subsequent analysis and processing. In this embodiment, the high-speed data interface 21 uses a MIPI interface. In another specific embodiment of the present invention, the high-speed data interface 21 uses a USB interface. In yet another specific embodiment of the present invention, the camera 2 and the image processing and control unit 4 transmit image data wirelessly, specifically using Wi-Fi or Bluetooth protocols. No physical cable connection is required between the camera 2 and the image processing and control unit 4, suitable for scenarios requiring increased operational flexibility.

[0028] The image processing and control unit 4 is signal-connected to the image acquisition device 1. It is configured to identify the type of earwax, the degree of blockage, and the state of the ear canal wall based on real-time images, generate a flushing strategy based on the identification results, and, after flushing, re-acquire real-time images of the ear canal through the image acquisition device 1 to evaluate the flushing effect and adjust the flushing strategy based on the evaluation results. In this embodiment, the image processing and control unit 4 is implemented using a main controller integrated on the back of the control panel 20. This main controller integrates an image processing chip and a frequency converter chip. The image processing chip is a TMS320C6713 DSP chip, and the frequency converter chip is an IR2104 high-voltage half-bridge frequency converter chip.

[0029] Specifically, when the image processing and control unit 4 analyzes real-time images, it classifies and identifies earwax based on the texture, color, and morphological features of the images, classifying earwax into dry earwax, wet earwax, or mixed earwax. Dry earwax appears as a light yellowish-white color, with a dry, dull surface and a loose, flaky texture in the image; wet earwax appears as a deep yellow color, with a sticky, glossy surface and a clump-like texture; mixed earwax exhibits characteristics of both types, and the classification is based on the proportion of dry and wet components.

[0030] In this embodiment, the image processing and control unit 4 performs the following steps in the analysis and processing of real-time images: First, the original image is preprocessed with noise reduction filtering and contrast enhancement to improve the signal-to-noise ratio of the image acquired in the dark environment of the ear canal; then, the region of interest is extracted from the preprocessed image, which is separated from the image background by a preset ear canal wall boundary segmentation algorithm; within the region of interest, color histogram features, gray-level co-occurrence matrix texture features, and morphological contour features are extracted respectively, and the above three types of features are combined into a feature vector; finally, the feature vector is input into a pre-trained classification model, and the recognition results of earwax type, degree of blockage, and ear canal wall state are output. As a specific implementation of this embodiment, the extracted color histogram features (256 dimensions), gray-level co-occurrence matrix texture features (including four dimensions: contrast, correlation, energy, and homogeneity), and morphological contour features (including three dimensions: area, perimeter, and roundness) can be combined into a 263-dimensional feature vector. This vector is then input into a pre-trained SVM classifier, which outputs classification results for earwax type (dry, wet, or mixed), degree of blockage (mild, moderate, or severe), and ear canal wall condition (normal, congested, or damaged). The specific implementation of the above feature extraction method and classification model belongs to conventional techniques in this field. Those skilled in the art can choose appropriate solutions based on actual hardware computing power and accuracy requirements. For example, the classification model can use a convolutional neural network or a support vector machine, and feature extraction can employ manually designed feature engineering or end-to-end deep learning feature extraction.

[0031] Meanwhile, the image processing and control unit 4 quantifies the degree of blockage by the proportion of earwax in the cross-sectional area of ​​the ear canal, classifying the blockage into mild, moderate, or severe blockage. In this embodiment, earwax accounting for less than 30% of the cross-sectional area of ​​the ear canal is considered mild blockage, 30% to 60% is considered moderate blockage, and greater than 60% is considered severe blockage. Furthermore, the image processing and control unit 4 assesses the condition of the ear canal wall by examining its color and texture features, identifying whether the ear canal wall is congested or damaged. Congestion is characterized by localized redness and coarsened texture on the ear canal wall, while damage is characterized by broken or abnormally shaped areas on the surface of the ear canal wall.

[0032] In this embodiment, the earwax removal rate is defined as (the percentage of earwax in the ear canal cross-sectional area before rinsing - the percentage of earwax in the ear canal cross-sectional area after rinsing) / the percentage of earwax in the ear canal cross-sectional area before rinsing × 100%. This quantifies the cleaning effect of a single rinse by comparing the relative reduction in blockage before and after rinsing. This definition is consistent with the blockage quantification method based on cross-sectional area percentage in this embodiment.

