Fine control oral cavity cleaning system and cleaning method

Abstract

The invention belongs to the technical field of oral cavity cleaning, and particularly relates to a fine-control oral cavity cleaning system and method, and the method comprises the following steps: obtaining images of teeth and gingiva in an oral cavity; acquiring a predetermined cleaning position based on a first image processing algorithm, and analyzing the predetermined cleaning position to generate a cleaning path; a cleaning device is controlled to aim at the preset cleaning position, and preliminary cleaning is conducted; the actual cleaning position of the cleaning device is obtained through a second image processing algorithm and compared with the preset cleaning position, the deviation between the actual cleaning position and the preset cleaning position is calculated, and the cleaning path is adjusted based on the deviation; and according to the cleaning path, the movement track and speed of the cleaning device are controlled, and fine cleaning is completed. The oral cavity cleaning device can realize accurate positioning and oral cavity cleaning.

Description

Technical Field

[0001] The present invention belongs to the technical field of oral cleaning, and particularly relates to a fine-controlled oral cleaning system and a cleaning method. Background Art

[0002] Oral cleaning is a key oral hygiene process. Its main purpose is to remove food residues on teeth and gums in the oral cavity, dental plaque and tartar on teeth and gums, so as to prevent oral diseases such as dental caries, periodontal diseases and bad breath. This process can be carried out in various ways, including manual brushing, using dental floss, interdental brushes, and oral cleaning devices such as oral irrigators or mouthwashes.

[0003] Problems existing in the prior art: During the oral cleaning process, how to accurately locate and clean teeth while avoiding damage to gums or wounds. In the existing oral cleaning devices, during the cleaning process, due to inaccurate positioning or improper control of the cleaning force, the cleaning effect is often not ideal or the oral cavity is damaged. Summary of the Invention

[0004] The purpose of the present invention is to provide a fine-controlled oral cleaning system and a cleaning method, which can accurately locate and clean the oral cavity.

[0005] The technical solutions adopted by the present invention are specifically as follows: A fine-controlled oral cleaning method includes the following steps: Obtain images of teeth and gums in the oral cavity; Based on the first image processing algorithm, obtain a predetermined cleaning position, and generate a cleaning path by analyzing the predetermined cleaning position; Control the cleaning device to aim at the predetermined cleaning position for preliminary cleaning; Obtain the actual cleaning position of the cleaning device through the second image processing algorithm, compare it with the predetermined cleaning position, calculate the deviation between the actual cleaning position and the predetermined cleaning position, and adjust the cleaning path based on the deviation; According to the cleaning path, control the movement trajectory and speed of the cleaning device to complete fine cleaning.

[0006] The preliminary cleaning uses a small-force water flow to clean the predetermined cleaning position in the oral cavity, and optimally adjusts the water flow force according to the characteristics of the predetermined cleaning position for subsequent cleaning of the predetermined cleaning position.

[0007] The method for judging the predetermined cleaning position includes the following steps: Establish a coordinate system, analyze multi-view images of the predetermined cleaning position, and obtain the coordinates of the predetermined cleaning position through triangulation. The multi-view images can be obtained by a multi-camera or by moving a single camera. Set the motion trajectory to include the following steps: Based on the acquired images of teeth and gums in the oral cavity, the coordinate system established for the predetermined cleaning positions, and the data information of the predetermined cleaning positions, obtain the cleaning strategies and methods for teeth and gums in the oral cavity during the actual cleaning process; Set the cleaning motion trajectory according to the cleaning strategies and methods, where the cleaning strategies at least include the Manhattan strategy and the Euclidean strategy, The Manhattan strategy forms a right-angled broken line trajectory by moving step by step to connect adjacent cleaning points; The Euclidean strategy forms the shortest path trajectory by moving in a straight line to connect adjacent cleaning points.

[0008] The algorithm optimizes the motion control of the cleaning device, including the following settings: State: cleaning target state, target water spraying position, and actual water spraying position; Action: cleaning device state and motor motion control parameters; Reward mechanism: the deviation between the target water spraying position and the actual water spraying position, including sparse rewards based on cleaning efficiency and cleaning effect; In the reinforcement learning training stage, optimize its decision-making strategy by repeatedly attempting cleaning and learning to maximize the cumulative reward.

[0009] Apply the algorithm to achieve feedback on the motion control strategy of the cleaning device, and adjust the decision according to the real-time feedback to optimize the cleaning efficiency and cleaning completion degree of the cleaning device by feedback.

[0010] A fine control oral cleaning system includes a pre-positioning system, the pre-positioning system includes an image acquisition device, a cleaning device, and a control system for the operation of the cleaning device. The control system uses the above oral cleaning method, and the image acquisition device includes at least one camera.

[0011] It includes: a transmission module, a storage module, and a central control unit. All modules are connected through standard interfaces to ensure real-time data transmission and feedback optimization.