[0033] Based on the above recognition results, the image processing and control unit 4 generates a flushing strategy. This flushing strategy includes at least pulse frequency parameters and water flow pressure parameters. In this embodiment, the flushing strategy is automatically calculated and generated by the image processing and control unit 4 based on the recognition results. In another specific embodiment of the invention, the flushing strategy is obtained by matching a preset flushing strategy library. The preset flushing strategy library stores multiple sets of preset flushing strategy parameters corresponding to combinations of recognition results related to earwax type, degree of blockage, and ear canal wall condition. The image processing and control unit 4 retrieves the closest matching preset strategy from the strategy library based on the current recognition result. This method involves less computation and has a fast response time, making it suitable for scenarios with limited processor computing power.

[0034] In this embodiment, the pulse frequency parameter in the rinsing strategy is determined according to the type of earwax: dry earwax corresponds to a first frequency range, and wet earwax corresponds to a second frequency range. The lower limit of the first frequency range is greater than the upper limit of the second frequency range, so that the pulse frequency corresponding to dry earwax is higher than that corresponding to wet earwax. Specifically, the first frequency range is 2.0 to 3.0 Hz, and the second frequency range is 0.8 to 1.5 Hz; the two ranges do not intersect on the frequency axis. The principle behind this design is that dry earwax is harder and requires a higher pulse frequency to generate periodic impact force to break up hard earwax fragments; wet earwax is sticky, and a lower pulse frequency is used to gradually dissolve and remove the sticky earwax through continuous rinsing. The water pressure parameter in the flushing strategy is determined jointly based on the degree of blockage and the condition of the ear canal wall: the water pressure increases with the degree of blockage, but does not exceed a preset safe pressure upper limit; the water pressure corresponding to mild blockage is 8 to 15 kPa, the water pressure corresponding to moderate blockage is 15 to 25 kPa, and the water pressure corresponding to severe blockage is also 15 to 25 kPa; when the ear canal wall is identified as congested or damaged, the water pressure is limited to below the preset safe pressure value. In this embodiment, the preset safe pressure value is 3 to 8 kPa.

[0035] After rinsing, the operator removes the rinsing handle 18 and re-inserts the image acquisition device 1 into the ear canal to acquire images of the ear canal interior. The image processing and control unit 4 evaluates the rinsing effect based on the newly acquired images, including assessing the earwax removal rate and changes in the ear canal wall condition. In this embodiment, if the earwax removal rate does not meet a preset standard (≥85%), the cleaning is considered complete when the proportion of earwax in the ear canal cross-sectional area drops below 15% of the initial assessment value. In other embodiments, this preset standard can be set to other values ​​according to clinical needs, such as 80% or 90%. The image processing and control unit 4 adjusts the parameters of the rinsing strategy based on the evaluation results, such as adjusting at least one of the pulse frequency and water flow pressure for use in the next round of rinsing. When the ear canal wall condition is identified as congested or damaged, at least one of the water flow pressure and pulse frequency is reduced or rinsing is terminated. Cleaning is considered complete when the earwax removal rate reaches the preset standard (≥85%). Through the above-mentioned post-rinse evaluation and strategy iteration mechanism, the system can gradually optimize parameters in multiple rinsing cycles, taking into account both the safety of a single rinse and the cumulative cleaning effect.

[0036] The pulse water flow execution unit 5 is connected to the image processing and control unit 4 and is configured to output a pulse water flow with adjustable pulse frequency and water pressure according to the rinsing strategy. In this embodiment, the pulse water flow execution unit 5 includes a peristaltic pump 6 and a motor 7 that drives the peristaltic pump 6. The pump body of the peristaltic pump 6 does not contact the rinsing fluid; the rinsing fluid only flows within the constant temperature water tank 12, the pipeline 16, and the nozzle 17. The peristaltic pump 6 pushes the liquid flow by squeezing the external hose. In another specific embodiment of the present invention, the pulse water flow execution unit 5 uses a diaphragm pump. The diaphragm pump generates a pulse water flow through the periodic reciprocating motion of the diaphragm. The pulse frequency is determined by the frequency of the diaphragm's reciprocating motion, and the water pressure is determined by the diaphragm stroke and the driving air pressure. A pressure sensor 15 is installed on the output pipeline of the diaphragm pump, and its output signal is connected to the comparison circuit 10 of the safety monitoring unit 8. The normally closed contact 23 of a relay 22 is connected in series in the power supply circuit 11 of the diaphragm pump's drive circuit. Its shutdown control method is consistent with the peristaltic pump scheme, ensuring that the safety monitoring unit 8 can still achieve direct shutdown protection without image processing and control unit 4 under this alternative pump type. This method is suitable for scenarios requiring a higher pressure range. In another specific embodiment of the invention, the pulse water flow execution unit 5 uses a plunger pump, which generates pulse water flow by driving the reciprocating motion of the plunger through a stepper motor. The step angle and drive frequency of the stepper motor directly determine the pulse frequency and impulse accuracy. A pressure sensor interface is integrated into the output pipeline of the plunger pump for easy connection to the sensor 9 of the safety monitoring unit 8. The normally closed contact 23 of the safety monitoring unit 8 is also connected in series in the power supply circuit 11 of its stepper motor driver, allowing direct shutdown when physical parameters exceed limits. This scheme is superior to the peristaltic pump scheme in terms of pulse frequency control accuracy.

[0037] In this embodiment, motor 7 is a brushless motor, driven by square wave start and sensorless FOC magnetic field orientation control, achieving 0.1-degree micro-step positioning, making the frequency and impulse adjustment of the pulsed water flow more precise. In another specific embodiment of the invention, motor 7 is a stepper motor, using a microstepping drive method. By adjusting the frequency and microstepping level of the stepping pulses, the output pulse frequency and single impulse of the peristaltic pump 6 are controlled. This method has high pulse frequency control accuracy and is suitable for scenarios requiring a wide range of frequency adjustment.

[0038] The image processing and control unit 4 sends control signals to the motor 7 via a control bus. The motor 7 drives the peristaltic pump 6 to output pulsed water flow according to the pulse frequency and water pressure set by the rinsing strategy. In this embodiment, the control bus uses an SPI bus. The inlet of the peristaltic pump 6 is connected to the outlet of the constant temperature water tank 12 via a detachable pipe 16, and the outlet of the peristaltic pump 6 is connected to the nozzle 17 at the front end of the rinsing handle 18 via the detachable pipe 16. The nozzle 17 is made of a flexible material and has a quick-release connection structure with the pipe 16. The flexible material can deform upon contact with the ear canal wall, preventing injury to the ear canal wall.

[0039] The safety monitoring unit 8 includes at least one sensor 9 independent of the image processing and control unit 4, configured to detect the physical parameters of the flushing fluid, and directly shut off the water flow output of the pulse water flow execution unit 5 when the physical parameters exceed a preset safety threshold, wherein the physical parameters include at least one of water flow pressure and temperature. (Refer to...) Figure 8 In this embodiment, the safety monitoring unit 8 includes a temperature sensor 14 and a pressure sensor 15, independent of the image processing and control unit 4. Since the image acquisition device 1 and the rinsing handle 18 are separate mechanisms, the internal state of the ear canal cannot be monitored in real time through image recognition during rinsing. Therefore, the safety monitoring unit 8 acts as the sole safety guarantee mechanism during rinsing, operating independently of the image processing and control unit 4 to ensure the physical safety of the rinsing operation. In another specific embodiment of the invention, the safety monitoring unit 8 only includes the pressure sensor 15, suitable for scenarios where liquid heating is not involved and room temperature rinsing fluid is sufficient. In this case, the physical parameters only include water flow pressure. In yet another specific embodiment of the invention, the safety monitoring unit 8 only includes the temperature sensor 14, suitable for scenarios sensitive to changes in the rinsing fluid temperature.

[0040] Temperature sensor 14 is used to monitor the temperature of the flushing fluid. In this embodiment, temperature sensor 14 is an NTC thermistor. In another specific embodiment of the present invention, temperature sensor 14 is a PTC thermistor. In yet another specific embodiment of the present invention, temperature sensor 14 is an infrared temperature sensor, which obtains the temperature of the outer wall of the pipeline through non-contact measurement, avoiding direct contact with the flushing fluid, and is suitable for scenarios with high sterility requirements.

[0041] Pressure sensor 15 is used to monitor water flow pressure and is installed on the output pipe 16 of pulse water flow actuator 5. In this embodiment, pressure sensor 15 is a piezoelectric pressure sensor. In another specific embodiment of the present invention, pressure sensor 15 is a strain gauge pressure sensor. In yet another specific embodiment of the present invention, pressure sensor 15 is a MEMS pressure sensor, which is small in size and has a fast response speed, making it suitable for scenarios with limited installation space.