[0012] The pre-positioning system includes: An input module for inputting the images of teeth and gums in the oral cavity acquired by the image acquisition device; A transmission module for transmitting digital signals of the control terminal, the oral cleaning device, and the pre-positioning system; A storage module for storing the images of teeth and gums in the oral cavity, the program for operating the pre-positioning system, and the cleaning strategies and methods for teeth and gums in the oral cavity; A CPU, connected to and controlling the above input module, transmission module, storage module, and used to execute the instructions for running the program; A control terminal, which is used to control the cleaning device and adjust a predetermined cleaning position, and realizes the transmission and processing of data information through an interaction module.

[0013] A precise oral cleaning device, characterized by comprising at least one oral cleaning device for executing the precise oral cleaning method, and the oral cleaning device executes the operation program and program instructions of the precise oral cleaning system.

[0014] The technical effects achieved by the present invention are as follows: In the present invention, through an image processing algorithm, the recognition and positioning of teeth and gums inside the oral cavity can be realized, and through the set cleaning movement trajectory and algorithm of the cleaning device, it is used to achieve accurate cleaning of the cleaning position. And during cleaning, by predicting the optimal movement trajectory and speed, damage to teeth and gums can be effectively avoided.

[0015] In the present invention, when the cleaning system judges whether to stop or complete an instruction, when the cleaning system approaches the threshold of the pre-cleaning position, an automated learning of the cleaning system is realized by designing a reward function. In the environment of the oral cleaning system, the reward function is used to encourage the cleaning device to accurately and safely clean the teeth and gums in the oral cavity, avoid damaging the teeth and gums, and effectively remove food residues, dental plaque and dental calculus.

[0016] In the present invention, by repeating the trajectory movement and gradually reducing the deviation between the cleaning device and the target cleaning area, the cleaning accuracy and efficiency can be significantly improved, so as to achieve the thoroughness of the cleaning effect. Description of the Drawings

[0017] Figure 1 is the flowchart of the oral cleaning method of the present invention; Figure 2 is the flowchart of reinforcement learning in the present invention; Figure 3 is the flowchart of the reinforcement learning system in the present invention; Figure 4 is the flowchart of the method for judging a predetermined cleaning position in the present invention; Figure 5 is the schematic diagram of the predetermined cleaning position and the actual cleaning position in the analysis of the image processing algorithm of the present invention. Detailed Embodiments

[0018] In order to make the purpose and advantages of the present invention clearer, the present invention will be specifically described below in conjunction with embodiments. It should be understood that the following text only describes one or several specific implementation manners of the present invention, and does not strictly limit the specific scope of protection requested by the present invention.

[0019] It should be noted that in the description and claims of the present invention and the above-mentioned drawings, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0020] According to an embodiment of the present invention, a fine control oral cleaning system and a cleaning method are provided. It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system including at least one set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0021] The associated method embodiment can also be executed in an electronic device including a memory and a processor, a similar control device, or in the cloud. Taking the electronic device as an example, the electronic device can include one or more processors and a memory for storing data. Optionally, the above-mentioned electronic device can also include a communication device for communication functions and a display device. Those of ordinary skill in the art can understand and know that the above structural description is only illustrative and does not limit the structure of the above-mentioned electronic device. For example, the pre-positioning system can also include more or fewer components than the above structural description, or have a configuration different from the above structural description.

[0022] In this application, the used processor CPU can include one or more processing units. For example: the processor can include a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processing (DSP) chip, a microcontroller unit (MCU), a field-programmable gate array (FPGA), a neural-network processing unit (NPU), a tensor processing unit (TPU), a processing device such as an artificial intelligent (AI) type processor, etc. Among them, different processing units can be independent components or integrated in one or more processors. In some instances, the control end or the input module can also include one or more processors.

[0023] The storage module can be used to store computer programs, such as storing the computer program corresponding to the pre-positioning system in the embodiments of the present invention. The processor realizes the positioning result of the above-mentioned pre-positioning method by running the computer program stored in the memory. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory may further include a memory remotely provided relative to the processor, and these remote memories can be connected to the electronic device through a network. Examples of the above network include but are not limited to the Internet, intranet, local area network, mobile communication network, and combinations thereof.

[0024] The transmission module is used to receive or send data via a network. Specific examples of the above network may include the wireless network provided by the communication provider of the control end. In one instance, the communication mode of the transmission module includes wired and wireless, and the communication device includes a network interface controller (NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In this embodiment, for example, a remotely controlled robot, robotic arm, or micro cleaning device is used. In one instance, the communication device can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly. In some embodiments of this solution, the communication device is used to connect to the control end device, and instructions can be sent to the electronic device through a mobile device. In one instance, the wired transmission mode can be a direct connection of wires, which is used for an oral cleaning device integrally produced or an oral cleaning device controlled by wire transmission. In some embodiments of this solution, the wired transmission mode is used to connect to the control end device, and instructions can be sent to the electronic device through a mobile device.