[0042] The sensors of the safety monitoring unit 8 are connected to the power supply circuit 11 of the pulse water flow execution unit 5 via an independent comparison circuit 10. The comparison circuit 10 receives the output signals from the temperature sensor 14 and the pressure sensor 15, and compares the detected values ​​with a preset safety threshold. When the detected physical parameter exceeds the preset safety threshold, the comparison circuit 10 directly drives the relay 22 to operate, causing the normally closed contact 23 to open, thereby cutting off the power supply to the power supply circuit 11 and stopping the motor 7 and the peristaltic pump 6. This process does not require processing by the image processing and control unit 4, and there is no signal connection between the comparison circuit 10 and the image processing and control unit 4, forming a hardware-level safety baseline independent of the main control system.

[0043] In this embodiment, the preset safety thresholds include a temperature threshold of 42°C and a water pressure threshold of 30 kPa. These thresholds are exemplary values ​​and can be adjusted according to specific application scenarios. Those skilled in the art can determine appropriate safety thresholds based on the tolerance limits of the ear canal tissue and the characteristics of the rinsing solution.

[0044] The system also includes a constant-temperature water tank 12. The constant-temperature water tank 12 houses a heating / cooling assembly 13 and a temperature sensor. The temperature sensor in the constant-temperature water tank 12 is used in the constant-temperature control loop, and is an independent sensing element from the temperature sensor 14 in the safety monitoring unit 8. The constant-temperature water tank 12 is configured to allow the addition of ear canal care medication, achieving uniform medication delivery and integrated rinsing under the control of the rinsing strategy. The heating / cooling assembly 13 maintains the rinsing fluid temperature at 36.5℃±2℃ using a PID temperature control algorithm.

[0045] The system also includes a human-machine interface module 19, which is connected to the image processing and control unit 4 and integrated into the control panel 20. The human-machine interface module 19 provides a selection interface for preset rinsing modes, including a regular cleaning mode, a deep cleaning mode, and a sensitive soothing mode, and displays real-time images acquired by the camera 2, current rinsing strategy parameters, and system status information.

[0046] The working process of this embodiment is as follows: First, the user inserts the image acquisition device 1 into the ear canal and activates the device. The camera 2 and the supplementary light source 3 begin operating, acquiring real-time images of the inside of the ear canal and transmitting them to the image processing and control unit 4. The image processing and control unit 4 analyzes the real-time images, identifying the type of earwax, the degree of blockage, and the condition of the ear canal wall. Based on the identification results, the image processing and control unit 4 automatically generates a flushing strategy, determining parameters such as pulse frequency and water pressure. Simultaneously, the constant temperature water tank 12 maintains the flushing fluid at a preset temperature through the heating / cooling component 13.

[0047] The user then removes the image acquisition device 1 and inserts the nozzle 17 at the front end of the rinsing handle 18 into the ear canal. The pulse water flow execution unit 5 controls the peristaltic pump 6 to operate according to the rinsing strategy, outputting a pulse water flow with corresponding pulse frequency and water pressure. The rinsing fluid is delivered to the nozzle 17 through the tubing 16 to rinse the ear canal.

[0048] During the rinsing process, the safety monitoring unit 8 continuously monitors the signals from the temperature sensor 14 and the pressure sensor 15 via an independent comparison circuit 10. If an abnormal temperature or excessive pressure is detected, the comparison circuit 10 directly cuts off the power supply circuit 11, stopping the peristaltic pump 6, without the need for intervention from the image processing and control unit 4. Since the image acquisition device 1 has been removed from the ear canal during rinsing, the safety monitoring unit 8 is the sole safety guarantee during the rinsing phase.

[0049] After a single round of rinsing, the user removes the rinsing handle 18 and inserts the image acquisition device 1 into the ear canal again to acquire an image of the inside of the ear canal. The image processing and control unit 4 evaluates the rinsing effect based on the newly acquired image, including calculating the earwax removal rate and assessing changes in the ear canal wall condition. If the earwax removal rate does not reach the preset standard and the ear canal wall condition is normal, the rinsing strategy parameters are adjusted according to the evaluation results, and the user removes the image acquisition device 1 again and inserts the rinsing handle 18 to perform the next round of rinsing. If the ear canal wall condition is identified as congested or damaged, the parameters are reduced or rinsing is terminated. When the earwax removal rate is detected to have reached the preset value, the cleaning is considered complete.

[0050] After rinsing is completed, the system automatically generates a rinsing report, which includes comparison images of the ear canal before and after rinsing, the parameters of the rinsing strategy used, the number of rinsing cycles, and the rinsing duration.

[0051] Reference Figure 7 This embodiment provides an adaptive ear canal cleaning method based on visual recognition, which includes the following steps: Step S1: Acquire a real-time image of the inside of the ear canal using the image acquisition device 1. The image acquisition device 1 includes a camera 2 and a supplementary light source 3, which provides illumination inside the ear canal. In this embodiment, the camera 2 is a miniature CMOS image sensor, and the real-time image data is transmitted to the image processing and control unit 4 via a MIPI high-speed data interface.

[0052] Step S2: Perform image analysis on the real-time images to identify the type of earwax, the degree of blockage, and the condition of the ear canal wall. Specifically, earwax is classified into dry earwax, wet earwax, or mixed earwax based on the texture, color, and morphological features of the images; the degree of blockage is quantified as mild, moderate, or severe blockage based on the proportion of earwax in the cross-sectional area of ​​the ear canal; and the presence of congestion or damage to the ear canal wall is determined based on the color and texture features of the ear canal wall.

[0053] Step S3: Automatically generate a flushing strategy based on the identification results. The flushing strategy includes at least pulse frequency parameters and water flow pressure parameters. Specifically, the pulse frequency parameter is determined according to the type of earwax: the pulse frequency corresponding to dry earwax is higher than that corresponding to wet earwax. Dry earwax corresponds to a first frequency range, and wet earwax corresponds to a second frequency range, with the lower limit of the first frequency range being greater than the upper limit of the second frequency range. The water flow pressure parameter is determined according to the degree of blockage and the condition of the ear canal wall. The water flow pressure increases with the degree of blockage but does not exceed a preset safe pressure upper limit. When the ear canal wall is identified as congested or damaged, the water flow pressure is limited to below the preset safe pressure value regardless of the degree of blockage.

[0054] Step S4: Control the pulse water flow execution unit according to the rinsing strategy to output a pulse water flow with adjustable pulse frequency and water pressure to rinse the ear canal. In this embodiment, the pulse water flow execution unit is a peristaltic pump 6, driven by a brushless motor 7, which outputs pulse water flow by squeezing an external hose.

[0055] Step S5: After rinsing, real-time images of the inside of the ear canal are reacquired using the image acquisition device 1 to evaluate the rinsing effect. Evaluating the rinsing effect includes assessing the earwax removal rate and changes in the ear canal wall condition. Based on the evaluation results, the rinsing strategy is adjusted, including adjusting at least one of the pulse frequency parameter and water flow pressure parameter according to the earwax removal rate, and reducing at least one of the water flow pressure and pulse frequency or terminating the rinsing process when the ear canal wall condition is identified as congested or damaged. If the rinsing effect does not meet the preset standard, the process returns to step S4 to perform the next round of rinsing using the adjusted rinsing strategy.

[0056] Step S6: During the rinsing process, independent of steps S2 to S4, the physical parameters of the rinsing fluid are monitored. When a physical parameter exceeds a preset safety threshold, the output of the pulsed water flow is forcibly interrupted via an independent hardware circuit. The physical parameters include at least one of water pressure and temperature. Specifically, the temperature and water pressure of the rinsing fluid are monitored by temperature sensor 14 and pressure sensor 15, independent of the steps described above. When either parameter exceeds the preset safety threshold, the power supply circuit 11 of the pulsed water flow execution unit is directly cut off via an independent comparison circuit 10, which is not connected to steps S2 to S4, causing the peristaltic pump 6 to stop. Since the image acquisition device 1 has been removed from the ear canal during the rinsing process, the safety monitoring in step S6 is the only safety guarantee during the rinsing stage.

[0057] Based on the above embodiments, those skilled in the art, after reading this specification, can reasonably foresee the following variations that do not depart from the concept of the present invention: (1) Based on the above embodiments, the image processing and control unit 4 can also be configured to estimate the inner diameter of the ear canal at the current depth based on the size ratio of a known reference object, and adjust the single impulse parameter according to the estimated inner diameter of the ear canal so that the single output volume of the flushing fluid matches the cross-sectional area of ​​the ear canal, thereby avoiding insufficient impulse for large-diameter ear canals or excessive impulse for small-diameter ear canals.