[0025] The control end includes a manipulation button or a display device with a manipulation function. The display device can be a touch screen type liquid crystal display (LCD) and a touch display (also referred to as a "touch screen" or "touch display screen"). The display enables the user to interact with the user interface of the electronic device. In some embodiments, the above-mentioned electronic device has a graphical user interface (GUI), and the user can perform human-computer interaction with the GUI by touching and / or gesturing with a finger on the touch-sensitive surface. The executable instructions for performing the above human-computer interaction function are configured / stored in a computer program product or readable storage medium executable by one or more processors.

[0026] In addition to being applicable to this embodiment, this application is also applicable to various oral cleaning devices, including but not limited to smart electric toothbrushes, smart oral irrigators, oral cleaning robots, etc.

[0027] As Figure 1 shown, a method for fine-controlled oral cleaning includes the following steps: S1. Obtain images of teeth and gums in the oral cavity; S2. Based on the first image processing algorithm, obtain the predetermined cleaning positions, and analyze the predetermined cleaning positions to generate a cleaning path; S3. Control the cleaning device to aim at the predetermined cleaning positions for preliminary cleaning; S4. Through the second image processing algorithm, obtain the actual cleaning positions of the cleaning device, compare them with the predetermined cleaning positions, calculate the deviation between the actual cleaning positions and the predetermined cleaning positions, and adjust the cleaning path based on the deviation; S5. Control the movement trajectory and speed of the cleaning device according to the cleaning path to complete fine cleaning.

[0028] In step S1, for obtaining images of teeth and gums in the oral cavity, the specific obtaining methods include but are not limited to image photographing devices, ultrasonic photographing devices, infrared photographing devices, 3D scanners, etc., and combining with a ring-shaped LED light source to eliminate shadows, ensuring that the images cover the entire tooth and gum areas of the whole mouth. In this embodiment, the image photographing method is used to obtain images of teeth and gums in the oral cavity. In addition to gums and teeth, the photographing area can also be applied to parts inside the oral cavity such as the tongue coating.

[0029] Among them, the above images not only include the surfaces of teeth but also the details of gums, providing comprehensive visual information for subsequent generation of cleaning paths and improving the accuracy of image processing.

[0030] In step S2, the first image processing algorithm includes but is not limited to using algorithms such as image enhancement algorithms, texture analysis, image segmentation, image feature extraction, image matching algorithms, and color analysis algorithms. Using the above algorithms, for the images of teeth and gums obtained at the predetermined cleaning positions, use the image enhancement algorithm to process the clarity of the images, and respectively use texture analysis, image segmentation, image feature extraction, image matching algorithms, and color analysis algorithms to identify and distinguish the teeth, gums, and residues to be cleaned and extract elements, so as to provide predetermined cleaning points for generating the cleaning path, and through the sequential entry of the points at the predetermined cleaning positions, according to the positional relationship between the points, to guide the cleaning device to generate a cleaning path.

[0031] The above algorithms analyze the colors, textures, and shapes in the images compared with healthy tooth images, identify the parts to be cleaned such as dental plaque and dental calculus, and plan the optimal cleaning path.

[0032] Please refer to Figure 5, in the figure, the five-pointed star represents the predetermined cleaning position analyzed by the first image processing algorithm, and the circle represents the actual cleaning position analyzed by the second image processing algorithm.

[0033] As an embodiment: The first image processing algorithm needs to accurately identify dirty areas such as dental plaque and tartar on the tooth surface and provide target coordinates for the cleaning path.

[0034] The implementation steps are as follows: S301. Use a target detection algorithm such as the combination of an encoding-decoding structure like U-Net to achieve high-precision target detection and pixel-level segmentation. By collecting and training an oral image library, the labeled categories include dental plaque (red label), tartar (yellow label), and healthy tooth surface (green label). Randomly rotate, adjust the contrast, and add Gaussian noise to simulate the illumination changes in the oral cavity; S302. Use an image enhancement algorithm to enhance the local contrast of the image; S303. Extract the initial boundary between the gums and teeth through HSV color space segmentation; S304. Use an algorithm such as YOLOv5 to detect the bounding boxes of tartar and dental plaque and output candidate regions; S305. Perform semantic segmentation on the candidate regions through U-Net to generate a pixel-level mask for the dirty areas; S306. For the segmented dirty areas, extract the centroid coordinates as the predetermined cleaning points.

[0035] In steps S3 and S4, in the preliminary cleaning stage, the cleaning device will perform preliminary cleaning on the predetermined cleaning positions on the teeth and gums according to the generated cleaning path. The preliminary cleaning is mainly to remove large pieces of dental plaque, tartar, or food residues, etc., to prepare for the subsequent fine cleaning.

[0036] Alternatively, in the preliminary cleaning stage, the cleaning device performs preliminary cleaning on the predetermined cleaning positions on the teeth and gums according to the generated cleaning path. Subsequently, use the image processing algorithm in step S1 to obtain the images of the teeth and gums in the oral cavity and step S2 to identify whether there are still dental plaque, tartar, or food residues, etc. on the surfaces of the teeth and gums after preliminary cleaning. If so, directly start the fine cleaning method according to steps S4 and S5.

[0037] In step S4, the deviation between the actual cleaning position and the predetermined cleaning position, that is, the distance between the cleaning device and the predetermined cleaning position. When the actual cleaning position moves to the predetermined cleaning position, at this time, the deviation is the thickness of the remaining dental plaque, tartar, or food residues, etc. at the predetermined cleaning position. The cleaning method and path use different cleaning methods according to different cleaning types, including: For example, when the predetermined cleaning position is dental calculus, the ultrasonic cleaning device is used to directly impact the dental calculus part, or the ultrasonic cleaning device is used to vibrate and clean on the surface of the dental calculus in a fitting manner; For example, when the predetermined cleaning position is food residue, the cleaning device is used to rinse the predetermined cleaning position. The rinsing liquid includes but is not limited to water, mouthwash, etc. By increasing the water flow, the rinsing of the predetermined cleaning position is achieved. During the cleaning process, the food residue is advanced and cleaned from one side along the tooth gaps or one side of the tooth, and during the cleaning process, rotational cleaning for the position of the food residue, repeated spot rinsing cleaning, etc. are also used; For example, when the predetermined cleaning position is dental plaque, an ultrasonic cleaning device and a cleaning equipment are used in cooperation to achieve the cleaning purpose.

[0038] The above cleaning method is used to match different cleaning methods based on the magnitude of the deviation value. Based on this deviation value, the cleaning device will set the cleaning movement trajectory to ensure that the cleaning device can accurately clean each position that needs to be cleaned. The deviation calculation method uses the Euclidean strategy, that is, the following distance formula is used to achieve the quantitative calculation of the deviation: ; When the deviation is greater than a constant, local path replanning is triggered, such as inserting correction path points, and the constant is adjusted according to different cleaning types and tooth and gum types.

[0039] In addition, when in a sensitive area, switch to the Manhattan strategy to reduce the impact risk.

[0040] As an embodiment: The actual water spraying position of the cleaning device is tracked in real time through the second image processing algorithm and compared with the predetermined position. The implementation steps are as follows: S401. Obtain and track the dynamic trajectory of the water spraying flow; S402. Combine the HSV color space to extract the water flow area; S403. Align the real-time water spraying area with the predetermined coordinates through template matching; S404. Obtain the deviation value using the deviation calculation formula; S405. When the deviation value exceeds the predetermined value, trigger the controller to adjust the nozzle angle and correct the trajectory.

[0041] In step S5, in order to further optimize the cleaning effect, the cleaning device will also optimize the movement trajectory of the cleaning device based on the optimization algorithm, speculate and optimize the movement trajectory and speed to reduce the deviation value between the actual cleaning position and the predetermined cleaning position. When the deviation value approaches the minimum threshold, the cleaning process ends, thereby achieving fine cleaning.

[0042] Among them, the optimization algorithms include, but are not limited to, reinforcement learning algorithms, deep learning algorithms, policy gradient learning algorithms, etc.

[0043] According to the above steps, first, through the image processing algorithm, the recognition and positioning of teeth and gums inside the oral cavity can be achieved, and the cleaning movement trajectory and algorithm of the cleaning device are set to accurately clean the cleaning position. During cleaning, by predicting the optimal movement trajectory and speed, damage to teeth and gums can be effectively avoided.

[0044] Furthermore, in the above steps, before processing through the image processing algorithm, it also includes using the cleaning device to perform preliminary cleaning of the target in the oral cavity with a small - force water flow to remove some impurities, avoiding affecting the processing results during image recognition and the use of image algorithms, improving the accuracy of positioning the predetermined cleaning position, enhancing the efficient targeting of cleaning, and using a large - force water flow to clean the predetermined cleaning position in the oral cavity during the process of optimizing the movement control of the cleaning device using the algorithm. During the process of the cleaning device cleaning along with the predetermined cleaning position, the predetermined cleaning position in the oral cavity is cleaned by using a large - force water flow. By repeating the trajectory movement and gradually reducing the deviation between the cleaning device and the target cleaning area, the cleaning accuracy and efficiency can be significantly improved, thereby achieving the thoroughness of the cleaning effect.

[0045] The above - mentioned large - force water flow and small - force water flow are in a relative state. The specific water flow size is affected by factors such as water pressure, flow rate, flow velocity, and water flow cross - section. When targeting different people with oral diseases, or oral damage and infections, etc., it is necessary to judge according to the actual water flow force that can be tolerated, or the water flow force can be gradually increased until the water flow force that can achieve the removal of dirt at the cleaning position is reached.