[0058] (2) Based on the above embodiments, the safety monitoring unit 8 may also include an isolated leakage protection device, which detects leakage current through a current transformer and realizes automatic power-off and self-locking through a relay control circuit. Users can manually reset to release the protection, providing electrical safety protection in heating scenarios.

[0059] (3) Based on the above embodiments, the temperature sensor 14 may include a mechanical dual-temperature control switch and an electronic NTC thermistor. The two work together to sample and realize dual temperature detection and graded response control at both the hardware and software levels, preventing temperature monitoring failure due to the failure of a single sensor. Furthermore, the safety monitoring unit 8 may also include a sensor self-test circuit, which injects a known test signal into the temperature sensor 14 and the pressure sensor 15 and detects their output response when the system starts up. When the deviation of the output of any sensor from the expected value exceeds the allowable range, an alarm is issued and the flushing mode is prohibited, so as to ensure that the safety monitoring unit is in normal working condition before flushing.

[0060] (4) Based on the above embodiments, the human-computer interaction module 19 can also communicate with external terminal devices via wired or wireless means, transmit real-time images and rinsing parameters to the external terminal for display, and receive user instructions through the external terminal. The external terminal device is specifically a smartphone or tablet computer.

[0061] (5) Based on the above embodiments, step S5 may also include setting the maximum number of cycles or the maximum total rinsing time as the rinsing termination condition. When the number of cycles or the cumulative rinsing time reaches the preset upper limit, the rinsing will automatically terminate and a prompt will be issued regardless of whether the earwax removal rate reaches the preset standard, so as to prevent over-rinsing.

[0062] (6) Based on the above embodiments, the image acquisition device 1 and the rinsing handle 18 can also be integrated into a single structure. In this case, the image acquisition device 1 is embedded in the front end face of the rinsing handle 18, and can simultaneously acquire ear canal images during the rinsing process to achieve real-time dynamic adjustment during the rinsing process. This single structure is suitable for professional medical scenarios with high requirements for rinsing accuracy.

[0063] The above-described specific embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention.

Claims

1. An adaptive ear canal cleaning system based on visual recognition, characterized in that, include: An image acquisition device, including a camera and a supplementary light source, is configured to acquire real-time images of the inside of the ear canal; An image processing and control unit, connected to the image acquisition device, is configured to: identify the type of earwax, the degree of blockage, and the state of the ear canal wall based on the real-time image; generate a flushing strategy based on the identification results; the flushing strategy includes at least pulse frequency parameters and water flow pressure parameters; and after the pulse water flow execution unit finishes flushing according to the flushing strategy, re-acquire a real-time image of the inside of the ear canal through the image acquisition device to evaluate the flushing effect; the evaluation of the flushing effect includes at least evaluating the earwax removal rate and adjusting the flushing strategy based on the evaluation results. A pulse water flow execution unit, connected to the image processing and control unit, is configured to output a pulse water flow with adjustable pulse frequency and water pressure according to the rinsing strategy; as well as The safety monitoring unit includes at least one sensor independent of the image processing and control unit, configured to detect physical parameters, and when the physical parameters exceed a preset safety threshold, to directly shut off the water flow output of the pulse water flow execution unit through hardware circuitry independent of the image processing and control unit.

2. The system according to claim 1, characterized in that, The pulsed water flow execution unit includes a rinsing handle. The image acquisition device and the rinsing handle are independent and separate mechanisms. The image acquisition device and the rinsing handle are used alternately in different operation stages. The image acquisition device is used to insert into the ear canal to acquire images before and after rinsing, and the rinsing handle is used to insert into the ear canal to perform pulsed water flow rinsing during rinsing.

3. The system according to claim 1, characterized in that, The evaluation of the flushing effect includes evaluating the earwax removal rate and changes in the ear canal wall condition; the adjustment of the flushing strategy based on the evaluation results includes adjusting at least one of the pulse frequency parameter and the water flow pressure parameter according to the earwax removal rate, and reducing at least one of the water flow pressure and the pulse frequency or terminating the flushing when the ear canal wall condition is identified as congested or damaged.