[0046] Please refer to Figure 2 and Figure 3 , in order to achieve the automatic recognition and cleaning of teeth and gums in the oral cavity, an algorithm is used to realize the recognition and cleaning of teeth and gums in the oral cavity. The optimization algorithm is used to optimize the movement control of the cleaning device, specifically including: S101, State representation: the cleaning target state, the target water - spraying position, and the actual water - spraying position; S102, Action selection: According to the current state of the cleaning device, set the state of the cleaning device and the motor movement control parameters. The control parameters include the cleaning action trajectory and cleaning speed of the cleaning device; S103, Reward mechanism: The deviation between the target water - spraying position and the actual water - spraying position, including sparse rewards based on cleaning efficiency and cleaning effect. Design a reward function to evaluate the cleaning effect of each action. The reward function is based on the cleaning effect and the degree of protection of the gums; S104, Reinforcement learning: By repeatedly attempting cleaning and learning, optimize its decision-making strategy to maximize cumulative rewards; S105, Real-time adjustment: During each cleaning process, the algorithm adjusts its decisions based on real-time feedback to optimize the cleaning effect.

[0047] According to the above, the cleaning device can adapt to different oral environments and user needs, significantly improving the cleaning accuracy and efficiency. Through continuous learning and adjustment, the algorithm can predict the optimal movement trajectory and speed to minimize the error of the cleaning position until it approaches the cleaning threshold of the set cleaning position. When the cleaning device uses the algorithm to clean dirt layer by layer, use the cleaning device until it approaches the threshold of the pre-positioned cleaning position. The threshold is the boundary value close to the tooth-gum body, that is, the boundary data of the model is the boundary of the tooth and gum. By setting the cleaning threshold, it is possible to effectively avoid damage to the teeth and gums during the cleaning process by the cleaning device, that is, the minimum threshold of the deviation value in steps S5 and S103 above. This threshold can be negative.

[0048] If, when the threshold is positive, the cleaning position is easy to clean, use the preliminary cleaning or rinsing method to clean any position on the tooth surface, gum surface or in the oral cavity, and there is no residue after cleaning.

[0049] If, when the threshold is zero, no black spots or necrotic lesions are found on the tooth or gum surface after cleaning, use a friction cleaning device including an ultrasonic cleaning device to friction clean the tooth surface.

[0050] If, when the threshold is negative, lesions or black spots appear on the tooth or gum, use a grinding or cutting device to clean the necrotic lesion position.

[0051] A method for determining whether it is close to the pre-cleaning position threshold; If so, feedback to stop running; If not, then feedback when approaching the threshold, select whether to continue to execute the running instruction. The judgment method includes manual judgment, and then the cleaning system learns from the manual judgment result to improve the automatic judgment of the cleaning system; Until a stop or completion instruction is received.

[0052] Furthermore, the stop instruction includes being implemented by a button connected to the cleaning device or being realized by the algorithm obtaining the judgment after cleaning is completed.

[0053] Further, when the cleaning system determines whether to stop or complete an instruction, when the cleaning system approaches the pre-cleaning position threshold, an automated learning of the cleaning system is achieved by designing a reward function. In the environment of the oral cleaning system, the reward function is to encourage the cleaning device to accurately and safely clean the teeth and gums in the oral cavity, avoid damaging the teeth and gums, and effectively remove food residues, dental plaque, and dental calculus.

[0054] Furthermore, an algorithm is applied to achieve feedback on the motion control strategy of the cleaning device, and the decision is adjusted according to the real-time feedback to optimize the cleaning efficiency and cleaning completion degree of the cleaning device through feedback.

[0055] Combined with Figure 2 and Figure 3 , the specific steps are as follows: Ss1. Determine the goals of the reward function, including cleaning efficiency, safety, and cleaning effect; Ss2. Determine the reward and punishment mechanism. When the cleaning device correctly cleans the target area, a positive reward is given; if the cleaning action may harm the gums, a negative punishment is given; when the cleaning task is completed, an additional reward is provided; the reward is adjusted according to the speed of completing the task to encourage fast but safe cleaning; Ss3. Set environmental constraints, such as the movement range of the cleaning device, as well as the movement amplitude and standard movement actions during the cleaning process of the cleaning device; Ss4. Determine the reward function, convert the goals in the above steps into quantitative indicators, such as cleaning area, usage time, and approaching the pre-cleaning position threshold; and unify the reward and punishment criteria in the above steps, which is conducive to the optimization and adjustment of the cleaning algorithm of the cleaning device; Ss5. Simulate testing and actual training adjustment. Test the reward function in the simulation environment and observe whether the behavior of the intelligent agent meets the expectations; adjust the reward function in the real world to obtain the best performance; Ss6. Update and optimize. Collect data during the cleaning process, such as image comparison before and after cleaning, cleaning path, user feedback, cleaning effect comparison, deviation from the pre-cleaning positioning position, etc. Analyze and improve the reward function based on the collected data to achieve the improvement of the pre-positioning method of the cleaning device.

[0056] Please refer to Figure 4 , the predetermined cleaning position judgment method includes the following steps: S201. Establish a coordinate system; S202. Analyze multi-perspective images of the predetermined cleaning position, such as images of teeth and gums in the oral cavity, and the cleaning position of the current cleaning device. The methods for obtaining multi-perspective images include but are not limited to multi-eye cameras or moving single-eye cameras; S203. Obtain the coordinates of the predetermined cleaning position through triangulation, and generate data information for the predetermined cleaning position.