4. The system according to claim 1, characterized in that, The sensor of the safety monitoring unit is connected to the power supply circuit of the pulse water flow execution unit through an independent comparison circuit; the comparison circuit is configured to directly cut off the power supply of the power supply circuit when the physical parameter exceeds the preset safety threshold, and the cutting-off action does not need to be processed by the image processing and control unit; the physical parameter includes at least one of water flow pressure and temperature.

5. The system according to claim 4, characterized in that, The safety monitoring unit includes a temperature sensor and a pressure sensor, independent of the image processing and control unit. The temperature sensor is used to monitor the temperature of the flushing fluid, and the pressure sensor is used to monitor the water flow pressure.

6. The system according to claim 1, characterized in that, In the flushing strategy, the pulse frequency parameter is determined according to the earwax type, wherein dry earwax corresponds to a first frequency range, wet earwax corresponds to a second frequency range, and the lower limit of the first frequency range is greater than the upper limit of the second frequency range. The first frequency range is 2.0 to 3.0 Hz, and the second frequency range is 0.8 to 1.5 Hz. The water flow pressure parameter is determined according to the degree of blockage and the ear canal wall condition, wherein the water flow pressure increases with the degree of blockage, but does not exceed a preset safe pressure upper limit. When the ear canal wall condition is identified as congested or damaged, regardless of the degree of blockage, the water flow pressure is limited to below a preset safe pressure value, wherein the preset safe pressure value is less than the preset safe pressure upper limit. The types of earwax include dry earwax, wet earwax, and mixed earwax; the degree of blockage is quantified as mild blockage, moderate blockage, or severe blockage based on the proportion of earwax in the cross-sectional area of ​​the ear canal.

7. An adaptive ear canal cleaning method based on visual recognition, characterized in that, Includes the following steps: S1. Acquire real-time images of the inside of the ear canal using an image acquisition device, the image acquisition device including a camera and a supplementary light source; S2. Perform image analysis on the real-time images to identify the type of earwax, the degree of blockage, and the condition of the ear canal wall; S3. Automatically generate a flushing strategy based on the recognition results, wherein the flushing strategy includes at least pulse frequency parameters and water flow pressure parameters; S4. Control the pulse water flow execution unit according to the flushing strategy to output pulse water flow with adjustable pulse frequency and water pressure to flush the ear canal; S5. After rinsing, real-time images of the inside of the ear canal are re-acquired through the image acquisition device to evaluate the rinsing effect, and the rinsing strategy is adjusted based on the evaluation results. If the rinsing effect does not meet the preset standard, the process returns to step S4 and the next round of rinsing is performed with the adjusted rinsing strategy. as well as S6. During the rinsing process, independent of steps S2 to S4, physical parameters are monitored, and when the physical parameters exceed a preset safety threshold, the output of the pulsed water flow is forcibly interrupted by an independent hardware circuit.

8. The method according to claim 7, characterized in that, The evaluation of the flushing effect includes evaluating the earwax removal rate and changes in the ear canal wall condition; the step S5 of adjusting the flushing strategy based on the evaluation results includes adjusting at least one of the pulse frequency parameter and the water flow pressure parameter according to the earwax removal rate, and reducing at least one of the water flow pressure and the pulse frequency or terminating the flushing when the ear canal wall condition is identified as congested or damaged.

9. The method according to claim 7, characterized in that, The forced interruption of the pulsed water flow output in step S6 includes directly cutting off the power supply to the pulsed water flow execution unit through an independent comparison circuit that is not connected to the signals in steps S2 to S4.

10. The method according to claim 7, characterized in that, The automatic flushing strategy generated in step S3 includes: determining the pulse frequency parameter based on the earwax type, wherein the pulse frequency corresponding to dry earwax is higher than that corresponding to wet earwax, the pulse frequency corresponding to dry earwax is 2.0 to 3.0 Hz, and the pulse frequency corresponding to wet earwax is 0.8 to 1.5 Hz; determining the water flow pressure parameter based on the degree of blockage and the ear canal wall condition, wherein the water flow pressure increases with the degree of blockage but does not exceed a preset safe pressure upper limit; when the ear canal wall condition is identified as congested or damaged, the water flow pressure is limited to below a preset safe pressure value regardless of the degree of blockage, wherein the preset safe pressure value is less than the preset safe pressure upper limit; the earwax type includes dry earwax, wet earwax, and mixed earwax; the degree of blockage is quantified as mild blockage, moderate blockage, or severe blockage based on the proportion of earwax in the ear canal cross-sectional area.

Images