[0057] According to the above steps, the coordinate system establishment method includes: For example, when the coordinate system measurement method is the triangulation method of monocular multi-view, a high-definition imaging device is used. One camera obtains images of multiple views for triangulation. By capturing high-definition images of the same object at different positions, the geometric relationship between the obtained images is used to determine the position and size of the object. Further, the calibration method of camera arrangement at different angles and the depth of camera photography, as well as the acquisition method and processing flow of camera collection, are used to obtain the image of the cleaning position. The specific processing method is to use image processing algorithms to achieve the positioning of the cleaning position and the differentiation of different elements in the cleaning position. The elements include teeth, gums, tongue coating, dirt, etc. Further, an optimization algorithm is used to achieve the precise positioning of the specific cleaning position, and the speculated movement trajectory and speed are used to clean the cleaning position for further fine cleaning of the cleaning position.

[0058] For example, when the coordinate system measurement method is the 2D coordinate initial positioning and precise adjustment method, through the positioning of a single or multiple cleaning positions, the fine adjustment of the cleaning device is used to achieve the fine cleaning of the cleaning position. Specifically, a 2D coordinate system established between the predetermined cleaning positions on a single plane or multiple single planes and the cleaning device, and between multiple predetermined cleaning positions and the cleaning device respectively is used. The cleaning device is used to perform targeted cleaning on the predetermined cleaning positions to ensure the targeted cleaning of each predetermined cleaning position.

[0059] For example, when the coordinate system measurement method is the three-dimensional coordinate modeling method, the cleaning position is scanned and a three-dimensional model or a three-dimensional point model is established. The predetermined cleaning positions in the three-dimensional model are input into the cleaning system of the cleaning device. By inputting the predetermined cleaning positions in the three-dimensional model or the three-dimensional point model into the cleaning device, and by establishing a cleaning movement trajectory, each cleaning position is cleaned one by one.

[0060] Through the above steps, a unified multi-view spatial reference framework for teeth and gums in the oral cavity is established to ensure the operability of subsequent coordinate calculation and path planning of the cleaning position.

[0061] Furthermore, three-dimensional information of the predetermined cleaning position is obtained through multi-view images to solve the problems of occlusion and depth blur of single-view images.

[0062] In step S203, the coordinates of the predetermined cleaning position are obtained through triangulation, and the three-dimensional coordinates of the predetermined cleaning position are accurately calculated by using the geometric relationship of multi-view images.

[0063] Further, assume that the optical centers of two cameras are O1 and O2 respectively, and the projection points of the same stain point in the two images are p1 and p2. Then, the three-dimensional coordinates P(x, y, z) of this point can be calculated by solving the following system of equations: ; where B is the known baseline distance between the two cameras, is the camera focal length, and are the image coordinates of the two cameras respectively, is the parallax, based on the horizontal direction difference, is the principal point coordinate, that is, the intersection point of the camera optical axis and the image plane, the image center.

[0064] Among them, in triangulation, the difference in the vertical direction coordinates and is eliminated by camera calibration, and only the horizontal parallax is used for depth calculation.

[0065] It should be noted that if more than two perspectives are used, the coordinate calculation results can be optimized by the least squares method to reduce the influence of noise, and the RANSAC (Random Sample Consensus) algorithm is used to exclude mis-matched feature points.

[0066] It should be further noted that if there is a vertical offset in the camera installation, needs to be used to calculate the vertical parallax, but in practical applications, this difference is usually eliminated by position calibration. The value and the value are default equal.

[0067] Through the above steps, high-precision three-dimensional positioning of the cleaning position in the oral cavity can be achieved, providing reliable data support for subsequent path planning.

[0068] As an alternative embodiment, according to step S202 and step S203, the monocular camera ranging scheme adopts an orbital movement method to obtain the three-dimensional coordinates of the predetermined cleaning position through triangulation. This scheme uses multi-perspective images taken by the camera at different positions and combines the target detection algorithm to extract the target point coordinates to achieve high-precision depth calculation.

[0069] Assume that the camera moves along the orbit, and its optical center is at O1 and O2 at two different positions respectively. The projection points corresponding to a certain stain point P in the images taken at these two positions are p1 and p2 respectively. Then, the three-dimensional coordinates of this point can be calculated according to the camera imaging geometric model.

[0070] Further, camera pose recording: Install an encoder on the track to record the displacement ΔB of the camera in real time, ensuring the acquisition of accurate positions from two different perspectives.

[0071] Furthermore, object detection and feature point extraction: Use the first image processing algorithm (such as YOLO, Faster R-CNN, Mask R-CNN) to detect the stain area and extract the central coordinates p1 and p2 of the stain.

[0072] Even further, parallax calculation: Calculate the horizontal parallax of the target point in two frames of images, and its value is obtained from the following relationship: d = ∣u1 - u2∣; where u1 and u2 are the horizontal coordinates of the center of the stain point obtained by the first image processing algorithm in two frames of images respectively.

[0073] Even further, depth calculation: The depth Z of the target point P is calculated by the following formula: ; is the focal length of the camera, B is the baseline displacement of the camera between two shootings, and d is the calculated parallax. Combining the internal and external parameter information of the camera, calculate the complete three-dimensional coordinates of the target point P(x, y, z) and perform spatial correction.

[0074] By calculating with images taken at multiple consecutive positions, the least squares method can be used to optimize the coordinate solution to improve the measurement accuracy. Use the RANSAC algorithm to remove object detection errors and improve the accuracy of feature point matching.

[0075] As an alternative embodiment, setting the motion trajectory includes the following steps: Based on the images of teeth and gums in the oral cavity, the coordinate system established for the predetermined cleaning position, and the data information of the predetermined cleaning position, obtain the cleaning strategy and method for teeth and gums in the oral cavity during the actual cleaning process, and set the cleaning motion trajectory according to the cleaning strategy and method, where the cleaning strategy at least includes the Manhattan strategy and the Euclidean strategy. The Manhattan strategy forms a right-angled broken line trajectory by moving step by step to connect adjacent cleaning points. The Euclidean strategy forms the shortest path trajectory by moving in a straight line to connect adjacent cleaning points.

[0076] According to the above, the recognition and input of teeth and gums in a clean oral cavity are realized through an image method. After collecting the image data of teeth and gums, a data model with teeth and gums as the boundary is established. The data model includes a coordinate system and data information of a predetermined cleaning position. According to the data information of the predetermined cleaning position, the deviation between the data model and the data information of the predetermined cleaning position is obtained to judge the specific position that needs to be cleaned. The position information is input into the cleaning device to generate a cleaning strategy and method for further setting and optimizing the cleaning movement trajectory of the cleaning device.

[0077] The above deviation also includes food residues, dental calculus, etc. protruding from the tooth and gum surfaces.

[0078] In order to further improve the cleaning device during the cleaning process and in the process of fine cleaning of the cleaning position, by generating a detailed cleaning strategy according to the input different tooth and gum models, the fine degree of cleaning is improved.

[0079] Furthermore, the above tooth and gum models include information such as the shape of the tooth crown and the extension curvature of the tooth crown surface, the thickness of the gum at the connection between the gum and the tooth, and the gap formed between the tooth and the gum. Since the connection position of the tooth and the gum, the protrusion of the gum, and the gap between two adjacent teeth vary from person to person. For example, the tooth surface curvature from the tooth to the gum position is a smooth curved surface. According to the connection position between the tooth and the gum, a surface curvature model of the tooth can be generated, which can be used to judge the existing dental calculus and food residues according to the connection position between the tooth and the gum, which is beneficial to image judgment and more beneficial to the optimization of the cleaning path and strategy of the cleaning device. And the above learning method can identify and input different-shaped tooth models or the actually scanned tooth shapes through an image recognition device to establish a model database of various tooth and gum shapes.

[0080] Even further, according to the above, the training of the motion trajectory algorithm includes the following steps: Obtain the cleaning strategies for different types of dirt during the actual cleaning process; Obtain the dirt position, establish a coordinate system with the predetermined cleaning position, and the positional relationship between the cleaning device and the predetermined cleaning position, and obtain the dirt position. With the set motion trajectory of the cleaning device, according to the boundary model data of the tooth and the gum and the dirt position, adjust the precise motion trajectory.

[0081] A fine-controlled oral cleaning system includes a pre-positioning system, which in turn includes an image acquisition device, a cleaning device, and a control system for the operation of the cleaning device. The control system uses the cleaning method of the oral cleaning device, and the image acquisition device includes at least one camera.

[0082] A precise control oral cleaning system further includes a transmission module, a storage module, and a central control unit. All modules are connected through a standard interface to ensure real-time data transmission and feedback optimization, and are used to install the cleaning system. The cleaning system is used to perform a cleaning movement on the pre-positioned position of the pre-positioning system, and to execute a cleaning strategy and a cleaning movement trajectory.

[0083] Further, the pre-positioning system includes: An input module, including one or more image acquisition devices, and is used to input the images of teeth and gums in the oral cavity obtained by the image acquisition devices into the storage module and the CPU for further data processing; A transmission module, used to transmit digital signals of the control terminal, the oral cleaning device, and the pre-positioning system; A storage module, used to store the images of teeth and gums in the oral cavity, the program for running the pre-positioning system, and the cleaning strategies and methods for teeth and gums in the oral cavity; The CPU is connected to and controls the above-mentioned input module, transmission module, and storage module, and is used to execute the instructions of the running program; The control terminal is used to control the cleaning device and adjust the predetermined cleaning position, and realizes the transmission and processing of data information through the interaction module.

[0084] A precise control oral cleaning device includes at least one oral cleaning device for performing a precise control oral cleaning method. The oral cleaning device executes the running program and program instructions of the precise control oral cleaning system.

[0085] According to the above structure, by running the running program and program instructions of the precise control oral cleaning system through the cleaning device, a rotating shaft with universal adjustment on the cleaning device is used to adjust the rotation and telescoping of the cleaning head of the cleaning device. When there are two cleaning devices, the cooperation between the two cleaning heads can be achieved. In addition, when there are two cleaning devices, one of them can be an ultrasonic cleaning device for collaborative cleaning with the water flow rinsing head to make the cleaning efficiency higher.

[0086] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. The structures, devices, and operation methods not specifically described and explained in the present invention, unless otherwise specified and limited, are implemented according to the conventional means in the art.

Claims

1. A finely controlled oral cleaning method, characterized in that: The steps include: Obtain images of teeth and gums inside the mouth; Acquiring a predetermined cleaning position based on a first image processing algorithm, and analyzing the predetermined cleaning position to generate a cleaning path; Controlling the cleaning device to aim at the predetermined cleaning position to perform preliminary cleaning; Acquire the actual cleaning position of the cleaning device by a second image processing algorithm, compare it with the predetermined cleaning position, calculate the deviation between the actual cleaning position and the predetermined cleaning position, and adjust the cleaning path based on the deviation; The movement trajectory and speed of the cleaning device are controlled according to the cleaning path to complete fine cleaning.

2. A finely controlled oral cleaning method according to claim 1, characterized in that: The preliminary cleaning uses a small force water flow to clean the predetermined cleaning position in the oral cavity, and optimizes and adjusts the water flow force according to the characteristics of the predetermined cleaning position to perform subsequent cleaning on the predetermined cleaning position.

3. A finely controlled oral cleaning method according to claim 1, characterized in that: The predetermined cleaning position determination method comprises the following steps: A coordinate system is established, and the multi-view images of the predetermined cleaning position are analyzed, and the coordinates of the predetermined cleaning position are obtained by triangulation, wherein the multi-view images can be obtained by a multi-camera or by moving a monocular camera.

4. A finely controlled oral cleaning method according to claim 3, characterized in that: Setting the motion trajectory includes the following steps: Based on the acquired images of teeth and gums in the oral cavity, the coordinate system established at the predetermined cleaning position, and the data information of the predetermined cleaning position, a cleaning strategy and method for the teeth and gums in the oral cavity during the actual cleaning process is obtained; The cleaning trajectory is set according to the cleaning strategy and method, wherein the cleaning strategy includes at least the Manhattan strategy and the European strategy. The Manhattan strategy connects adjacent cleaning points by step-by-step movement to form a right-angled polyline trajectory; The Euclidean strategy connects adjacent cleaning points by moving in a straight line to form a shortest path trajectory.

5. A finely controlled oral cleaning method according to claim 1, characterized in that: The algorithm optimizes the motion control of the cleaning device, including the following settings: Status: cleaning target status, target water spraying position and actual water spraying position; Action: cleaning device status and motor motion control parameters; Reward mechanism: the deviation between the target water spraying position and the actual water spraying position, including sparse rewards based on cleaning efficiency and cleaning effect; During the reinforcement learning training phase, its decision-making strategy is optimized through repeated attempts to clean and learn in order to maximize the cumulative reward.

6. A finely controlled oral cleaning method according to claim 5, characterized in that: The algorithm is applied to achieve feedback on the motion control strategy of the cleaning device, and the decision is adjusted according to the real-time feedback to feedback and optimize the cleaning efficiency and cleaning completion of the cleaning device.

7. A finely controlled oral cleaning system, characterized in that: It comprises a pre-positioning system, which comprises an image acquisition device, a cleaning device, and a control system for the operation of the cleaning device, wherein the control system uses the oral cleaning method according to any one of claims 1 to 6, and the image acquisition device comprises at least one camera.

8. A fine-control oral cleaning system according to claim 7, characterized in that: Also includes: Transmission module, storage module and central control unit, all modules are connected through standard interfaces to ensure real-time data transmission and feedback optimization.

9. A finely controlled oral cleaning system according to claim 7, characterized in that: The pre-positioning system comprises: An input module, used to input the images of teeth and gums in the oral cavity acquired by an image acquisition device; A transmission module, used to transmit digital signals of the control terminal, the oral cleaning device and the pre-positioning system; A storage module, used for acquiring the intraoral teeth and gums images, running the program of the pre-positioning system and the intraoral teeth and gums cleaning strategy and method; CPU, connected to and controlling the above-mentioned input module, transmission module, storage module, and instructions for executing the running program; The control end is used to control the cleaning device and adjust the predetermined cleaning position, and realize the transmission and processing of data information through the interactive module.

10. A finely controlled oral cleaning device, characterized in that: The invention comprises at least one oral cleaning device for executing the fine-control oral cleaning method, wherein the oral cleaning device executes the operation program and program instructions of the fine-control oral cleaning system.

Images